Service Delivery by Health Facilities in Papua New Guinea Report based on a countrywide health facility survey Report No: AUS0000113 © 2017 The World Bank 1818 H Street NW, Washington DC 20433 Telephone: 202-473-1000; Internet: www.worldbank.org Some rights reserved This work is a product of the staff of The World Bank. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of the Executive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and Permissions The material in this work is subject to copyright. Because The World Bank encourages dissemination of its knowledge, this work may be reproduced, in whole or in part, for noncommercial purposes as long as full attribution to this work is given. Attribution—Please cite the work as follows: “World Bank. 2018. Service Delivery by Health Facilities in Papua New Guinea. © World Bank.” All queries on rights and licenses, including subsidiary rights, should be addressed to World Bank Publications, The World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522-2625; e-mail: pubrights@worldbank.org. Service Delivery by Health Facilities in Papua New Guinea CONTENTS Abbreviations xi Acknowledgments xiii Overview 1 Chapter 1: Introduction and Health System of Papua New Guinea 15 Challenges to PNG’s Fiscal Space for Health 15 The Basics of PNG’s Health System 17 The National Health Plan and Role of Level 5 to 7 Facilities 20 Objectives of the Study and Report Outline 20 Chapter 2: Survey Design 21 Survey Instruments 21 Sample and Sampling Design 21 Limitations 23 Chapter 3: Input-Related Aspects of Health Facilities 25 Health Facility Infrastructure Inputs 25 Instruments and Medical Supplies 27   Maternal and Neonatal Equipment and Materials 27   Family Planning, Vaccines, and Other Items 27   Availability of Drugs 28 Clinical and Patient Education Protocols 30 Service Availability in Health Facilities 32 Human Resources 36   Working Hours and Workload 38   Health Care Provider Satisfaction 38   Provider Knowledge Index 38 Summary 40   Facility Readiness 40   Human Resources for Health 41 iii iv Contents Chapter 4: Management, Access, Equity, Quality, and Safety of Health Service Delivery 43 Leadership and Governance 43 Access to Health Care Services 46 Safety-Related Indicators 48 Patient Satisfaction 50   Outpatient services 50   Inpatient Services 50 Equity-Related Indicators 53 Summary 53   Health Facility Autonomy: Expenses, Personnel, and Services 53   Health Facility Safety 54   Patient Satisfaction 54  Equity 54 Chapter 5: Cost and Output of Facilities 55 Outputs of Health Facilities: Quantities of Various Services Produced 55 Budget, Expenditure, and Financing of Health Facilities 59 Budget and Expenditure Comparisons 60 User Fees and Revenue Generation 61 Summary 61   Outpatient Consultation and Inpatient Admissions 63   Outreach Activities 63   Cost of Running a Health Facility 63   Budget and Expenditure 63 Chapter 6: Efficiency of Health Facilities Surveyed 65 Specific Ratios and Proportions as Efficiency Measures 65   Bed Occupancy Rate 65   Personnel-Based Performance Measures 66   Cost-Based Performance Indicators 68   Efficiency Measure Based on Multiple Ratios 69 Production and Cost Function-Based Efficiency Measures 71 Summary 75 Chapter 7: Policy Implications 77 Policy Recommendations to Improve the Key Enablers for Service Delivery 77 Policy Recommendations on Key Priority Service Areas 78 Policy Recommendations to Improve Efficiency 79 Implications of Technical Assistance to NDOH 80 References 81 Annex A: World Health Organization Building Block Framework 83 World Health Organization Building Block Framework for Performance Measurement 83 Indicators of Performance of Health Facilities Based on WHO Framework 83 Annex B: Provider Interview Results 85 Methodology of Selecting Health Care Providers 85 Demographic Characteristics of Health Care Providers Surveyed 85 Type of Services Provided by Health Care Providers 86 Staff Trainings 88 Contents v Health Care Provider Satisfaction 88 Motivation of Health Care Providers 90 Salary and Benefits of Providers 90 Provider Knowledge Index 91 Hospital Bed Occupancy Rate 94 Annex C: Exit Interview of Patients or Caretakers 95 Figures O.1 Infrastructural Indexes for Health Facilities Surveyed 2 O.2 Comparison of Availability Indexes by Facility Type 3 O.3 Comparison of Service Availability and Readiness Indexes by Health Facility Level and Type 3 O.4 Distribution of Human Resource Cadres by Facility Level 4 O.5 Average Weekly Hours Worked at Health Facility and Outreach by Facility Level 5 O.6 Patient Satisfaction by Health Facility Level 7 O.7 Common Reasons for Seeking Treatment at Outpatient Clinics 8 O.8 Breakdown of Health Facility Costs by Health Facility Level 9 O.9 Plot of Health Facilities by Bed Occupancy and Turnover Rates 10 1.1 Total Health Expenditure as a Proportion of GDP, 1996–2014 16 1.2 Health Financing Mix in PNG, 2014 16 1.3 External Resources on Health as a Percent of Total Health Expenditure, 1995–2014 17 1.4 PNG National Health Plan Framework 20 3.1 Infrastructural Indexes for Health Facilities Surveyed 27 3.2 Comparison of Availability Indexes, by Facility Type 29 4.19 Percentage of Clients from Low-Income Households, by Facility Level 53 5.1 Percent of Inpatients in Health Facility Wards, by Facility Type 59 5.2 Average Cost of Running a Facility in PNG (in Millions of PNG Kina [PGK]), 2014 60 5.2 Percent Distribution of Recurrent Health Facility Expenses, by Facility Type 61 6.1 Cost of Inputs and Outputs per Unit, by Facility Type 68 6.2 Comparison of Cost per Unit Output for Church-Run and Public Level 3 and 69   4 Health Facilities in PNG (2014 Costs) 6.3 Cost per Output for Level 5 to 7 Facilities in PNG (Cost for the Year 2014) 70 6.4 Plot of PNG Health Facilities, by Bed Occupancy and Bed Turnover Rates 70 6.5 Relationship Between Total Production and Total Number of Beds 72 6.6 Relationship Between Total Production and Total Personnel 72 6.7 Relationship Between Total Production and Total Number of Annual Patrols Conducted 73 6.8 Free-hand Production Frontier (Output vs. Cost) and the Relative Inefficiency 75   of Health Facilities in PNG A.1 WHO System Building Blocks and the Overall Health System Goals and Outcomes 84 Tables O.1 Percent of Budgeted Vacant Positions by Personnel Category 4 O.2 Total Number of Females in Each Personnel Category by Facility Type 5 O.3 Measure of Facility Authority on Fund Allocation, Personnel, and Service Delivery 6 O.4 Comparison of Health Facility Safety Indicators 6 O.4 Estimated Average Annual Output of Health Facilities by Level 7 O.6 Average Annual Number of Outreach Activities by Health Facility Level 8 O.7 Average Annual Cost to Run a Health Facility 9 O.8 Average User Fees by Health Facility Level 9 O.9 Average Bed Occupancy Rates by Method 9 vi Contents 3.1 Basic Infrastructural Variables by Facility Category 26 3.2 Infrastructure-Related Variables Affecting Provision of Services and Quality of Care 26 3.3 Availability of Maternal Health-Related Equipment and Supplies 28 3.4 Availability of Family Planning, Vaccine, Test Kits, and Other Items, by Facility Category 29 3.5 Facilities Reporting Continuous Drug Availability Over Past 30 Days, by Facility Category 31 3.6 Average Number of Drug Kits Received per Facility, by Facility Category, 2014 32 3.7 Types of Protocols Present by Facility Category 33 3.8 Availability of Various Registry and Reports in Facilities, by Facility Type 33 3.9 General Indicators of Service Availability by Level and Ownership 34 3.10 Availability of Laboratory Tests in Surveyed Facilities, by Level and Ownership 34 3.11 Service Provision Readiness Index, by Facility Level 35 3.12 Average Number of Personnel in Level 3 and 4 Facilities, by Personnel Type 36 3.13 Average Number of Personnel in Level 5 to 7 Facilities, by Personnel Type 37 3.14 Total Number of Female Personnel and Percentage of Female Personnel, 37 by Facility Category and Cadre 3.15 Percent of Providers Reporting a Second Job Outside the Surveyed Facility, 38 by Provider Type and Facility Level 3.16 Health Care Providers Satisfaction Scores, by Facility Type 39 3.17 Three Most Important Factors Affecting Provider Satisfaction 39 3.18 Motivation Index, by Dimension, Facility Type, and Provider Type 39 3.19 Percentage of Health Staff Providing Priority Health Services 40 3.20 Knowledge Index Related to Child Health Services, by Provider Type and Facility Level 40 3.21 Knowledge Index Related to Maternal Health Services, by Provider Type and Facility Level 41 4.1 Health Facility Management and Governance Index, by Facility Level and Ownership 44 4.2 Facility Authority Index Related to Decision Making on Expenditure, Personnel, 45 and Service Delivery 4.3 Days Since Last Supervisory Interaction, by Facility Type 45 4.4 Supervisor Activities at Most Recent Supervisory Visit 46 4.5 Types of 24/7 Services Available in Surveyed Health Facilities, by Region 47 4.6 Patient Access to Health Care Facilities, by Region 47 4.7 Health Service Coverage Provided in Catchment and Remote Areas, by Region 47 4.8 Percentage of Health Facilities with Expired Drugs Present 48 4.9 Percentage of Health Facilities Providing Immunization Without Specific 48 Cold-Chain Equipment 4.10 Percentage of Health Facilities Providing Obstetric First Aid, Basic Emergency 49 Obstetric Care, and Comprehensive Emergency Obstetric Care 4.11 Medical Waste Disposal Equipment and System Availability, by Level and Ownership 50 4.12 Health Facility Disposal Practice for Sharps 50 4.13 Health Facility Disposal Practice for Biomedical Waste 50 4.14 Outpatient Clinic Satisfaction Scores 51 4.15 Inpatient Admission Process Satisfaction Scores 51 4.16 Inpatient Satisfaction Scores for Doctors’ Services 52 4.17 Inpatient Satisfaction Scores for Nurse and Support Staff, Housekeeping, and Food Services 52 4.18 Inpatient Perceptions of Health Facility Quality and Security 53 5.1 Annual Average Maternal and Child Health Services Delivered, by Facilities 56 5.2 Annual Average Outpatient Visits, by Facility Type 56 5.3 Annual Average Discharges, by Facility Category 57 5.4 Average Annual Health Facility Outputs, by Facility Type 58 5.5 Estimated Child Vaccinations Delivered, by Facility and Outreach Activities 59 5.6 Average Facility-Level Costs, by Cost Categories (PNG Kina) 61 5.7 Budget and Expenditure Comparison for Level 3 and 4 Health Facilities, PNG Kina 62 Contents vii 5.8 Budget and Expenditure Comparison for Level 5 and 6 and Level 7 Health Facilities, 62 PNG Kina 5.9 Average Annual User Fees Collected, by Facility Type, PNG Kina 62 6.1 PNG Hospital Inpatient Days and Bed Occupancy Rates, 2014 66 6.2 Select Performance Measures Based on Inputs and Output 66 6.3 Productivity of Health Care Providers in Surveyed Facilities by Facility Level 67 (Output per Day), 2014 6.4 Distribution of Health Facilities, by Bed Size and Efficiency 70 6.5 Distribution of PNG Health Facilities, by Location, Ownership, and Efficiency 71 6.6 Distribution of Health Facility Efficiency Scores, by Level and Ownership 74 B.1 PNG Health Care Providers Surveyed, by Provider Type and Facility Level 86 B.2 Distribution of PNG Health Care Providers Interviewed, by Gender and Provider Type 86 B.3 Age Distribution of Interviewed PNG Health Care Providers, by Provider Type 86 B.4 Types of Primary Health Care Provider Services in Level 3 and 4 Facilities, by Provider Type 87 B.5 Types of Primary Health Care Provider Services in Level 5 to 7 Facilities, by Provider Type 87 B.6 Percentage of PNG Health Care Providers Reporting No Training in Previous 3 Years, 88 by Facility Level B.7 PNG Health Care Provider Training Categories, by Subject Matter 88 B.8 PNG Health Care Providers’ Satisfaction Scores, by Facility Type and Health Care 89 Provider Type B.9 Factors Affecting PNG Health Care Provider Satisfaction, by Provider Type and 89 Situation and Aspect Type B.10 Top Three Factor Categories Affecting PNG Health Care Provider Satisfaction 90 B.11 Index of Motivation, by Provider and Facility Types 90 B.12 Percent of PNG Health Care Providers Involved in the Provision of Priority Health Services 91 B.13 Knowledge Index Related to Child Health Services, by Provider Type 91 C.1 Frequencies, by Level and Ownership 95 C.2 Mean Number of Medicines Listed in the Prescription, by Level and Ownership 95 C.3 Mean Medication Dispensed from Health Facilities, by Level and Ownership 95 C.4 Mean Satisfaction Score, by Level and Ownership 95 C.5 Mean Overall Satisfaction, by Level and Ownership 95 C.6 Reasons for Seeking Treatment for Children, by Level and Ownership 96 Map 2.1 Map of PNG with Districts Visited (green) and Location of Surveyed Facilities 22 (red and blue squares) ABBREVIATIONS AIDS acquired immunodeficiency syndrome ANC antenatal care ARB Autonomous Region of Bougainville ARI acute respiratory illnesses BCG bacillus Calmette-Guérin (a vaccine for tuberculosis) BEmOC basic emergency obstetric care CEmOC comprehensive emergency obstetric care CEO chief executive officer CHW community health worker DFF direct facility funding EmOC emergency obstetric care EPI extended program on immunization FBB facility-based budgeting GDP gross domestic product GFATM Global Fund to Fight AIDS, Tuberculosis and Malaria HEO health extension officer HIV human immunodeficiency virus IMCI integrated management of childhood illnesses IUD intrauterine device KRA key result areas MDR TB Multidrug-resistant tuberbulosis MO Medical officer NDOH National Department of Health NGI New Guinea Islands NHIS National Health Information System NHP National Health Plan NHSS National Health Service Standards NO Nursing officer OBGYN obstetrics and gynecology ix x Abbreviations OECD Organisation for Economic Co-operation and Development OOP out-of-pocket PFM public financial management PHA Provincial Health Authority PMGH Port Moresby General Hospital PNC postnatal care PNG Papua New Guinea PPP purchasing power parity RHEO resident health extension officer RMO resident medical officer RDT rapid diagnostic test SHP Southern Highlands Province SMO specialist medical officer SSMO senior specialist medical officer STI sexually transmitted infection TA technical assistance TB Tuberculosis TT tetanus toxoid WDI World Development Indicators WHO World Health Organization WHP Western Highlands Province XDR TB extensively drug-resistant tuberculosis ACKNOWLEDGMENTS This analytical and advisory activity is one of the sub- Department of Foreign Affairs and Trade for pro- tasks under the Papua New Guinea Programmatic viding thorough comments and peer reviewers Health Advisory Services and Analytics. Reem Hafez (Senior Health Economist) and Mickey This report was prepared by Xiaohui Hou (Senior Chopra (Lead Health Specialist and Global Solution Economist and Task Team Leader), M. Mahmud Lead in Service Delivery). Khan (Principal Investigator, University of South Toomas Palu (Practice Manager of East Asia and Carolina, Columbia, SC, USA), Justin Pulford Pacific Region, Health, Nutrition and Population (Co-PI, Papua New Guinea Institute of Medical Global Practice) provided technical comments and Research), Olga Saweri (Papua New Guinea Institute overall supervision on this report. The team also of Medical Research), Ibrahim Demir (University acknowledges Michel Kerf (Country Director for of South Carolina, Columbia, SC), Rifat Haider Papua New Guinea and Pacific Islands, East Asia (University of South Carolina, Columbia, SC), and and Pacific Region) and Patricia Veevers-Carter Shakil Ahmed (Nossal Institute for Global Health, (Country Manager for Papua New Guinea) for their University of Melbourne, Australia). overall guidance and support. The team thanks Colin Wiltshire (consultant) and The authors sincerely thank Papua New Alan Cairns (consultant) for their contribution in Guinea National Department of Health, Papua drafting the questionnaires and providing comments New Guinea Institute of Medical Research, health on earlier drafts. The team also thanks Aneesa development partners in Papua New Guinea and Arur, Laurin Janes, Maude Archambault, Nicolas health facilities surveyed for their assistance and Rosemberg, Pranita Sharma, Susan Ivatts and Venki insightful comments provided in various stages of Sundararaman for providing useful comments, this project. Melody Molinoff for her editing, and Tasha Sinai Financial support for this work was received from and Abigail Abbott Blenkin for their assistance. The the Australian Department of Foreign Affairs and team also thanks colleagues from the Australian Trade. M. Mahmud Khan, Xiaohui Hou, Olga PM Saweri, Ibrahim Demir, Rifat Haider, Shakil Ahmed and Justin Pulford, December 2017. Service Delivery by Health Facilities in Papua New Guinea, Washington DC: The World Bank Group. xi OVERVIEW Introduction facing underscore the need to prioritize financing within the health sector and improve efficiencies to Papua New Guinea (PNG) has seen some improve- make the most of the allocated funding. ments in health indicators over the past 25 years, A better understanding of service delivery but the pace of improvements is not as robust is critical to help PNG achieve the goals set in as expected. PNG has fallen short of Millennium the PNG National Health Plan (2011–2020). The Development Goal targets on maternal and under- Health Plan was developed to achieve two main five mortality rates. The high rates of multidrug- goals: strengthen primary health care (for all) and resistant and extensively drug-resistant tuberculosis improve service delivery to the rural majority and (TB) point to a very weak health system, particularly the urban dis­ advantaged. These goals imply that the delivery of frontline health services. The health quality health care services should be readily acces- system in PNG is likely to weaken further because sible and affordable for all, especially the poorer public spending on health is projected to decline segments of the population. The plan also identi- significantly in the short to medium term. fied priority health care services, which included Public expenditure accounts for about 80 per- maternal health care, child health, and control of cent of total health expenditure in PNG, implying infectious diseases. relatively low out-of-pocket spending. In real terms, The purpose of this study is to investigate the total health expenditure per capita has been declin- inputs and outputs of health services delivery and ing since 2004, in part because the absolute increase the relative efficiency of resources used across in total health expenditure has been offset by pop­ the facilities. Specifically, the study examines the ulation growth. Due to present fiscal constraints, it extent to which the upper-level primary and sec- is unlikely that the budgetary allocation to health ondary PNG health care facilities have an adequate will see a significant increase in the near future. level of resources, both human resources and other Total health expenditure per capita is already low material inputs, to be able to deliver quality health by regional standards and is expected to decline care services and to determine whether the inputs further. are combined in an efficient manner to produce External financing on health is declining. On quality health care services. average, 20 percent of total annual health spend- A primary survey was conducted at all second- ing in PNG is from development partners; fund- ary (levels 5 and 6), tertiary (level 7), and a random ing is volatile with regard to levels, sources, and selection of functional upper-primary-level health recipients. Australia is the largest bilateral donor. facilities (levels 3 and 4). To understand differences The Global Fund to Fight AIDS, Tuberculosis and in the management structure and efficiency of pub- Malaria (GFATM) and Gavi, the Vaccine Alliance, have lic sector and church-run level 3 and 4 facilities, an become increasingly important partners in health attempt was made to sample equal numbers of pub- service delivery in PNG. The challenges that PNG is lic and church facilities at these levels. 1 2 Service Delivery by Health Facilities in Papua New Guinea Figure O.1  Infrastructural Indexes for Health Facilities Surveyed items to level 3 and 4 public health facilities. Level 3 100.0 100.0 100.0 and 4 public facilities show lower availability for all 100.0 100.0 the different item types compared with level 3 and 4 89.7 90.0 church-run facilities (figure O.2). A significant num- 79.5 80.0 75.6 73.7 ber of health facilities lack very basic maternal and 70.0 65.6 66.0 neonatal equipment. The vaccine availability index 62.0 varied from 73.7 percent for level 3 and 4 public sec- 57.6 60.0 51.3 tor facilities to 100 percent for the level 7 facility. The 50.0 46.4 38.6 38.6 most limited vaccine supply, as reported by the facili- 40.0 ties, was the bacillus Calmette-Guérin (BCG) vaccine. 30.0 Level 5 to 7 facilities are usually better off with regard 20.0 to availability of items and equipment, and level 7, in general, shows the best availability indexes. 10.0 Higher-level public health facilities can provide 0.0 Condition of No/few repairs Condition: other Condition: more types of services than lower-level facilities, toilet needed infrastructural overall and church health facilities provide more service Level 3-4 Public Level 3-4 Church Level 5-6 Level 7 types than public health facilities at comparable Note: higher scores indicate better condition. levels (figure O.3, column 1). The overall service availability index was 55.9 for level 3 and 4 public facilities, 67.4 for level 3 and 4 church-run facilities, 82.5 for level 5 and 6 facilities, and 100.0 for the Key Survey Results level 7 facility. Again, among level 3 and 4 facilities, Inputs Used in the Production church-run facilities appear to be better prepared to of Health Care Services provide various types of medical care services. More than 90 percent of level 5 to 7 health facili- Facility Readiness ties reported the ability to do all the laboratory tests The overall health facility infrastructure condition listed in the table, but the ability of public level 3 and is quite poor. Most facilities reported the need for 4 facilities to perform basic tests was woefully lack- major building repairs and emphasized the lack of ing (figure O.3, column 2). Almost all level 3 and 4 adequate toilets, stable electricity supply, and consis- facilities reported the ability to conduct malaria test- tent water supply. Many of the level 3 and 4 health ing, either by rapid diagnostic test or by microscopy,2 facilities require better connectivity to electricity and but were not able to conduct many other critical clean water supply. Figure O.1 shows the infrastruc- tests, such as pregnancy and anemia tests. The labo- tural indexes1 by facility type. Higher index values ratory service availability index (below) summarizes imply better infrastructural condition. On average, the availability of selected pathology services in the the indexes are lower for level 3 and 4 public health facilities and shows the proportion of laboratory facilities than for level 3 and 4 church-run facilities. tests that the facilities in the specific category could An assessment of family planning items, vaccines, do on the day of the survey. The index for laboratory test kits, and other related items stocked in health tests was found to be 100 for the level 7 facility and facilities revealed the need to improve the supply of 95 for level 5 and 6 facilities, implying that all the tests considered were available in the level 7 facility; 1 95 percent of the tests, on average, were available in The indexes are measured to show the desirable or positive aspects of infrastructural condition or availability. For example, level 5 to 7 facilities. The indexes were 54.5 for level to construct the infrastructural index for toilets the following vari- 3 and 4 church-run facilities, and 29.7 for level 3 and ables were combined: adequate number of toilets present for the 4 public facilities. use of patients (= 1, if not adequate = 0), most toilets are func- tional, not many repairs needed in toilets of the facility, and most The quantity and quality of different types of toilets have water. Similarly, infrastructural variables included service provision crucially depend on the facil- the variables such as the facility has visible signboard, facility has ity’s readiness to provide the clinical services. A electric supply from electric main, facility ground is clean, there is staff housing on campus, the facility has designated space for clini- cians to provide services, the facility has an ambulance, and main 2 building is accessible to persons with disabilities. The questionnaire was not specific. Overview 3 Figure O.2  Comparison of Availability Indexes by Facility Type 0 10 20 30 40 50 60 70 80 90 100 Antenatal care items Maternal/neonatal items (excluding ANC) Family Planning items Vaccine Test kits and other items Level 3-4 Public Level 3-4 Church Level 5-6 Level 7 Note: Best possible index is 100. Higher value of index implies availability of more items among all possible items considered. ANC = antenatal care. readiness index3 describes the scale to which health types of drugs reported to be supplied in insufficient facilities are capable of providing services. Overall, quantities were antibiotics, analgesics and antipyret- level 3 and 4 facilities, both public and church-run, ics, antimalarials and antacids. score quite low. The readiness indexes were found to be 40.3 for level 3 and 4 public sector facilities, 48.6 Human Resources for Health for level 3 and 4 church-run facilities, 84.6 for level 5 PNG has an acute shortage of doctors, particu- and 6 facilities, and 100 percent for the level 7 facility. larly in lower-level facilities. Figure O.4 illustrates Drug availability indexes4 were quite low for all the average number of personnel employed at the the health facility categories (figure O.3, column 4). various facility levels. Nursing officers and commu- For example, the drug availability index was only nity health workers (CHWs) make up the major- 49.1 for level 3 and 4 public sector health facilities, ity of the workforce employed at health facilities, 54.3 for level 3 and 4 church-run facilities, 59.9 for if nonclinical staff are not counted. The number of level 5 and 6 facilities, and 62.2 for the level 7 facility. nonclinical staff is quite high in all facility types. Some basic drugs were reported not available on a Level 5 to 7 health facilities tend to complement continuous basis in many health facilities, including doctors with trained clinical “specialist” staff, such as the level 7 facility. Poor availability of paracetamol elixir across different levels of health facilities also points to the management and supply issues of basic Figure O.3  Comparison of Service Availability and Readiness drugs. In general, drug stock-outs at the facilities Indexes by Health Facility Level and Type surveyed indicate massive improvement required for 120 supply chain and drug procurement. The top three 100 80 3 The readiness index reflects the percent of all readiness aspects 60 or variables satisfied by the health facilities in a specific category. 40 For example, if the readiness index of a facility category is 32.5, it means that 32.5 percent of all the “readiness” indicators were satis- 20 fied on the average for the facility category. 4 Drug availability is defined as continuous availability of each of 0 General Service Laboratory Services Readiness Drugs Availability the drugs in the past 30 days without any stock-outs. In the survey, Availability Availability Index Index data were collected on drug availability for 45 different drugs and Index Index all these 45 drugs are listed in the table 3.5. This index was con- structed when all 45 drugs were considered. 3-4 Public 3-4 Church Level 5-6 Level 7 4 Service Delivery by Health Facilities in Papua New Guinea Figure O.4  Distribution of Human Resource Cadres by Facility Level 0.27 0.88 Level 7 11.96 3.94 43.03 15.98 5.78 18.15 0.42 2.15 2.62 Level 5-6 7.71 29.46 18.75 7.54 31.34 2.05 2.05 1.79 3-4 Church 31.11 32.39 5.12 25.48 2.11 3.08 3-4 Public 5.70 22.02 39.20 6.29 21.59 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Doctors Residents HEOs Nurses CHWs Other clinical staff Allied health workers Non-clinical staff radiographers, physiotherapists, and anesthetic tech- nurses, CHWs, midwives, and kitchen staff. At the nical officers, but these types of specialist staff are time of data collection, there were only two female often not present at level 3 and 4 facilities. chief executive officers (either at district or provin- A significant number of budgeted facility posi- cial level). tions remained vacant (table O.1). Only a small per- Health workers have similar weekly working centage of level 3 and 4 health facilities were allocated hours with the exception of doctors at the level 7 medical officer positions, but more than 40 percent of facility; outreach varies by facility level (figure O.5). these positions remained unfilled. Even in level 5 to In terms of hours spent on outreach activities, nurses 7 facilities, about a quarter of medical officer posi- and CHWs in level 3 and 4 health facilities allocated a tions were vacant, implying an overall shortage of higher number of hours for outreach patrols, whereas doctors in the country. Overall, more than 20 percent doctors reported greater involvement with outreach of nursing officer (NO) and health extension officer activities at level 5 and 6 facilities. The level 7 health (HEO) positions were not filled. facility did not conduct outreach in 2014. Gender disparities are apparent in employment Health workers consider their workload mod- patterns or personnel composition (table O.2). erately high or very high. This perception may not Females comprise 26 percent of all medical officers in be a direct reflection of hours worked, but perhaps PNG. Even among residents, only about 34 percent the patient load. Despite the perceived high work- were female, implying that the male-to-female ratio load, health workers are reasonably satisfied with of doctors will improve in the future but by a rela- their job. The determinants of satisfaction are not tively small amount. Overall, about 60 percent of all the same across health worker cadres or across health facility personnel were female and most were health facility levels. In the survey, health care pro- viders mentioned a number of specific aspects that affect their level of satisfaction, including the condi- Table O.1  Percent of Budgeted Vacant Positions by tion of health facility infrastructure and the avail- Personnel Category ability of medicine, equipment, and supplies. Level 3 Level 3 Most health care providers scored poorly in and and Level 5 4 public 4 church and 6 Level 7 the knowledge index, especially for maternal Medical 46.67 43.75 26.36 24.43 and child health issues. Health care providers were officers asked a series of questions to determine their level of HEOs 21.95 31.25 –3.91 25.00 knowledge on priority health services of PNG. The NO 25.95 20.16 16.09 31.28 questions were focused on three areas: child health, CHW 10.68 15.02 9.32 8.09 maternal health, and sexually transmitted infections Other clinical 20.00 10.00 16.93 32.94 (STIs). Most health care providers, including the Total 15.06 15.36 12.94 23.25 fully trained doctors, scored poorly in the knowledge personnel index, especially for maternal and child health issues. Overview 5 Table O. 2  Total Number of Females in Each Personnel Category by Facility Type Level 3 and 4 Level 3 and 4 public facilities church facilities Level 5 and 6 facilities Level 7 facility (tertiary) Personnel % % % % type Total Female Female Total Female Female Total Female Female Total Female Female Medical 7 0 0.00 5 2 40.00 279 74 26.52 129 33 25.58 officers Other doctors 1 0 0.00 1 0 0.00 55 18 32.73 4 2 50.00 Residents 0 0 — 0 0 — 30 10 33.33 40 14 35.00 HEO and 32 17 53.13 11 4 36.36 133 64 48.12 3 1 33.33 RHEO Nursing 117 94 80.34 190 151 79.47 1345 1155 85.87 168 104 61.90 officers Community 251 160 63.75 208 143 68.75 960 759 79.06 216 127 58.80 health workers Midwife 12 12 100.00 16 16 100.00 97 97 100.00 8 8 100.00 Dental 7 3 42.86 3 1 33.33 45 20 44.44 15 9 60.00 personnel Pharmacy 17 2 11.76 21 5 23.81 161 69 42.86 31 19 61.29 and lab personnel Office staff 23 15 65.22 24 15 62.50 276 176 63.77 34 20 58.82 Driver 30 1 3.33 27 0 0.00 104 0 0.00 13 0 0.00 Cleaner/ 41 16 39.02 50 26 52.00 267 118 44.19 12 4 33.33 handyman Kitchen staff 8 2 25.00 8 5 62.50 111 68 61.26 7 2 28.57 Other staff 68 19 27.94 73 18 24.66 1362 639 46.92 440 214 48.64 Total 614 341 55.54 637 386 60.60 5225 3267 62.53 1120 557 49.73 Note: RHEO = resident health extension officer. Figure O.5  Average Weekly Hours Worked at Health Facility and Outreach by Facility Level 70 60 50 40 30 20 outreach outreach 10 outreach outreach 0 3-4 Public 3-4 Church Level 5-6 Level 7 Doctors HEOs Nurses CHWs 6 Service Delivery by Health Facilities in Papua New Guinea Table O.3  Measure of Facility Autonomy on Fund Allocation, 4 church-run facilities, and 55.2 percent of level 3 and Personnel, and Service Delivery 4 public sector facilities reported their satisfaction. Level 3 Level 3 Level 5 to 7 and level 3 and 4 church-run health and and Level 5 facility personnel are more satisfied with the level of Authority related to 4 public 4 church to 7 autonomy that they have compared with level 3 and Funds 30.5 47.3 75.4 4 government-run health facilities. Personnel 47.0 54.5 82.9 Services 44.8 48.0 80.5 Health Facility Safety Satisfaction with level of 55.2% 84.0% 89.5% Level 5 to 7 health facilities scored better than lower- authority facilities have level facilities on a health facility safety assessment, highlighting important safety gaps in emergency obstetric care at lower levels (table O.4). This study Knowledge indices results were better for STIs, measured a number of safety indicators using the although the scores remained around 80 (80 out facility survey data. Most of level 3 and 4 facilities did of 100). Strengthening in-service trainings should be not have a blood bank. The presence of expired drugs considered to improve the knowledge base of health in more than 10 percent of facilities is also a concern. care providers. Few facilities did not have all the equipment and sup- plies needed for a child vaccination program. The Health Facility Autonomy, Safety availability of emergency obstetric care (EmOC) was and Client Satisfaction relatively low in level 3 and 4 health facilities. Nonavailability of comprehensive EmOC Health Facility Autonomy: Expenses, (CEmOC) in most level 3 and 4 facilities is especially Personnel, and Services concerning. Given that PNG has a high fertility rate, Level 5 to 7 health facilities have relatively more emergency obstetric service is essential for improving autonomy than level 3 and 4 facilities. In general, maternal health services. Policy makers should care- church facilities reported a higher degree of auton- fully evaluate how to strengthen the availability and omy than the level 3 and 4 public facilities (table O.3). provision of emergency obstetric services. Health facility autonomy is determined by the inde- pendent “control” that the management has over health Patient Satisfaction facility expenses, staff, and the services offered. The Patient satisfaction was similar across the facility survey also asked the facility managers whether they types, although level 3 and 4 church-run facilities were happy with their level of autonomy; 89.5 percent and the level 7 facility performed better than others. of level 5 to 7 health facilities, 84 percent of level 3 and Patient satisfaction was evaluated using a wide range of factors such as ease of access, cost, cleanliness, and Table O.4  Comparison of Health Facility Safety Indicators interactions with the health facility staff (figure O.6). Level 3 Level 3 Level 3 and 4 public sector facilities show the lowest and and Level 5 average satisfaction followed by level 5 and 6 facilities. Safety indicator 4 public 4 church to 7 In one specific dimension, the dimension related to Blood bank present 3.4% (1) 8% (2) 100.0% (18)* cost of obtaining services, level 3 and 4 public sector Percent of expired drugs 70% 40% 50% facilities performed better than others. found in ≥ 10% of facilities Cold chain-related indicators Equity % facilities with functional 88.9% 92.0% 100.0% The survey found that poorer households are vaccine thermometer more likely to seek medical care from level 3 and % facilities with functional 92.6% 88.0% 100.0% vaccine fridge 4 health facilities than other facility types. The frac- % facilities with functional 96.3% 88.0% 100.0% tion of treatment seekers who considered themselves vaccine carrier “poor”drastically declines with the increasing level of EmOC-related indicators the health care facility. The percentage of treatment Basic EmOC 63.0% 60.0% 88.9% seekers who are poor was 73 percent at level 3 health Comprehensive EmOC 11.1% 12.0% 83.3% facilities, whereas it was 10 percent at level 7. If wealth *Information on one facility was missing for level 5 to 7 facilities. scores are used (based on the ownership of assets Overview 7 Figure O.6  Patient Satisfaction by Health Facility Level 90 76.6 80 70.5 65.5 65.9 70 58.8 60 50 40 30 20 10 0 3-4 Public 3-4 Church Level 5 Level 6 Level 7 Type of services Admission Doctors' service Nurses service Treatment cost Overall visit Average score Note: Other includes diseases such as malnutrition, anemia, and skin diseases. and access to various amenities of life), the average gynecology wards were also important with regard to wealth score was the lowest for the users of level 3 and inpatient days. 4 church-run facilities and the wealth score increased The most important reason for seeking out- with increasing level of facilities. patient care was infectious disease (figure O.7). Respiratory illnesses were the leading cause of seek- Cost and Output of Health Facilities ing treatment from health facilities. For all health facility levels, aside from level 7, the second leading Outpatient Consultation and Inpatient Admissions cause of seeking treatment was malaria. At the ter- tiary hospital (level 7 facility), malaria was the sixth Health facility output is measured by the number leading cause of seeking care. Other common causes of outpatient consultations and inpatient days of of seeking treatment are diarrhea and accidents and stay. Annual output per health facility was estimated injuries. by using the output information of 1 representative month for each quarter (4 months total). As expected, outpatient visits and inpatient days vary widely by Outreach Activities facility level. Level 5 to 7 health facilities had significantly Level 3 and 4 public health facilities performed higher inpatient days than level 3 and 4 facilities poorly in conducting outreach activities. Health (table O.5). Average inpatient days in level 6 facili- facilities organize outreach activities to improve acces- ties were quite high compared with level 5 facili- sibility to health care services, especially for popula- ties. However, one should keep in mind that there tions living in remote areas. Outreach in PNG is quite are only three facilities in level 6 and 15 facilities in expensive and hard to sustain. Table O.6 shows the level 5. The length of stay also varies by the ward of percentage of facilities organizing outreach patrols admission, which is probably more of a reflection of and the average number of patrols planned and con- number of beds in different wards than the length ducted in 2014. Level 3 and 4 church-run facilities of stay in wards. Among the different wards, the organized the highest number of outreach patrols in medical and surgical wards experienced the highest 2014, followed by level 5 and 6 facilities. Even though level of inpatient stays. Pediatric and obstetrics and level 3 and 4 public sector facilities planned to conduct Table O.5  Estimated Average Annual Output of Health Facilities by Level Level 3 and Level 3 and 4 public 4 church Level 5 Level 6 Level 5 to 7 Total outpatient consultations 28,769 19,991 99,681 135,189 124,807 Total inpatient days 1,871 3,809 25,091 46,082 38,447 8 Service Delivery by Health Facilities in Papua New Guinea Figure O.7  Common Reasons for Seeking Treatment health facilities show lower average expenditure on at Outpatient Clinics salary and compensation than the public sector facili- 100 ties. Further, training is significantly lower in level 3 and 4 facilities compared with that for level 5 to 7 facilities, implying greater emphasis on training at 80 upper-level facilities. Drug expenses are relatively low in all facility types, especially in level 5 to 7 facili- 60 ties. The capital expenditure (without annualizing) was quite high for level 3 and 4 church-run facilities. 40 If the capital cost items are excluded, church facili- ties show the highest operational expenses in relative terms. These high expenses are probably due to the 20 significantly higher number of outreach patrols that church facilities have organized compared with other 0 facility types. 3-4 Public 3-4 Church Level 5-6 Level 7 infectious diseases Accident/Injury Budget and Expenditure NCDs Other* The percentage of budgets expended differs quite Note: Other includes diseases such as malnutrition, anemia, and skin diseases. drastically between health facility levels. Level 3 and 4 public health facilities spent 28 percent of their budget, whereas level 3 and 4 church health facili- ties spent 52 percent of their budget. These figures a relatively high number of outreach patrols, they only do not necessarily imply that the facilities were not conducted 27 percent of those planned. The most able to spend the money allocated to them. It is more common reason provided for “not conducting all of likely that the facilities were never actually given the the planned outreach activities” was lack of fund- authority to use the funds budgeted or the budget ing. Other factors included fuel problems and staff allocation was never received. Budgeting is actually constraints. done at the provincial or central levels and individ- ual facilities may have no control on the budgeted Cost of Running a Health Facility amount. Level 5 and 6 health facilities spent, on average, 92 percent of their budgets, and the level Excluding capital expenses, the average cost of facil- 7 health facility overspent its budget by about one- ity varied from K 1.6 million for a level 3 or 4 church- fifth. It should be noted that some of the health facil- run facility to K 57.7 million for the level 7 facility. ity expenses are covered by user fees, which may not The range of services provided by health facilities be accounted for when the budgets are set. Therefore, clearly indicate that even the level 3 and 4 facilities the facility level expenditure as percent of budget are quite complex production units, and, therefore, it would have been even lower if the facilities did not is difficult to collect comprehensive information on collect any user fees. production and costs. Table O.7 reports the average User fees are still charged by some lower-level annual cost of running a health facility by level. facilities, despite the policy of free primary health Salaries, operations, and utilities, rent, and main- care and subsidized secondary care. Table O.8 tenance make up a large component of health facil- reports the percentage of health facilities charging ity expenses (figure O.8). Level 3 and 4 church-run user fees; the average consultation, outpatient, and Table O.6  Average Annual Number of Outreach Activities by Health Facility Level Level 3 and Level 3 and Outreach patrols organized 4 public 4 church Level 5 and 6 Level 7 All facilities Percent of facilities organizing outreach patrols 82.8 92.0 50.0 0.0 76.7 Average number of outreach patrols planned for 2014 18.2 51.2 17.6 0.0 29.2 Average number of outreach patrols actually conducted  4.9 38.7 14.1 0.0 18.9 Percent of planned patrols carried out in 2014 26.9 75.7 79.8 — 64.5 Overview 9 Table O.7  Average Annual Cost to Run a Health Facility Figure O.8  Breakdown of Health Facility Costs by Health PNG K millions Facility Level Level 3 Level 3 100.0 and and Level 5 7.4 4 public 4 church and 6 Level 7 90.0 4.3 Average total 3.2 2.3 16.5 88.5 80.0 30.4 26.8 11.2 34.8 cost Average total 3.2 1.6 15.8 57.7 70.0 10.8 cost (excl. 60.0 19.4 3.5 capital) Note: PNG K = Papua New Guinea kina. 50.0 18.2 4.4 40.0 30.0 63.8 59.7 registration fee; and the estimated yearly income 20.0 44.4 39.9 generated through the user fees. As expected, level 5 to 7 health facilities collected significantly higher 10.0 user fees than level 3 and 4 facilities. In general, 0.0 however, collected user fees were very low com- Levels 3-4 Public Levels 3-4 Church Levels 5-6 Public Level 7 pared with the cost of running a health facility. Even Salary Operational Drugs for level 5 to 7 facilities, total user fees collected (in Benefits Capital Other expenses PNG kina)were only K 183,169. User fees are mostly Utility, rent & maintenance Training retained by health facilities and funneled back into the health facility either to meet shortfalls or to cover unplanned expenses. Health facilities noted that the user fees generally went to improving infra- Input-Output and Cost-Output Ratios structure; buying drugs or other medical supplies A number of other ratios were used to estimate rela- and equipment; paying for casual labor, adminis- tive efficiency of health facilities. These ratios are tration, and other office expenses; and purchasing either input-output based or cost-output based. The food for patients. number of beds per doctor at level 5 to 7 facilities was about six to seven, and for level 3 and 4 facilities the ratios were much higher because fewer doctors are Efficiency Measures employed at these levels. The number of personnel There are a number of efficiency measures estimated per bed was the highest in level 6 facilities, followed by in this report. The efficiency measures are either some simple ratios based on cost and production or a deri- vation of efficiency scores using econometric estima- Table O.8  Average User Fees by Health Facility Level tion of production and cost functions. Level 3 and Level 3 and 4 public 4 church Level 5 to 7 The Bed Occupancy Rate Percent charging user fees 28.6% 60% 79% In general, upper-level facilities (level 5 to 7) are Average outpatient fee K 1.60 K 2.70 K 6.00 able to use hospital inputs more efficiently than Yearly user fee collection K 1,690 K 19,084 K 183,169 other level facilities. A health facility is likely to be more “efficient” if the bed occupancy rate is higher, Table O.9  Average Bed Occupancy Rates by Method simply because hospital bed is one of the most Level 3 and Level 3 and expensive inputs in the production of health ser- 4 public 4 church Level 5 and 6 Level 7 vices. Independent of various methods, level 3 and Method 1 39.9 37.6 70.6 81.3 4 bed occupancy rates are quite low, both for public Method 2A 22.7 27.7 42.5 80.0 Method 2B 20.6 24.9 41.7 54.2 and church health facilities (table O.9). The ques- Note: Occupancy rates can be determined using two methods. The first method uses the number of tion is how to better identify the needed number of occupied beds on any given day and the second method estimates the number of days of each stay beds in level 3 and 4 facilities to ensure and sustain within a given year. In this report, the second method produced two efficiency measures, 2A and 2B, because two data options were available. Method 2A estimated the yearly admission days from access to inpatient services at lower-level health 1 month of administratively reported data and 2B used 4 months of data on hospital admission days facilities. collected from the health facility by the survey team. 10 Service Delivery by Health Facilities in Papua New Guinea the level 7 facility. The lowest personnel-to-bed ratio “efficient” health facilities. The figure indicates that was found in church-run facilities. The total cost per 6.9 percent (n = 2) of level 3 and 4 public health facili- bed was the lowest for church-run level 3 and 4 health ties are Lasso efficient, in comparison to 16 percent facilities. The cost per output was also the lowest for (n = 4) of level 3 and 4 church health facilities. church-run facilities and the highest for level 5 facili- ties. The cost per clinical personnel was the lowest for Policy Recommendations level 6 facilities followed by church-run facilities. Considering the cost per unit of output, the most The policy recommendations from the analysis are efficient facility category appears to be church-run grouped into four major areas. facilities. As expected, the highest cost facilities were level 5 to 7 facilities. Since these facilities provide a Policy Recommendations to Improve the Key wider range of services and cater to more severely ill Enablers for Service Delivery patients, it is not surprising to see relatively higher cost per output in these facilities, if case-mix is not A well-functioning health facility should be well taken into account. It was not possible to correct maintained, and adequate resources should be allo- output values because of the severity of medical con- cated for facility maintenance, especially for lower- ditions of patients. Therefore, this ratio may not rep- level facilities. Infrastructural conditions of the resent relative efficiency of facilities across different surveyed facilities indicate that about 30 to 40 percent levels but can be used as a proxy for efficiency among of level 3 to 6 facilities require major repairs. More health facilities within a specific level. than 60 percent of level 3 and 4 facilities did not have A greater fraction of level 5 to 7 health facili- an adequate number of toilets (for level 5 and 6, the ties was found to be efficient; by the Lasso measure percentage was 40 percent). Therefore, decision mak- about 63 percent of facilities surveyed were efficient. ers working in the PNG health sector need to deal Health facilities with fewer beds (between 50 and 100) with two aspects of infrastructural issues: infrastruc- appear to be less efficient, implying that efficiency ture maintenance and health facility design. improves with bed size of the facilities. The Lasso Getting the ‘right amount’ to the ‘right place’ and model incorporates the bed occupancy rate, bed turn- ‘on time’ needs to be a guiding priority. Budgets and over rate, and length of stay into a hospital efficiency funding streams need to be appropriately aligned to measure. Bed turnover is plotted against occupancy, ensure the right amount gets to the right level. Getting and average values of bed occupancy and turnover the money right is not only a matter of providing the are used to define four quadrants of optimal or sub­ right amount, it’s also a matter of getting the money to optimal combinations of bed turnover and occupancy. the right place and to the people who ultimately need Quadrant 3, labeled Zone 3 in figure O.9, contains all to spend it. The World Bank report “Below the Glass Floor: the analytical review of expenditure by provin- cial administrations on front line rural health” (2013) Figure O.9  Plot of Health Facilities by Bed Occupancy and “Financing the frontline: an analytical review of and Turnover Rates provincial administrations in Papua New Guinea’s rural health expenditure 2006–2012 (2015)” provides detailed analyses and policy recommendations. Stable water and electricity supply needs to be addressed for all facilities. Although most of the level 100 5 to 7 facilities were connected with an electric supply Bed turnover line, most of the level 3 and 4 facilities surveyed were Zone 3 not. About 15 to 20 percent of level 3 and 4 facilities 50 Zone 2 did not have water for health care providers to use in treating patients. PNG health system should ensure Zone 1 that all health facilities have access to a reliable source Zone 4 of water. 0 The availability of essential drugs at all health 0 20 40 60 80 100 facility levels needs to be improved. Using the list Occupancy rate of essential and commonly prescribed drugs in PNG, Overview 11 the survey found that drug availability on a continu- increase the supply of nurses and other allied health ous basis is poor at all health facility levels, including workers. Over the years, the supply of these personnel the level 7 facility. The nonavailability of drugs affects has consistently declined, implying that the human the quality of health services as well as satisfaction resource situation will continue to worsen, especially and motivation of providers. Therefore, nonavail- at level 3 and 4 facilities. The fact that facilities con- ability of drugs adversely effects both patients and sistently have unfilled positions, even for semiskilled health care providers. Drugs and medical supplies are support staff, indicates a systemwide problem of ordered bimonthly from the national drug and medi- personnel management. cal supplies’ procurement body, the Area Medical In-service training on priority health care ser- Store, which is supplemented by a “push system,” vices needs to be strengthened and expanded for standardized drug and medical supply packs sent to all health cadres. The evaluation of health care pro- all operational health facilities. Depending on need, viders’ knowledge about different types of priority direct requisition and purchase by health facilities health care services showed an interesting pattern. also takes place. An overhaul of the drug and medical The survey asked a number of knowledge questions supply system is required by defining the standard- related to child health care, maternity care, and ized supply kits by regions or provinces rather than STIs. Maternal care-related knowledge was low defining a national standardized kit. This system may for all provider types, including doctors. On aver- also discourage procurement from local distributors, age, the correct responses were only slightly above which often charge significantly higher prices. 50 percent. Similar results were found for child The drug monitoring and accounting system care services. However, the knowledge indexes were needs to be strengthened. In the survey, facilities found to be relatively high for STIs, in between 80 reported the number of drug and medical supply and 90 percent. Given that these are priority health boxes received in the previous one year as well as services, the knowledge indexes of health provider the drugs procured outside of the push system. The types imply that any of these personnel will be amount of drugs obtained or procured, as reported, equally effective (or ineffective) in the provision of was significantly lower than expected. In most cases, the basic priority services. The in-service trainings drug costs were less than 3 percent of the total facil- should be strengthened on these specific areas, and ity operation cost. It is likely that the facilities do with quality training, nurses or CHWs should be not keep detailed information on drugs obtained or able to provide some of these services at the same procured. The system of monitoring drug supplies effectiveness level as doctors. This is not to imply needs strengthening. Moreover, the push system of that policy makers should not consider appointing drug supply may create a significant misallocation of doctors at level 3 and 4 facilities; however, the rec- scarce drugs among health facilities. Simple inventory ognition of the Master of Medicine in a rural health management should be implemented in all health program would also improve the rural placement facilities to track drugs received through the push sys- of doctors. It is important to have doctors in refer- tem, procurement, and the utilization of drugs. ral facilities among level 3 and 4 facilities, but given The number of human resources in health facili- the overall shortages of doctors, it will be appropri- ties needs to be improved to match budget targets. ate to consider appointing other health personnel Human resource availability at health facilities is to provide maternal, child health, and STI-related lower than the number budgeted for personnel. At services. level 3 and 4 facilities, about 40 percent of budgeted Institute the user fee waiver policy for patients doctor positions were not filled. The percent of posi- at lower-level health facilities. Users of level 3 and 4 tions not filled was also high for nurses and CHWs facilities are predominantly poor, as indicated by the (about 25 percent). Even at higher levels (level 5 to 7 client surveys from the health facilities. The facilities facilities), more than 20 percent of doctor and nurse in the survey reported that they collect only a small positions were vacant. It appears unlikely that PNG amount of money in user fees. User fees collected, as will be able to meet the need for doctors at level 3 reported by facilities, represent less than 2 percent of and 4 facilities within a reasonable period of time facility expenditure. Level 3 and 4 church facilities and policy makers should carefully consider how to collected about 2 percent of facility expenses in user ensure access to care in the absence of trained doc- fees. Level 3 and 4 public facilities collected less than tors. Policy makers will also have to decide how to 0.3 percent in user fees. It is possible that facilities 12 Service Delivery by Health Facilities in Papua New Guinea underreport user fee collection. If the user fees are of obstetric first aid and basic EmOC at level 3 and 4 that low, especially for level 3 and 4 public facilities, it facilities is a major concern. The health sector of PNG may not be worth collecting the fees because they do should carefully evaluate the gaps in EmOC at level 3 not represent a substantial stream of operating rev- to 7 facilities so that the system can be strengthened. enue for the facility, but may represent a substantial The PNG health system should better integrate barrier to care for the poor population. outreach patrol planning into the health sector planning process to consistently reach segments of Policy Recommendations on Key the population in remote locations. The majority of Priority Service Areas the PNG population live in rural areas, many of which are not easily accessible. Making health care available The provision of antenatal care in all health facili- to the entire population will require a well-structured ties needs to be prioritized. Maternal and neona- outreach activity plan. Some facilities are able to tal health is one of the priority health areas for the reach remote population groups through outreach. health system of PNG because of the high fertility Facilities mentioned lack of funding, lack of fuel for rate and relatively high infant and maternal mortal- transportation, and lack of personnel as important ity. It appears that PNG has been successful in ensur- factors that limited the number of outreach activi- ing supply of basic instruments and supplies needed ties. By determining goals at the national and regional for maternity care. The items needed for antenatal or provincial levels, based on need or type, it will be care (ANC) should be available in all health facili- possible to estimate the resource requirements and ties in PNG, but the index of ANC items was 71 per- improve health care access to rural PNG. cent, 82 percent, and 81 percent for level 3 and 4 public facilities, level 3 and 4 church-run facilities, Policy Recommendations to Improve Efficiency and level 5 and 6 public facilities, respectively. Some very basic ANC instruments and supplies were miss- Health facilities should use existing resources more ing or not functional, including weighing scales and efficiently. In the survey, it was found that although tape measures. Inability of a health facility, even a about 30 percent of level 3 and 4 facilities had oper- level 3 health facility, to provide ANC is an impor- ating theaters, less than half of the facilities actually tant concern for any health care system. Health use the operation theaters for conducting surgeries policy makers will have to ensure the availability of on a regular basis. Nonuse of the facilities highlights ANC in all facilities. the need to ensure availability of complementary Child immunization services must be strength- resources such as skilled personnel and other instru- ened. All the required childhood vaccination sup- ments and supplies. Another similar issue is the avail- plies (vaccines) were not available in most of the ability of shelving for storing drugs. Most facilities facilities. The most critical shortage appears to be reported having adequate space for storing drugs, with the BCG vaccine (consistent with worldwide but some of the drug boxes were placed on the floor shortage of the vaccine). The child immunization because of lack of shelving. Policy makers should program is highly cost-effective and so the system ensure that all complementary inputs are supplied to should be able to improve its efficiency and effective- effectively use available resources. ness by ensuring the timely vaccination of children. From an allocative efficiency perspective, the The supply chain needs to be strengthened and the range of important medical care services in level 3 immunization program should develop an informa- and 4 facilities needs to be expanded. A number tion system to identify locations with shortages and of important medical care services are not available excess supplies. in most level 3 and 4 facilities, including laboratory All facilities at level 3 or above should be able and diagnostic services and pharmacy. Availability to provide obstetric first aid and basic emergency of such services at level 3 and 4 health facilities can obstetric care. Most facilities should also be able to improve the overall efficiency and quality of service provide comprehensive EmOC. The survey of facili- delivery. The range of health function tests at level 3 ties indicated that 20 percent of facilities at level 3 and and 4 facilities could be broadened to include simple 4 were not prepared to provide even obstetric first urine, stool, and blood glucose tests. aid, and 40 percent were not ready to provide basic The government needs to identify and resolve EmOC. Given the high fertility rate in PNG, the lack impediments to the provision of health care ser­ Overview 13 vices at level 3 and 4 facilities. Bed occupancy rates, A recommendation is that if FBB and DFF are to a measure of health facility utilization, are very low in be implemented, sufficiently skilled personnel are level 3 and 4 facilities, both public and church-run. employed at health facilities to ensure appropriate Low utilization rates are likely a reflection of non- management. availability of complementary inputs. Since level 3 and 4 facilities should provide more comprehensive Implications of Technical Assistance to NDOH health care services, policy makers should identify the resources in short supply. Without basic ame- A junior public financial management specialist nities, such as water and electricity, it is difficult to provides embedded technical assistance (TA) to the make health facilities more effective in the provision NDOH. The main goals of this TA are to improve of inpatient services. In addition to these basic ame- planning and prioritization of health expenditure, as nities, facilities also lack skilled human resources, well as strengthening health sector budget submis- instruments, and supplies. sions. In addition, the TA offers a good opportunity to The health information system needs to be follow-up on the recommendations of the analytical strengthened across all facility levels. Better data can work. There are four areas that have clear implications lead to better decision making and better outcomes. for public financial management TA at the NDOH: Many of the registries used for data collection by the health information system were not available at facil- • Costing of health services. The report offers a ity levels 3 to 6. Policy makers need to strengthen the detailed description of the outputs produced by information system and ensure that different reports, the different facility levels. Combined with costing forms, and registries are available in all facilities. Data data, this information can help assess the adequacy should be routinely used to generate evidence for bet- of facility budgets and increase knowledge about ter decision making. funding needs in the front lines. The government needs to improve the autonomy • Efficiency analysis. Given the health sector’s finan- of public health facilities to improve their effi- cial constraints, an important component of the ciency. After examining the various indicators of TA focuses on achieving efficiency gains in health health facility performance and service delivery, the expenditure. The report offers a set of efficiency results indicate that the church-run facilities perform indicators that can potentially help the NDOH better than other types of health facility categories advance in the discussions around a facility’s per- (level 3 and 4 public-run facilities and level 5 to formance benchmarking. 7 facilities). It is interesting that church-run facilities • Service delivery readiness. The report identifies key are not necessarily the better-endowed facilities with service delivery elements that need to be avail- regard to human resource availability or availability able for the effective provision of quality health of drugs and supplies. Church-run facilities are also services, including outreach activities. This infor- very active in trying to provide coverage to popula- mation can help the NDOH to develop a checklist tion groups living in remote areas. The administra- for supervision of a facility’s readiness. Readiness tive structure of church-run facilities and the degree gaps can be monitored and incorporated into the of autonomy they have may improve the efficiency planning process. and effectiveness of these facilities compared with • Health facilities’ autonomy. In the context of decen- other facility types. The National Department of tralization, the NDOH is exploring the poten- Health (NDOH) is exploring facility-based bud- tial for FBB and DFF. Understanding a facility’s geting (FBB) and direct facility funding (DFF) as desired level of autonomy, and identifying the instruments to potentially improve the autonomy areas where facilities would like to have more of and funding flow to the public health facilities. autonomy, will help steer this process. Chapter 1 INTRODUCTION AND HEALTH SYSTEM OF PAPUA NEW GUINEA Papua New Guinea (PNG) has seen steady improve- lowered economic growth. The short-term economic ments in health indicators over the past 25 years, outlook is not positive. A key factor dampening but the pace of improvements is not as robust as resource growth is the required fiscal consolidation expected. The under-five - mortality rate declined recently initiated by the government. from about 89 per 1,000 live births in 1990 to 57.3 Total health expenditure, as a proportion of GDP, in 2015. Life expectancy improved from less than has been stagnating between 4 and 5 percent in 40 years in 1960 to 63 years in 2014. The infant mor- recent years, despite steady growth. As figure 1.1 tality rate declined from 65 per 1,000 live births in illustrates, total health expenditure rose rapidly 1960 to 44.5 in 2015. Despite these positive trends, between 2000 and 2001, a 2.7 percent increase in one the lack of strong momentum prevented PNG from year. It peaked in 2004, at 8.4 percent, before falling meeting Millennium Development Goals. Further, by 4.4 percent between 2004 and 2007. In 2014, total PNG’s health outcomes do not compare favorably health expenditure (4.4 percent) was comparable to with other regional countries despite an increase the low-middle average of 4.5 percent and to other in gross domestic product (GDP) per capita from countries in the region: Fiji (4.5), Solomon Islands K 786 (US$655 at PPP5) in 2000, to K 6,114 (US$2,268 (5.1), Timor-Leste (1.5), and Vanuatu (5).6 It should at PPP) in 2014; a compound rate of growth of be noted here that lower percent of GDP allocated to about 3.8 percent over the 14-year period (World health does not necessarily imply absolute decline in Development Indicators [WDIs], http://databank. health expenditure per capita. A number of countries worldbank.org/). The PNG health sector is entering show declining or stagnant percent of GDP allocated a challenging period of constrained fiscal resources to health without showing any negative consequences as the government grapples with domestic revenue on health outcomes. Therefore, this particular indi- mobilization, achieving a balanced budget, and resto- cator of health system performance should be care- ration of macroeconomic stability. fully evaluated. Government plays a very important role in funding health care services. PNG spends about Challenges to PNG’s Fiscal Space US$94 per capita per year on health and the share for Health of public expenditure on health is about 81 percent PNG experienced relatively high economic growth (figure 1.2). Given the budget constraints that the from 2010 to 2014. However, the decline in com- PNG government is experiencing, it has become cru- modity prices and drought in 2015 and early 2016 cially important to ensure efficient and effective use have weakened the external and fiscal positions and of public sector resources so that the health system can continue to provide an adequate level of health 5 Purchasing power parity (PPP) data from: http://wdi.worldbank. org/table/4.16 . Note that 2014 PPP conversion factor value is the 6 average of the 2013 and 2015 values. WDIs. 15 16 Service Delivery by Health Facilities in Papua New Guinea Figure 1.1  Total Health Expenditure as a Proportion of GDP, 1996–2014 10.0% Health expenditure (% of GDP) 8.0% 6.0% 4.0% 2.0% 0.0% 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 Health expenditure, total (% of GDP) Health expenditure, public (% of GDP) Source: World Development Indicators database, October 2016. Figure 1.2  Health Financing Mix in PNG, 2014 in the region: Fiji (9 percent), Timor (31 percent), Solomons (56 percent), and Vanuatu (48 percent). PNG has relatively low out-of-pocket (OOP) 9% expenditure on health by international standards. 10% Given comparatively low access rates, there may be Public a significant degree of foregone care, which reduces OOP OOP spending, and low OOP does not necessar- Other private ily mean financial protection. Health care costs are cited as one of the main reasons the poor do not visit 81% health facilities in case of illness (Irava et al. 2015). Only 10 percent of total health spending is attributed to OOP payments and is equivalent to only 0.5 per- cent of total household expenditure. OOP expendi- Source: World Health Organization, October 2016. Note: OOP = out-of-pocket. ture was less than 30 percent of total consumption in any household in PNG, and only 0.02 percent of households spent more than 20 percent. Less than 1 percent of households spent more than 10 percent of care services without adversely affecting access, their budgets on OOP payments. However, an esti- quality, and timeliness of health service delivery. mated 0.3 percent of households were impoverished External resources on health,7 as a percentage as a result of OOP health spending, and 3 percent of of the total health expenditure, continue to play the poor were pushed further into poverty. a signi­ficant role in PNG’s health financing. In The government has renewed its intention to 2014, external resources accounted for 21 percent reduce OOP spending by introducing a fee-free (figure 1.3) and have stabilized around 20 percent basic and subsidized specialized health care policy, in recent years. This share is high when compared which requires all facilities to stop charging user with the low-middle income average of 3.3 percent. fees for primary care. K 20 million has been allo- However, the percentage is lower than for many others cated to compensate facilities for the loss of user fees in addition to the already existent health function grant, which funds facility operations. However, 7 External resources on health are funds or services in kind the persistent slow release of funding by central that are provided by entities not part of the country in ques- agencies and delays in channeling funds at the pro- tion. The resources may come from international organizations, vincial level means that facilities do not receive the other countries through bilateral arrangements, or foreign non­ governmental organizations. These resources are part of total financial support required to function without user health expenditure (WDI). fees. This problem has prevented the successful Introduction and Health System of Papua New Guinea 17 Figure 1.3  External Resources on Health as a Percent of Total Health Expenditure, 1995–2014 35 30 External resources on health 25 as % of THE 20 15 10 5 0 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 External resources on health as % of Total Health Expenditure (THE) Source: World Development Indicators database, October 2016. elimination of user fees at primary public health departments, hospitals, and district health services facilities to date. under one management board. However, this leg- islation is not compulsory. By 2016, 10 provinces fully established PHAs, and an additional five The Basics of PNG’s Health System provinces plan to establish a PHA by the end of 2017 (National Department of Health, 2017). Initial PNG has a government-funded health system implementation of the PHA was slow in the three pilot throughout much of the country. It is supple- provinces, because many staffing, financial manage- mented by government-subsidized health services ment, and institutional hurdles had to be overcome. provided by various Christian missions. At the time Since then, implementation has accelerated and an of independence, a centralized National Department independent review found early success in the finan- of Health (NDOH) managed the whole health cial management and service delivery support that system. Since independence, the government has PHAs are providing. However, the model suffers made successive attempts to decentralize the pro­ from an inconsistent legal and regulatory framework. vision of services to provincial and district govern- A key example is that provincial governments should ments, including the introduction of two significant immediately pass the funding for rural health service Organic (Constitutional) Laws (1977 and 1995). This delivery on to PHAs after it has been transferred from legislation attempts to give provinces and local-level the national level. However, this is only a voluntary governments increased control over health ser- arrangement, which allows some provinces to retain vice delivery and resources. The 1995 Organic Law a share of funding for their own activities. Future devolved primary health care services to the pro­ strengthening of the PHA model requires commit- vincial level. However, hospitals (including their ments to adequate levels of resources, timely release budgets, human resources, and payroll) are managed of funds, recruitment of skilled staff, improved com- by an autonomous board, which is independent of munication, coordination and effective reporting the NDOH. This structure contributes to a diffuse among all stakeholders, and sufficient levels of sup- and fragmented health delivery system. port from national departments. If such issues are To address health system fragmentation, the resolved and capacity constraints are removed, the 2007 Provincial Health Authority Act enabled the PHAs may prove to be a successful model that uni- provincial government to establish Provincial fies all service delivery functions within the province Health Authorities (PHAs) to be responsible for under one management structure. Note that the sur- both primary and secondary health care in the vey was carried out in 2015 and the survey collected provinces. This legislation is supposed to stream- information on calendar year 2014 when only a few line the provision of health services at the provin- provinces had fully implemented the devolution cial level and bring together the provincial health policy. 18 Service Delivery by Health Facilities in Papua New Guinea Church Health Services is the key partner in support, including leadership, teaching training, and delivering health services in PNG.8 Both Christian health information; (4) health workforce productiv- Health Services (CHS) and Catholic Church Services ity and capability; (v) drugs, equipment, and medical have negotiated a separate funding with NDOH. supplies; and (5) health infrastructure, that is, build- Overall, it is estimated that churches provide 47 per- ings that meet minimum standards.10 cent of primary health services, and a significant Health facilities are categorized according to share of secondary services, particularly in rural the services they provide and staff ceilings and are areas. The role of the CHS varies significantly across detailed in the NHSS. As mentioned above, the levels the country. In rural areas, CHS provide up to 60 per- of health facilities in PNG are defined by numbers cent of primary health services, and in some prov- 1 to 7, where levels 1 to 4 provide basic primary health inces such as West Sepik (Sandaun), CHS account for care, outpatient services, and limited clinical support 80 percent of all health services provided.9 services, such as laboratory, pharmacy, and radiog- The National Health Administration Act 1997 raphy services (only provided by a few level 4 health required the NDOH to develop a set of national facilities). Level 1 facilities, community aid posts, health standards to operationalize the National are managed by one health worker, either a com- Health Plan (NHP). In 2011, the Medical Standards munity health worker (CHW) or a nurse. There are Division, through extensive consultation with clini- no inpatient services at this level and, in many cases, cians, public health, and health service managers, the scope of health activities is very limited, deal- developed the National Health Service Standards ing with basic infectious diseases. Anything that is (NHSS) for PNG. These standards contribute to the more complex or requires admission is referred to a government’s Vision 2050 and were endorsed by the level 3 health facility. A level 2 health facility, health National Health Board at the March 2011 meeting. subcenter, employs about three staff members, usu- The NHSS outlined a structure for health service pro- ally two nurses and one CHW. The health facility only vision in PNG for the following 10 years. The stan- admits pregnant women for delivery of babies and dards are seen as (a) an important tool for national, provides prenatal, postnatal, and newborn services. If provincial, and district planning and delivery of the pregnant women develop complications, staff at health services; and (b) a blueprint for providing level 2 facilities are supposed to refer the women to safe, quality care consistent with the NHP objec- the closest level 5 health facility. Level 3 health facili- tives. The minimum standards envisage a hierarchical ties, or health centers, also provide basic inpatient structure for health services across the nation includ- services for deliveries and minor ailments requiring ing level 1 village aid posts, level 2 community health observation. Level 4 health facilities, district and rural posts, level 3 health centers and urban clinics, level 4 hospitals, provide general admissions and, depending district hospitals, level 5 provincial general hospitals, on the employment of a doctor, may provide surgi- level 6 regional referral hospitals, and level 7 national cal services if anesthetic officers and scrub nurses are referral hospitals offering secondary and complex employed and present in the facility. The range of sur- tertiary-level clinical services. The standards specify gical services is limited and includes predominantly the following: (1) core clinical and primary health cases such as appendectomies or cesarean sections. care services, such as medical, sexually transmitted Complex cases, such as hernias and open reductions infections, and maternal and child health; (2) support and internal fixations, are carried out at level 5 to services, such as diagnostics (medical imaging and 7 health facilities (or hospitals). Level 5 to 7 health pathology, pharmacy, drugs, and medical supplies), facilities provide secondary and tertiary health ser- infection control, and public health; (3) management vices. The types of services offered on a regular basis are determined by the skill and knowledge of the staff. Therefore, a health facility employing a number of 8 Christian Health Services only administer funding to various doctors [medical officers [MOs]), consultants (spe- church organisations organizations for the delivery of health cialist medical officers [SMOs]), or senior consultants services and liaise centrally with NDOH. They are not techni- cally service providers, but this report will refer to CHS as such throughout the report. 9 10 Christian Health Services Technical Assistance Mission Report, National Health Service Standards for Papua New Guinea 2013. Note that the Catholic Church is not part of CHS and has 2011–2020. A blueprint for providing safe, quality health services negotiated a separate MOU Memorandum of Understanding for as required by the National Health Plan 2011–2020 to transform funding. Catholics are the largest church provider. our health system, Government of Papua New Guinea, June 2011. Introduction and Health System of Papua New Guinea 19 (senior specialist medical officers [SSMOs]) have the (2-year training to receive a diploma in nursing). skill and knowledge to treat a wider range of (com- Nurses can further train in diploma courses to become plex) medical conditions compared with a health aesthetic officers, midwives, or specialist pediatric facility that employs one health extension officer nurses, for example. CHWs train for 1 or 2 years and (HEO). Further, level 5 to 7 health facilities provide receive a certificate in community health work. clinical support services, such as pharmacy services According to the health service delivery profile (services provided by a pharmacist, that is, a person report (WHO and NDOH, 2012), PNG has 22 pro- holding a Bachelor of Pharmacy degree), laboratory vincial hospitals (the number of functional hospi- services (provided by a pathologist, that is, a medi- tals in 2014 was 19), 14 district and rural hospitals, cal officer trained in pathology), and radiology (pro- 201 health centers, 428 health subcenters, and vided by medical officers trained in radiology), and 2,672 functional or open aid posts (as of 2008). most of these specialized providers are employed by Another report used information from PNG official level 6 and 7 health facilities. The level 7 health facil- documents; the number of hospitals was reported ity, the Port Moresby General Hospital, is the most as 19 with 719 rural health centers and 2,672 health advanced health care facility of PNG and provides posts (Janovsky and Travis, 2007). There appears comprehensive health care services. to be significant confusion even for basic health In PNG, health care facilities organize outreach sector information such as number of functional patrols to improve coverage of health facilities. health facilities in the country. Based on the reported Outreach is categorized either as integrated rural numbers of health facilities, population coverage patrols or community outreach. Integrated rural varies from 5,000 to 20,000 depending on the region integrated patrols are implemented by level 5 to 7 or area. health facilities and require a team of medical officers Population access to health services appears (usually one per specialty) to travel to level 2 and 3 in decline. Health posts offer greater access to ser- sub health centers and health centers to provide vices, but unfortunately the number of open or consultation to patients. Integrated rural patrols are functional aid posts appears to be in decline. The supposed to be organized once a quarter; however, in service delivery profile report also mentioned that a reality, this does not happen as expected because of number of aid posts were closed because of lack of logistics and time constraints. Community outreach, funding. A sector performance assessment report of on the contrary, is handled by level 3 and 4 health PNG (National Department of Health, 2016) also facilities. Most of these outreach patrols are day- reported further decline in aid posts and the amount clinics, usually maternal-child health clinics. They of outreach conducted in recent years. The reported are run in the catchment area of the health facili- decline represents further reductions in population ties and most patrols immunize infants and provide access to health care services. prenatal services to pregnant women. These com- Use of various health care services has remained munity outreach clinics are usually run by two or much lower than the optimal level. Such low use three nursing officers and CHWs. is not caused by improved health status, but rather In terms of educational and training of health a sobering picture of low incentives from demand care providers in PNG, a medical officer or registrar side to seek care and inadequate performance from needs to complete 5 years of university education the supply side. According to the World Health (bachelor of medicine and surgery) and 2 years Organization Health Profile of PNG, DTP3 immu- of residency to become a fully licensed doctor. To nization coverage by 1 year of age was about 70% in become an SMO, it takes another 5 years (master of 2015. Although there are no recent data on maternal medicine). An SMO may get promoted to SSMO service use, less than a third of pregnant women after significant years of experience, usually around probably obtained at least four antenatal visits dur- 15 years. An HEO studies for 4 years of university ing pregnancy. Contraceptive prevalence rate has also (bachelor of rural health) and then has 2 years of remained relatively low. Human resource availability residency before becoming a licensed HEO. An HEO in health remained quite low, which restricts the can further specialize by receiving diploma train- use of medical care services. Number of doctors per ings, which are the same specialized diplomas for 10,000 population is probably around 0.5, imply- nurses. Nursing officers study for 3 years (bachelor ing that if per capita doctor’s visit is one per year, a of nursing); some nurses have a diploma in nursing doctor will have to see about 80 patients per day on 20 Service Delivery by Health Facilities in Papua New Guinea the average. PNG has produced a significant number plan focuses on strengthening primary care for all, of other clinicians, such as nurses and community it also explicitly indicates the need for improved ser- health workers, to improve access to medical care. vice delivery for the rural majority and the urban The number of nurses and CHWs in PNG is prob- disadvantaged. Key result areas (KRAs) include the ably around 5.3 per 10,000 population, which should need for strengthening not only the provision of pri- lower patient load on doctors. mary health care services but also the development of well-integrated primary and secondary levels. To support and realize the goals and objectives of The National Health Plan and the NHP 2011–2020, primary and secondary health Role of Level 5 to 7 Facilities care facilities must be ready to provide services The National Health Plan 2011–2020 (Government to deal with priority health outcomes and health of Papua New Guinea, June 2010) defined PNGs service delivery related KRAs. The plan also defines health system goal and its specific objectives to specific components or dimensions of health sector achieve health targets. Figure 1.4, taken from the strengthening. Therefore, the hospital system must NHP 2011–2020, summarizes the health system be strengthened in a way that will help achieve the goal with specific objectives and strategies. The system-specific key results with a goal of improv- plan recognizes that PNG has not been successful in ing population health outcomes. Although public improving health status of the population and has facilities are allowed to generate revenue to cover the lagged behind comparable countries in basic health shortfall in operating budget, the facilities remain outcomes, such as the infant mortality rate and the subject to public sector control and management. maternal mortality rate (see also Howes 2014). The Therefore, although the hospital sector of PNG may plan is considered a part of a longer-term strategy and be considered dominated by not-for-profit facilities vision to improve health. Although the 2011–2020 (private sector being very small), the outcome of the upper-level primary and secondary facilities, in gen- Figure 1.4  PNG National Health Plan Framework eral, should not deviate significantly from the objec- tives and goals defined by the NHP 2011–2020. National Health Plan 2011–2020 Objectives of the Study and Report Outline Strengthened Primary Health Care for All and Improved Service Delivery for the Rural Majority and Urban Disadvantaged The purpose of this study is to investigate whether primary and secondary health care facilities (level 3 and 4 and level 5 to 7) in PNG have an adequate Address Priority Health Outcomes level of resources, both human resources and other material inputs, to deliver quality health care ser- vices and to determine whether the inputs are KRA 1 KRA 4 A Healthy and Prosperous Nation For all, both Now Improve Service Improve Child Survival combined in an efficient manner to produce health Delivery care services. KRA 5 Chapter 2 discusses the survey design and meth- and for Future Generations Improve Maternal Health ods. The report followed an input-output frame- Improve Service Delivery to Achieve KRA 6 Reduce the Burden of Communicable Diseases work to present the study results. Chapter 3 discusses KRA 2 Strengthen Partnerships input-related aspects of health facilities, including and Coordination KRA 7 Promote Healthy Lifestyles health facility infrastructure, instruments and medi- with Stakeholders cal supplies, clinical and patient education protocols, KRA 8 Improve our Preparedness for Disease Outbreaks service availability, and human resources. Chapter 4 and Emerging Population Health Issues analyzes management, access, equity, quality, and safety of health service delivery in PNG. Chapter 5 KRA 3 discusses the costs and outputs of PNG’s health Drug and Medical Leadership and Infrastructure Information Strengthen Governance Workforce Financing Supplies Health facilities. Chapter 6 analyzes the efficiency of health Health (ICT) Systems facilities using the resources to produce the outputs. Last, Chapter 7 concludes and provides policy rec- Source: Government of Papua New Guinea, June 2010. ommendations based on the findings. Chapter 2 SURVEY DESIGN Survey Instruments vious year to ensure that seasonality of service provision was captured when estimating annual The purpose of this study is to investigate whether production. the upper-level primary and secondary health care 3. Health care provider interview. The instrument facilities in PNG have an adequate level of resources, was directed toward registered clinicians, namely both human resources and other material inputs, medical officers, nursing officers, health extension to be able to deliver quality health care services and officers (HEOs), and community health workers to determine whether the inputs are combined in (CHWs); no resident medical officers and HEOs or an efficient manner to produce health care services. student nurses and CHWs were interviewed. The To achieve this objective, the study used the World survey collected basic demographic information, Health Organization (WHO) building blocks frame­ information on types of services provided, time work11 to conceptually identify important aspects or spent on service provision, and the clinicians’ dimensions of a health system and health facilities. ability to provide specific service types, as well These aspects are captured in the following four as their opinion on working conditions in the survey instruments: health facility. 4. Patient interviews. Two instruments were designed 1. Health facility assessment instrument. The instru­ for patient interviews: one for conducting out­ ment comprised of nine sections, each of which patient exit interviews and the other for inter­ was designed to collect information on one spe­ viewing admitted patients (inpatients). Both col­ cific operational aspect of the facility. In particu­ lected general patient information and patient lar, the instrument or questionnaire explored the satisfaction levels with the services received. The services offered by the health facilities as well as instruments also included questions on patient different inputs and resources available for deliv­ household asset ownership to understand the ering the services. economic and social status of the patients using 2. Health facility costing and output instrument. the facilities. The instrument collected information on health facility costs and outcomes. The facility respon­ dents were asked about the budget of the facili­ Sample and Sampling Design ties by line items as well as actual expenditures in the previous year. Output levels of the facilities To achieve the study objective, all level 5 to 7 facili­ were collected for 4 specific months in the pre­ ties operational in PNG at the time of the survey were selected, including the national referral hospital (Port Moresby General Hospital), the three regional hospitals, and 15 provincial hospitals. Sixty level 3 11 Refer to Annex A for details of WHO building block framework. to 4 facilities were selected from 11 provinces and 21 22 Service Delivery by Health Facilities in Papua New Guinea Map 2.1  Map of PNG with Districts Visited (green) and Location of Surveyed Facilities (red and blue squares) Note: Red squares represent the level 5 to 7 facilities surveyed, while the blue squares represent the level 3 to 4 facilities. stratified by public and church health facilities.12 squares represent the level 3 to 4 facilities. Green Church-run facilities play a very important role in color represents the districts visited by the survey providing services to rural and remote areas (Ascroft team to conduct the survey of facilities. Some of et al. 2011) . Therefore, equal representation of the health facilities selected for the survey were church-run health facilities was important. located in quite remote areas requiring careful Map 2.1 shows the location of the health facili­ evaluation of travel time and cost. ties surveyed in PNG. Red squares represent The sampling followed a two-step selection pro- the level 5 to 7 facilities surveyed, while the blue cess to capture an equal amount of publicly and church-run facilities. In step one, between six (New 12 The selection of level 3 and 4 facilities was carried out from Guinea Islands) and eight (Southern, Highlands, Chimbu, Southern Highlands Province (SHP), Western Highlands and Momase) districts across the 11 provinces were Province (WHP), Milne Bay, Oro or Northern Province, Western, randomly selected per region. 13 In step two, the East Sepik, Madang, the Autonomous Region of Bougainville (ARB), Manus, and New Ireland. In an earlier study, Howes et al. largest government- and church-administered health analyzed the expenditures in rural health services in PNG between facilities (one of each type based on outpatient num­ 2002 and 2012. This report serves to complement Howes et al. bers per annum as reported by the National Health (Howes 2014) by surveying level 3 and 4 health facilities in provinces not covered by the earlier study. Therefore, expenditure patterns Information System) were purposely selected per reported in the earlier report can be reviewed to supplement the district. This method resulted in 12 level 3 and 4 information reported in this study. Also note that Hela Province health facilities from the New Guinea Islands region, was sampled as part of Southern Highlands because there was no segregation made between Hela and Southern Highlands two each from all six districts in the participating with regard to health facility listing and numbers. Note that Tari 13 General Hospital, a level 5 health facility, was surveyed. Similarly, The total number of districts per region across the 11 provinces Jiwaka Province was sampled with Western Highlands Province. is as follows: Southern (Western, Milne Bay, Oro) = 9; Highlands The Provincial Health Authority (PHA) of Jiwaka was established (Southern Highlands, Western Highlands, Simbu) = 15; Momase in April of 2017 and at the time of the survey all health facilities in (Madang, East Sepik) = 12; and Islands (Manus, New Ireland, the province were being managed by the Western Highlands PHA. Autonomous Region of Bougainville) = 6. Survey Design 23 provinces, and 16 level 3 and 4 health facilities from inpatient interviews were planned at each level 5 to 7 eight districts in all or most14 of the participating facility and five interviews at each level 3 and 4 provinces in each of the Southern, Highlands, and facility. The survey team made an attempt to inter­ Momase regions. Although, theoretically, this sam­ view at least one participant from each of the core pling procedure allowed selection of 30 publicly run hospital wards (e.g., medical, pediatric, surgical, and and 30 church-run facilities from the selected dis­ obstetrics and gynecology), if possible. Participant tricts, in reality not all these facilities were found to selection within a ward was based on simple random be operational. In some of the selected areas, there selection using total number of patients per ward were no other functional level 3 and 4 facilities as the sampling frame (e.g., in a hospital with five to be included in the survey, resulting in less than wards, two participants were systematically selected 60 facilities in the final sample. from each). The goal was set to interview 400 health care For outpatient interviews, the target was to providers from the surveyed facilities, includ- interview 400 users of outpatient services from ing 250 interviews in level 5 to 7 facilities and across all participating health facilities. The tar­ 150 interviews in level 3 and 4 facilities. Provider get was to interview 10 outpatients from each level interviews were completed with health facility staff 5 to 7 facility and five outpatients at each level 3 to that provided clinical services. The targeted numbers 4 facility. In facilities where two or more outpatient of provider interviews per cadre per facility were as departments were present, participants were sought follows: three medical officers (or less if fewer than from the two most highly used outpatient services. three employed), one HEO, two nursing officers, and In smaller facilities, which did not have multiple out­ two CHWs. These targets were based on the expecta­ patient wards or sections, all patients were recruited tion that medical officers might not be available at from the outpatient service provision area or location. most of the level 3 and 4 health facilities sampled The survey teams aimed to achieve an even split by (therefore, oversampling was sought at level 5 to 7 sex (male or female) and age (<5 years or 5+ years). facilities). The intent of participant selection within The interviewers were instructed to let the pros­ each cadre was to achieve the following number of pective interview candidate know immediately after interviews: 50 medical officers, 50 HEOs, 150 nursing arrival to the facility that she or he would be inter­ officers, and 150 CHWs across a range of specialities viewed. At that point, the interviewer explained the and clinical service delivery areas. The procedure fol­ purpose of the interview and solicited the patient’s lowed to select health care providers for interview consent to participate. The interviews were conducted was to approach a health care professional and ask only when written informed consent was obtained. her or his name and position. If the person fell into The first patient to exit the outpatient clinic was one of four categories (medical officer, nursing officer, usually the first patient interviewed. HEO, or CHW), the interviewer then ascertained The field survey started on February 9, 2015, that the person was not a trainee and provided services and ended on July 31, 2015. In total, 73 health facil­ to patients. ity survey assessments, 72 health facility costing and A total of 245 inpatient interviews were con- output instruments, 385 inpatient interviews, and ducted across all facilities in the survey.15 Ten 514 outpatient exit interviews were completed. 14 Using this selection strategy, it was possible to exclude one of Limitations the selected provinces in the Highlands region. 15 The procedure followed by the interviewers to select the A number of limitations of the study should be men­ inpatients was to enter the ward through the main entrance and tioned here. First, it is not possible to indicate over­ attempt to interview the first patient on the right-hand side of all geographic access to health care services and the the room. The only condition of eligibility was that the patient must have spent the previous night in the health facility. The inter­ effect of service provision on the population because viewer then assessed whether the patient or the caregiver of the the study did not collect information from lower- patient was able to answer the questions. If the patient appeared level health facilities (levels 1-2). A community-level too sick to respond to the questions, the interviewer moved to the household survey was not conducted and therefore, next patient adjacent to the first. Once the ability of patients to respond to questions was verified, the required number of patients it was not possible to indicate population coverage of was selected from among the eligible patients. health facilities. 24 Service Delivery by Health Facilities in Papua New Guinea Second, the survey is not a nationally represen­ Fourth, the wealth scores were calculated by using tative survey of level 3 and 4 facilities. The survey Principal Component Analysis (PCA) considering design selected larger, functional level 3 and 4 in the reported ownership of assets by patients inter­ target provinces. viewed at exit for users of outpatient services and in Third, severity of medical conditions of patients is the facility for inpatients. This may bias the measure­ not known. It is assumed that the average severity of ment of wealth scores as the patients self-select to medical conditions increases with the level of health visit a facility and the patient population may be a facility and the cost of providing services should also biased sample of underlying distribution of assets in increase with severity. the community. Chapter 3 INPUT-RELATED ASPECTS OF HEALTH FACILITIES This chapter describes the inputs used by health the characteristics of the health facility buildings, facilities in the process of producing health care construction materials used, number of toilets avail- services in PNG. In any production process, inputs able in the facilities, and the repair needs. Only are used to produce outputs. Health care facilities use 41 percent of level 3 and 4 public facilities and 56 percent facility infrastructure (the hospital building and other of church-run facilities had signboards with the name infrastructural resources); human resources, such as of the facilities on the main buildings. Presence of medical officers and nursing officers; health com- a signboard is not an input per se, but its absence modities, for example, drugs and medical supplies; could imply either no budgetary authority given to and the knowledge and skills of health care providers. the facility managers to erect a signboard or lack of Inputs are combined during the production process client orientation by the facilities. The majority of to produce the outputs of each facility. health facilities required a major repair. Although The results are presented by categorizing sur- the information is subjective, it indicates the degree veyed health facilities by administration type (pub- of repair needed in most facilities. Only about one- lic and church-run) and level. The health facilities third of facilities (both church-run and public health surveyed were grouped per level; level 3 and 4 health facilities) reported an adequate number of toilets. This facilities were grouped together, levels 5 and 6, and, information again indicates the urgency of improv- last, level 7. Although there is only one level 7 facil- ing the infrastructure for better quality of care. Most ity, its sheer size may prompt some distortions or facilities reported that the facility building is accessible bias in the facility data and therefore it is analyzed to patients with disabilities. separately. The level 3 and 4 group was further sub- Most level 3 and 4 health facilities require divided into public facilities and church-run facil- better connectivity to electricity and clean water ities. Since all the facilities at levels 5 to 7 were public, supply. Table 3.2 illustrates service-related infrastruc- categorization based on administration is not rel- tural variables across facility types and levels. The average evant for the upper-level facilities. Twenty-nine number of inpatient beds varied from 21.4 for level 3 public facilities and 25 church facilities at levels 3 and 4 public health facilities to 886 for the level 7 and 4 were surveyed. At the levels 5 and 6, all 18 facil- facility. Most facilities reported having an ambulance. ities in the survey were public facilities. However, more than half of the surveyed health facili- ties at levels 3 and 4 reported very poor connection to electricity: 35 percent for public health facilities and Health Facility Infrastructure Inputs 28 percent for church-run health facilities. In addition, Facility management at all facility levels (except over half of the level 3 and 4 health facilities surveyed level 7) reported the need for major building required repairs to their water supply systems. The repairs and emphasized the lack of adequate toi- overall infrastructure condition is much better for level lets, stable supply of electricity, and consistent water 5 and level 6 facilities; the level 7 health facility required supply. Table 3.1 reports basic infrastructural inputs: no infrastructure, electrical, or plumbing repairs. 25 Table 3.1  Basic Infrastructural Variables by Facility Category Level 3 and Level 3 and Level 5 Infrastructural variables 4 public 4 church and 6 Level 7 Percent of facilities with signboard with the name of facility on the main building 41.38 56.00 88.89 100.00 Average number of buildings or structures on the facility ground 11.34 15.04 27.33 16.00 Number of buildings or structures used for providing health care services 5.24 5.32 16.50 14.00 Percent of facilities indicating that buildings and infrastructure need major repairs 65.52 60.00 72.22 0.00 Percent of facilities with adequate number of toilets 34.48 36.00 61.11 100.00 Percent of facilities indicating that many repairs are needed in patient toilets 55.17 48.00 38.89 0.00 Average number of toilets in the facility 4.83 6.48 23.69 92.00 Average number of rooms in the main building (excluding closets and toilets) 7.03 9.46 13.61 26.00 Percent of facilities with main building roof made of tin or metal 93.1 100 100 100.00 Percent of facilities with main building wall made of cement and brick 10.34 24 38.89 100.00 Percent of facilities with main building floor made of cement and brick 58.62 68 61.11 100.00 Percent of facilities with main facility building accessible for patients with 68.97 72.00 88.89 100.00 disabilities Percent of facilities reporting that main building windows and doors need 48.28 48.00 22.22 0.00 many repairs Percent of facilities reporting that main building interior walls and roof need 44.83 52.00 22.22 0.00 many repairs Percent of facilities reporting need for many repairs in the electrical wiring of 44.83 56.00 11.11 0.00 main building Percent of facilities reporting need for many repairs of the water supply system 75.86 52.00 16.67 0.00 Table 3.2  Infrastructure-Related Variables Affecting Provision of Services and Quality of Care Level 3 and Level 3 and Infrastructural variables 4 public 4 church Level 5 and 6 Level 7 Average number of wards in the facilities 3.28 5.20 8.72 21.00 Average number of hospital beds 21.38 37.68 157.72 886.00 Floor space of facilities in square meters 301 714 3,249 105,114* Percent of facilities open 24 hours a day, 7 days a week 86.21 100.00 100.00 100.00 Percent of facilities providing inpatient services 86.21 96.00 100.00 100.00 Percent of facilities providing emergency services 75.86 100.00 94.44 100.00 Percent of facility connected to electric supply line 34.48 28.00 94.44 100.00 Percent of facilities with functional backup generator 48.28 76.00 100.00 100.00 Percent of facility connected with water supply line 6.90 8.00 83.33 100.00 Percent of facilities in which health care providers are able to use the 82.76 84.00 100.00 100.00 water source for treating patients Percent having telephone/two-way radio for communication 48.28 56 83.33 100.00 Percent of facilities with ambulance (car, boat, or plane) 89.66 96 94.44 100.00 No. of vehicles the facility has (including ambulance) 1.59 1.88 8.56 21.00 Percent of facilities with operation theater 27.59 36.00 100.00 100.00 Average number of operation theaters 0.28 0.40 2.17 7.00 Percent of facilities with facility ground clean 65.52 84.00 88.89 100.00 Percent of facilities with staff housing on facility ground 93.10 100.00 72.22 100.00 Percent of facilities that have designated spaces for clinicians to 68.97 80.00 100.00 100.00 provide services Percent of facilities with functional operation theater where surgeries 13.79 16.00 100.00 100.00 are done regularly Percent of facilities with separate room/space for drug storage 100.00 100.00 100.00 100.00 Percent of drug storage that has secure doors and windows 93.10 96.00 94.44 100.00 % drug storage area has ventilation 79.31 80.00 83.33 100.00 Drugs in the storage area protected from sun and rain 96.55 96.00 94.44 100.00 Percent of facilities with enough shelving in drug storage (no drugs or 13.79 32.00 11.11 0.00 supplies kept on the floor) *Does not include all the floor space of the national health facility. Only the floor spaces of major departments and wards are included Input-Related Aspects of Health Facilities 27 The summary analysis, reported by the infra- Figure 3.1  Infrastructural Indexes for Health Facilities Surveyed structure indexes,16 reconfirms the poor condition of 100.0 100.0 100.0 100.0 level 3 and 4 public health facilities. Figure 3.1 shows 100.0 89.7 the infrastructural indexes by facility type. On average, 79.5 the indexes are higher for level 3 and 4 church-run 80.0 75.6 73.7 65.6 66.0 62.0 facilities than level 3 and 4 public facilities. 60.0 57.6 51.3 46.4 Instruments and Medical Supplies 40.0 38.6 38.6 This section presents information on the avail- 20.0 ability of various equipment and supplies, which constitute an important aspect of inputs used by 0.0 health facilities in producing health care services. Condition Not many Condition: Other Condition: For the purposes of this study and in the context of of toilet repairs needed infrastructural overall commonly prescribed drugs in PNG, the essential Level 3-4 Public Level 3-4 Church Level 5-6 Level 7 medicine list can be found at http://www.who.int/ medicines/publications/essentialmedicines/en/. very basic items, such as a stethoscope, blood pressure- Maternal and Neonatal Equipment and Materials measuring cuff, or tape measures. A significant number of health facilities lack very The indexes of availability of ANC-related basic maternal and neonatal equipment. Table 3.3 equipment and supplies were 71.4 and 81.6 for shows the availability of maternal health-related equip- level 3 and 4 public and church-run facilities, ment and items. The items presented in the table can respectively. The indexes were 81.1 for level 5 and be grouped into two subcategories: items and sup- 6 facilities and 100 for the level 7 facility plies related to antenatal care (ANC) and items and The indexes of availability of delivery and neo- supplies related to maternal or neonatal service. The natal equipment and supplies were 64.7 and 76.0 ANC items include fetal stethoscope, stethoscope for level 3 and 4 public and church-run facilities, and blood pressure cuff, tape measure, scale for tak- respectively. The indexes were 89.9 for level 5 and ing weight, and ultrasound machine (and gel).17 All 6 facilities and 100 for the level 7 facility. health facilities, irrespective of their level, should have these basic items if they provide maternal health Family Planning, Vaccines, and Other Items services. However, not all facilities in the survey had An assessment of family planning items, vaccines, test kits, and other related items stocked in health facilities revealed the need to improve the supply 16 of items to level 3 and 4 public health facilities. The indexes are measured to show the desirable or positive aspects of infrastructural condition or availability. For example, to construct Table 3.4 reports the availability of family planning, an infrastructural index for toilets the following variables were com- vaccines, test kits, and other items. For each supply bined: adequate number of toilets present for the use of patients (= 1, category, an index was constructed to show the aver- if not adequate = 0), most toilets are functional, not many repairs needed in toilets of the facility and most toilets have water. Similarly, age availability of the items in the facilities by facility infrastructural variables included the variables such as the facility has category. The availability index of family planning visible signboard, the facility has electric supply from electric main, items was found to be 81.6 for level 3 and 4 public facility ground is clean, there are staff housing on campus, the facility has designated space for clinicians to provide services, the facility has sector health facilities, 84.0 for level 3 and 4 church- an ambulance, and the main building is accessible to persons with run facilities, and 88.9 for level 5 and 6 facilities. disabilities. Nonavailability of family planning injections in the 17 If a facility reports having the instrument, a value of 1.0 was assigned for the facility for that equipment. The availability index level 7 facility reduced the overall index of family was constructed to show the percent of listed equipment and items planning item availability for the facility to 66.7. As available in each of the facility categories. For example, if the ANC mentioned earlier, since there is only one facility at items availability index is 65 for level 3 and 4 facilities, it means that the facilities on the average had 65 percent of all the ANC equipment level 7, nonavailability of even a single item signifi- and items considered. cantly reduces the overall index. 28 Service Delivery by Health Facilities in Papua New Guinea Table 3.3  Availability of Maternal Health-Related Equipment and Supplies Level 3 and 4 Level 5 and 6 Level 7 ANC, delivery and neonatal % of % of equipment and supplies Public Public Church Church Public % Public % Foetal stethoscope (or monitor) 27 93% 25 100% 14 78% 1 100% Stethoscope and blood pressure cuff 22 76% 25 100% 16 89% 1 100% Tape measure 23 79% 22 88% 16 89% 1 100% Scale 28 97% 25 100% 16 89% 1 100% Ultrasound machine (and gel) 28 97% 25 100% 17 94% 1 100% Facility has at least two skilled birth 21 72% 19 76% 18 100% 1 100% attendants covering 24 hours a day Delivery kit (instruments, supplies) 25 86% 25 100% 18 100% 1 100% Stethoscope 10 34% 11 44% 15 83% 1 100% Partograph 11 38% 13 52% 14 78% 1 100% Pelvic procedure instruments such 18 62% 22 88% 16 89% 1 100% as speculum Delivery light 24 83% 24 96% 18 100% 1 100% Sterilizer 15 52% 16 64% 13 72% 1 100% Vacuum extractor 14 48% 21 84% 18 100% 1 100% Forceps 25 86% 25 100% 18 100% 1 100% Manual vacuum aspirator/suction bulb 27 93% 25 100% 17 94% 1 100% Resuscitation bag, newborn 4 14% 4 16% 15 83% 1 100% Eye drops or ointment for newborn 27 93% 25 100% 18 100% 1 100% Needles and syringes 28 97% 25 100% 18 100% 1 100% Sterile C-section instrument kits 28 97% 25 100% 18 100% 1 100% Cord supplies for newborn: clamps, ties, scissors 2 7% 5 20% 12 67% 1 100% IV sets, including sterilized needle and tube 27 93% 25 100% 17 94% 1 100% IV fluids, including normal saline and ringer lactate 29 100% 25 100% 18 100% 1 100% Index ANC item availability 71.41 81.60 81.11 100.00 Index maternal and neonatal items 64.66 76.00 89.93 100.00 availability (excluding ANC) The vaccine availability index varies from 73.7 this category. Again, when the sample is just one, the for level 3 and 4 public sector facilities to 100 per- availability index is either 100 percent or 0 percent, cent for the level 7 facility. The most limited vaccine which biases the mean availability index because the supply, as reported by the facilities, is the bacillus number of items considered is relatively small. Calmette-Guérin (BCG) vaccine. For test kits and other supplies, the index was 63.1 for level 3 and Availability of Drugs 4 public facilities and 77.7 for level 3 and 4 church- run facilities. Level 5 to 7 facilities reported having Drug availability indexes18 were quite low for all the almost all items considered in this category of test health facility categories. For example, the drug avail- kits and other items. ability index was only 49.1 for level 3 and 4 public sec- Level 3 and 4 public facilities show lower avail- tor health facilities, 54.3 for level 3 and 4 church-run ability for all the different item types compared with level 3 and 4 church-run facilities (figure 3.2). 18 Level 5 to 7 facilities are usually better in availabil- Drug availability is defined as continuous availability of each of ity of items and equipment; and level 7, in general, the drugs in the past 30 days without any stock-outs. In the survey, data were collected on drug availability for 45 different drugs and shows the best availability indexes. The indexes are all these 45 drugs are listed in table 3.5. This index was constructed biased for level 7 because there is only one facility in when all 45 drugs are considered. Input-Related Aspects of Health Facilities 29 Table 3.4  Availability of Family Planning, Vaccine, Test Kits, and Other Items, by Facility Category Level 3 and 4 Level 5 and 6 Level 7 Family planning, vaccine % of % of % of Level 5 % of and other items Public Public Church Church Public and 6 Public Level 7 Family planning items Oral pills 26 90% 21 84% 17 94% 1 100% FP injections 22 76% 22 88% 15 83% 0 0% Condoms 23 79% 20 80% 16 89% 1 100% Index of FP items 81.6 84.0 88.9 66.7 Vaccines and related items BCG vaccines 14 48% 15 60% 9 50% 1 100% Hepatitis B 23 79% 23 92% 16 89% 1 100% Vitamin A 23 79% 24 96% 17 94% 1 100% Sabin/Polio 23 79% 23 92% 15 83% 1 100% DTP/Hib 22 76% 21 84% 15 83% 1 100% Tetanus toxoid 22 76% 24 96% 16 89% 1 100% Measles 21 72% 21 84% 13 72% 1 100% Iron/Folic acid 23 79% 25 100% 14 78% 1 100% Index vaccine 73.7 88.0 79.9 100.0 Test kits and other items HIV/AIDS test kit 15 52% 16 64% 17 94% 1 100% Malaria test kit 22 76% 22 88% 16 89% 1 100% TB Category 1 kit 18 62% 20 80% 17 94% 1 100% TB Category 2 kit 13 45% 14 56% 17 94% 1 100% Oral rehydration salt 27 93% 24 96% 18 100% 1 100% HS Darrow’s 25 86% 20 80% 17 94% 1 100% Oxygen 8 28% 20 80% 18 100% 1 100% Index test kits and other items 63.1 77.7 95.2 100.0 Note: FP = family planning; DTP = diphtheria, tetanus, and pertussis; Hib = Haemophilus influenzae type b; HIV = human immunodeficiency virus; AIDS = acquired immunodeficiency syndrome; TB = tuberculosis. Figure 3.2  Comparison of Availability Indexes, by Facility Type 0 10 20 30 40 50 60 70 80 90 100 Index ANC items Index maternal and neonatal items (excluding ANC) Index of FP items Index vaccines Index test kits and other items Level 3-4 Public Level 3-4 Church Level 5-6 Level 7 30 Service Delivery by Health Facilities in Papua New Guinea facilities, 59.9 for level 5 and 6 facilities, and 62.2 for drugs that were reported to be supplied in insuf- the level 7 facility (table 3.5).19 ficient quantities were antibiotics, analgesics and Some basic drugs were reported not available antipyretics, and antimalarials and antacids. on a continuous basis in many health facilities, The PNG health sector has two congruent systems including the level 7 facility. Less than a quarter for distribution of pharmaceuticals and medical sup- of level 3 and 4 facilities reported nonavailability plies, the so-called pull and push systems. The pull of the following drugs: paracetamol elixir, amodia- system requires health facilities to conduct bimonthly quine (antimalarial agent), captopril (hyper­tension ordering from the Area Medical Store, which is sup- drug), 20 and simvastatin (cholesterol-lowering ported by the NDOH. The push system, which is a drug). Amodiaquine is no longer a first-line treat- system devised to supplement the national pharma- ment for malaria, and so nonavailability of the ceutical and medical supplies system, distributes drug is not surprising. The present first-line treat- standardized kits to all health facilities across the ment drugs for malaria, artemisinin combina- country. Table 3.6 reports the number of different tions, more specifically, artemether-lumefantrine, types of drug boxes received by surveyed facilities or mala-wan, were available in 59 percent and 86 in 2014. Contents of the drug boxes can be found percent of level 3 and 4 public facilities. Clearly, this in National Department of Health documents level of availability is not ideal for malaria prevalent (reproduced in Annex D) and the prices of drugs areas. The poor availability of simvastatin at level 3 are obtained from the Papua New Guinea Medical and 4 health facilities can be explained by two and Dental Catalogue (National Department of requirements. First, the Papua New Guinea Medical Health, 2012). and Dental Catalogue , 10th edition (National Department of Health, 2012) classifies the drug in category B, implying that only medical officers can Clinical and Patient Education Protocols prescribe it. Second, to prescribe the drug, a blood Higher-level health facilities demonstrated a bet- workup is required and pathology (or the health ter representation of protocols than mid- to lower- facility laboratory) needs the capability to run bio- level health facilities. The presence of protocols in chemistry tests to ensure the drug can be prescribed health facilities is the initial step toward improved safely and correctly. Poor availability of simvastatin quality of care. The survey asked the facility respon- at level 5 and 6 health facilities, however, cannot be dents about 11 widely used protocols, such as acute explained by the availability of laboratory services respiratory illnesses (ARI), Integrated Management or doctors. of Childhood Illnesses (IMCI), growth monitor- In general, drug stock-outs at the facilities sur- ing, and tuberculosis (TB) diagnosis and treatment. veyed indicate massive improvement required for Table 3.7 reports the availability of different clinical supply chain and drug procurement. Poor avail- and patient education protocols in the facilities at ability of paracetamol elixir across different levels of the time of the survey. The level 7 facility reported health facilities also points to the management and having most of the protocols, but two of the 11 pro- supply issues of basic drugs. The top three types of tocols were not present in the facility. In general, level 3 and 4 church-run facilities show better avail- ability of protocols than the level 3 and 4 public sector health facilities. The index of protocol avail- 19 Drug availability index was constructed on the basis of contin- ability was 63.6 for level 3 and 4 public health facili- uous availability of a specific drug, not the quantity of the drug available. If a drug was available on the day of the survey but the ties, 82.2 for level 3 and 4 church-run facilities, and drug was not available at any time during the previous 30 days, 84.8 for level 5 and 6 facilities. the drug is considered as “not available on a continuous basis” Higher-level public health facilities show better for the purpose of this availability measure. The underlying assumption is that if a particular drug was available on a contin- recording and registry; for the same level, church- uous basis in the previous 30 days, the quantity of the drug avail- run health facilities are doing better than public able in the facility must have been adequate to meet the need or health facilities. Table 3.8 shows the availability demand. 20 Despite the nonavailability of captopril, another anti­ hypertensive of reports by report or registry type and the over- drug, enalapril, is commonly prescribed in PNG. all availability of reports in the facility by using the Input-Related Aspects of Health Facilities 31 Table 3.5  Facilities Reporting Continuous Drug Availability Over Past 30 Days, by Facility Category Level 3 and 4 Level 3 and 4 Level 5 and 6 Level 7 % of % of % of % of Name of drug Public Public Church Church Public Public Public Public Panadol/paracetamol tablet 20 69% 22 88% 16 89% 1 100% Paracetamol liquid 7 24% 5 20% 6 33% 0 0% Diclofenac (pain/inflammation) 9 31% 9 36% 15 83% 0 0% capsule/tablet Pethidine 15 52% 16 64% 15 83% 0 0% Chloroquine 8 28% 5 20% 9 50% 1 100% Quinine injection 25 86% 24 96% 14 78% 1 100% Primaquine 22 76% 22 88% 17 94% 1 100% Amodiaquine 5 17% 5 20% 12 67% 1 100% Artemisinin combination 17 59% 21 84% 13 72% 1 100% Fansidar 27 93% 25 100% 18 100% 1 100% Sulfadoxine tablet 19 66% 13 52% 7 39% 0 0% Pyrimethamine tablet 14 48% 11 44% 7 39% 0 0% Rifampicin 10 34% 12 48% 8 44% 1 100% Isoniazid 13 45% 18 72% 7 39% 1 100% Ethambutol 9 31% 14 56% 8 44% 1 100% Pyrazinamide 9 31% 11 44% 8 44% 0 0% Streptomycin 15 52% 18 72% 7 39% 0 0% TB blister packs 19 66% 23 92% 10 56% 1 100% Oxytocin 25 86% 24 96% 18 100% 1 100% Ergometrine 23 79% 24 96% 16 89% 1 100% Lignocaine 28 97% 25 100% 18 100% 1 100% Depo-provera 22 76% 22 88% 17 94% 1 100% Ferrous sulfate 25 86% 22 88% 18 100% 0 0% Liniment 11 38% 5 20% 7 39% 0 0% Flagyl/Metronidazole 27 93% 24 96% 17 94% 1 100% Amoxicillin capsule or tablet 24 83% 22 88% 18 100% 1 100% Cotrimoxazole oral susp or tablet 23 79% 23 92% 17 94% 1 100% Diazepam capsule or tablet 19 66% 15 60% 16 89% 1 100% Amitriptyline capsule or tablet 17 59% 6 24% 17 94% 1 100% Paraldehyde 27 93% 21 84% 18 100% 1 100% Albendazole tablet 24 83% 25 100% 18 100% 1 100% Tinidazole tablet 26 90% 25 100% 18 100% 1 100% Ciprofloxacin capsule or tablet 21 72% 16 64% 11 61% 1 100% Crystalline penicillin injection 26 90% 25 100% 17 94% 1 100% Ceftriaxone 1 g/vial injection 11 38% 12 48% 14 78% 1 100% Chloramphenicol injection 26 90% 24 96% 17 94% 0 0% Mala wan (1) 25 86% 23 92% 17 94% 1 100% Captopril, 25 mg, capsule or tablet 7 24% 2 8% 1 6% 0 0% Atenolol, 50 mg, capsule or tablet 16 55% 13 52% 18 100% 1 100% Simvastatin, 20 mg, capsule or tablet 5 17% 1 4% 7 39% 1 100% Glibenclamide capsule or tablet 9 31% 5 20% 17 94% 1 100% (diabetes) Salbutamol, 4 mg, tablet 26 90% 21 84% 16 89% 1 100% CMP (Chlorpheniramine maleate) tablet 16 55% 11 44% 16 89% 1 100% Omeprazole capsule or tablet 5 17% 5 20% 4 22% 1 100% Antiretroviral for HIV/AIDS 12 41% 12 48% 10 56% 1 100% Drug availability index 49.11 54.31 59.87 62.22 32 Service Delivery by Health Facilities in Papua New Guinea Table 3.6  Average Number of Drug Kits Received per Facility, was 48.7 for level 3 and 4 public facilities, 56.6 for by Facility Category, 2014 level 3 and 4 church facilities, 71.8 for level 5 and Level 5 6 public facilities, and 100 for the level 7 facility. Levels 3 and 4 and 6 facilities facilities Level 7 facility Drug box Public Church Public Public Service Availability in Health Facilities types (drug (N = 28) (N = 25) (N = 18) (N = 1) box number) average average average average Higher-level public health facilities can provide Boxes 1–5 11 8 31 80 a higher number of service types than lower-level Boxes 6–14 10 6 33 80 facilities, and church health facilities provide a Boxes 15–22 12 7 25 80 higher number of service types than public health Box 23 7 6 14 80 Box 24 6 7 11 60 facilities at comparable levels. The overall service Box 25 7 6 10 80 availability index was 55.9 for level 3 and 4 public Box 26 1 1 4 80 facilities, 67.4 for level 3 and 4 church-run facili- Box 27 1 1 2 80 ties, 82.5 for level 5 and 6 facilities, and 100.0 for the Box 28 0.1 0.4 0.2 0 Box 29 0 1 0 0 level 7 facility. Again, among level 3 and 4 facilities, Box 30 0 1 0 0 church-run facilities appear to be better prepared for Box 31 0 1 0.5 0 the provision of various types of medical care services Box 32 0 1 0 0 Box 33 0 0.5 0 0 (table 3.9). For level 3 and 4 church-run facilities and Box 34 0 1 0 0 level 5 to 7 facilities, 100 percent reported providing Box 35 0 1 0 0 at least some medical care services 24 hours a day and Box 36 0 0.5 0 0 7 days a week (24/7). Five different types of 24/7 ser- Medical Box 1 0 0.4 0 0 vices are listed in the table and most facilities offer emergency and outpatient services around the clock. Half of level 5 and 6 facilities reported providing weighting factors mentioned above.21 About 96 percent laboratory and diagnostic services around the clock, of church-run level 3 and 4 facilities had antenatal but laboratory services were not available around the care (ANC) register available, while the availabil- clock in most of the level 3 and 4 facilities. Availability ity of ANC register in level 3 and 4 public facilities of 24/7 pharmacy services was also very low overall. was 83 percent and in level 5 and 6 public facilities Only about 39 percent of level 5 and 6 facilities, 31 per- it was 72 percent. Availability of delivery and mater- cent of level 3 and 4 public facilities, and 36 percent nity register was higher than ANC register in level 5 of level 3 and 4 church-run facilities reported offer- and 6 facilities (94 percent). The register and reports ing 24/7 pharmacy services. Only seven facilities out with relatively low rates of availability were opera- of 19 at levels 5 to 7 reported providing around the tion theater register, blood bank register, Extended clock snakebite services. For other services, it is clear Program on Immunization (EPI) report, malaria that some of the services are not routinely organized report, notifiable disease report, and IMCI report. by specific categories of facilities. For example, not Overall index of report and register availability shown in the table, only a few facilities at levels 5 to 7 indicated that they provided ANC at least 5 days a 21 Report and register availability is defined as actual visual veri- week, organized outreach ANC services, or provided fication of the presence of the reports. If the facility mentioned delivery services in the community. that it had the report but could not show the report to the data Level 5 to 7 health facilities are better equipped collection enumerators, it is considered not available. A report or register is considered available if the form or report is actually to conduct a broader range of pathology and labo- filled-in either fully or partially. To calculate the index of avail- ratory investigations compared with level 3 and ability of the Information System, reported availability categories 4 health facilities. An important part of health service were assigned specific weights based on availability and degree of provision is the ability to correctly diagnose a patient’s completion of the forms. The weights assigned to different cat- egories of availability to calculate the information system index condition. Table 3.10 reports the number and percent were as follows: reports and register available and completely of health facilities able to conduct, on the day of the filled in = 100, reports and register available but partially com- survey, a set of specific laboratory or pathological ser- pleted = 67, reports and register not shown but the facility reports vices. The table includes some investigations that can be that they are available =33, not available = 0. If the facility men- tioned that it does not provide the service, it is assigned a value carried out by rapid diagnostic tests (RDTs) and others of zero as well. that require some form of laboratory infrastructure. Input-Related Aspects of Health Facilities 33 Table 3.7  Types of Protocols Present by Facility Category Level 3 and 4 Level 3 and 4 public church Level 5 and 6 Level 7 Protocol type present Number % Number % Number % Number % IMCI protocol 16 55 19 76 17 94 1 100 ARI protocol 11 38 15 60 12 67 0 0 Diagnosis and treatment of diarrhea 17 59 23 92 13 72 1 100 Growth monitoring protocol 19 66 20 80 16 89 1 100 Treatment of severe malnutrition  9 31 18 72 14 78 1 100 TB diagnosis and treatment 20 69 23 92 18 100 1 100 NHIS guidelines 15 52 15 60 12 67 0 0 Malaria treatment protocol 21 72 25 100 17 94 1 100 Immunization schedule 27 93 23 92 15 83 1 100 Family planning 25 86 22 88 16 89 1 100 Patient education material 23 79 23 92 18 100 1 100 Protocol availability index 64 82 85 82 Note: NHIS = National Health Information System. Table 3.8  Availability of Various Registry and Reports in Facilities, by Facility Type Level 3 and 4 facilities Reports available Level 5 and 6 and fully or partially Public facilities Church facilities facilities Level 7 facility completed Number % Number % Number % Number % ANC register 24 83 24 96 13 72 1 100 Delivery, maternity register 24 83 24 96 17 94 1 100 Operation theater register 7 24 6 24 15 83 1 100 Blood bank register 4 14 2  8 17 94 1 100 HIV register 15 52 17 68 18 100 1 100 Tuberculosis register 17 59 22 88 16 89 1 100 Hospital activity report 12 41 14 56 11 61 1 100 EPI report 17 59 14 56 11 61 1 100 Family planning report 21 72 16 64 12 67 1 100 Leprosy report 8 28 4 16 9 50 1 100 Lab register 10 34 12 48 16 89 1 100 Mortality/morbidity report 19 66 15 60 13 72 1 100 Malaria/Kala Azar report 18 62 14 56 12 67 1 100 Notifiable disease report 14 48 14 56 13 72 1 100 IMCI report 12 41 12 48 13 72 1 100 Equipment status report 6 21 11 44 10 56 1 100 Index Information System 49 57 72 100 Diagnostic tools (RDTs) improve the accessibility of the morbidity and mortality of HIV, the growing testing at level 3 and 4 health facilities, as a malaria concern of TB (multidrug-resistant TB and exten- RDT or the RDT for HIV, for example, does not sively drug-resistant TB), and the added problems require a lab. of TB-HIV comorbidities, the number of level 3 Almost all level 3 and 4 facilities reported the and 4 health facilities that are able to do HIV RDTs ability to conduct malaria testing, either by RDT and TB smears is considerably low. Because HIV or by microscopy,22 but were not able to conduct RDTs are relatively simple, it is assumed that this many other critical tests such as pregnancy and ane- is a supply issue, whereas TB smears require a little mia test. Among level 3 and 4 facilities, only 28 per- more knowledge in addition to lab consumables. cent of public sector health facilities and 56 percent More than 90 percent of level 5 to 7 health facili- of church-run facilities were able to conduct preg- ties reported the ability to do all the laboratory tests nancy tests. Ability to do an anemia test was also listed in the table, but the ability of public level 3 and low in level 3 and 4 facilities. In addition, despite 4 facilities to perform basic tests was woefully lack- ing. The laboratory service availability index summa- 22 The questionnaire was not specific. rizes the availability of selected pathology services in 34 Service Delivery by Health Facilities in Papua New Guinea Table 3.9  General Indicators of Service Availability by Level and Ownership Level 3 and 4 Level 5 and 6 Level 7 % of % of Service category Public Public Church Church Public % Public % Facility open 24 hours, 7 days a week 25 86% 25 100% 18 100% 1 100% Emergency services available 24/7 22 76% 25 100% 17 94% 1 100% Outpatient clinics open 24/7 23 79% 23 92% 14 78% 1 100% Snakebite services available 24/7 5 17% 11 44% 6 33% 1 100% Laboratory and diagnostic services available 24/7 5 17% 9 36% 9 50% 1 100% Pharmacy services available 24/7 9 31% 9 36% 7 39% 1 100% Facility has blood bank 1 3% 2 8% 17 94% 1 100% Facility conducts blood transfusions 4 14% 3 12% 17 94% 1 100% Performs surgeries in the facility regularly 4 14% 4 16% 18 100% 1 100% Facility provides ANC services 29 100% 25 100% 15 83% 1 100% Facility provides delivery services (at facility or 27 93% 25 100% 17 94% 1 100% in community) Facility provides neonatal care 21 72% 24 96% 17 94% 1 100% Facility provides postpartum maternal 25 86% 25 100% 18 100% 1 100% examination and treatment Facility provides EPI services 27 93% 25 100% 17 94% 1 100% General service availability index 55.91 67.43 82.54 100.00 the facilities and shows the proportion of laboratory electricity and water in the facilities. Although 95 per- tests that the facilities in the specific category could do cent of level 5 to 7 facilities are connected to the main on the day of the survey. The index for laboratory tests electric supply grid, only about a third of level 3 and 4 was found to be 100 for the level 7 facility and 95 for facilities were connected. More than 80 percent of all level 5 and 6 facilities, implying that all the tests con- facilities connected to the electric supply experienced sidered in the table were available in the level 7 facility blackout at least for some time in the week previous to and 95 percent of the tests, on average, were available the survey. in level 5 to 7 facilities. The indexes were 54.5 for level The survey revealed that improvements to the 3 and 4 church-run facilities, and 29.7 for level 3 and electrical supply, both through main grid connec- 4 public facilities. tion and backup generator, are required at all levels, The quantity and quality of different types of particularly levels 3 and 4. All level 5 to 7 facilities service provision crucially depends on the facility’s had functional backup generators, while the percent- readiness to provide the clinical services. Table 3.11 age of level 3 and 4 facilities with backup generators reports a number of aspects related to facility readiness was 76 percent for church-run facilities and 48 percent in the provision of services. Some of the indicators are for public facilities. Half of the level 3 and 4 facilities related to power and water sources and availability of with backup generators reported having problems Table 3.10  Availability of Laboratory Tests in Surveyed Facilities, by Level and Ownership Level 3 and 4 Level 5 and 6 Level 7 Laboratory or rapid tests Public % Church % Public % Public % Malaria tests (blood test or RDT) 26 90% 24 96% 17 94% 1 100% HIV tests (blood test or RDT) 10 34% 14 56% 16 89% 1 100% TB Smears 7 24% 18 72% 18 100% 1 100% Pregnancy tests 8 28% 14 56% 18 100% 1 100% Anemia 6 21% 14 56% 18 100% 1 100% Urine test R/M/E 5 17% 10 40% 17 94% 1 100% Stool for RE 2 7%  4 16% 17 94% 1 100% Blood glucose 5 17% 11 44% 16 89% 1 100% Laboratory services availability index 29.74 54.5 95.13 100 Input-Related Aspects of Health Facilities 35 Table 3.11  Service Provision Readiness Index, by Facility Level Level 3 and 4 Level 5 and 6 Level 7 Readiness indicators Public (N ) % Church % N % N % Electricity connected to supply grid 10 34.48 7 28.00 17 94.44 1 100.00 Backup generator 14 48.28 19 76.00 18 100.00 1 100.00 Electricity availability (supply or generator) 20 68.97 21 84.00 18 100.00 1 100.00 Blackout last week 8 27.59 6 24.00 15 83.33 0 0.00 Problem last month running generator due 8 27.59 10 40.00 3 16.67 0 0.00 to fuel shortage Water from main line 2 6.90 3 12.00 15 83.33 1 100.00 Water shortage last year 18 62.07 11 44.00 7 38.89 0 0.00 Water available for use by health care 24 82.76 21 84.00 18 100.00 1 100.00 providers Water available in delivery room 13 44.83 18 72.00 18 100.00 1 100.00 Facility does direct blood transfusion 3 10.34 3 12.00 16 88.89 1 100.00 Blood transfusions done last month 3 10.34 3 12.00 18 100.00 1 100.00 Facility has designated space for clinicians 20 68.97 20 80.00 18 100.00 1 100.00 to provide service Facility has on-call room or space for health 3 10.34 11 44.00 10 55.56 1 100.00 care providers to take rest Facility has telephone or shortwave radio 14 48.28 14 56.00 15 83.33 1 100.00 Facility has ambulance 26 89.66 24 96.00 17 94.44 1 100.00 Ambulance out of service last year 11 37.93 10 40.00 6 33.33 0 0.00 Facility has other vehicles 8 27.59 5 20.00 18 100.00 1 100.00 Facility has operation theater 8 27.59 9 36.00 18 100.00 1 100.00 Index of readiness 40.26 48.56 84.57 100.00 in operating the generator in the previous month. not have running water in the facility’s delivery room. If this is the case, on average, only a very small pro- Level 3 and 4 public health facilities perform con- portion of level 3 and 4 facilities had operational and siderably poorly compared with church-run health usable backup generators. About 82 percent of facil- facilities. Only about 45 percent of public facilities ities were either connected to the main electric grid at levels 3 and 4 had water connected to the delivery or had generators. All level 5 to 7 facilities potentially room (and water available on the day of the survey) had electric supply (either main grid connection compared with 72 percent at level 3 and 4 church-run or generators); however, only 84 percent of level 3 health facilities and at all level 5 to 7 health facilities. and 4 church-run facilities and 69 percent of level 3 Water shortage is another major issue. About and 4 public facilities had a stable electric supply. 62 percent, 44 percent, and 38 percent of level 3 Most health facilities have access to water, but and 4 public, level 3 and 4 church, and level 5 and only about 45 percent of public facilities at levels 3 6 facilities, respectively, experienced water shortages in and 4 had water connected to the delivery room. 2014. The survey asked facility respondents whether Water is another critical component in health service the facility was responsible for maintaining the water provision. Facility respondents were asked whether supply system. About 45 percent of facilities were the health care providers had access to water on the responsible for maintaining the water supply system, day of the survey. One hundred percent of level 5 to while it was 64 percent for church-run level 3 and 4 7 facilities and about 80 percent of level 3 and 4 facili- facilities and 68 percent for level 5 to 7 facilities. ties reported that the providers had access to water. Lack of designated space is an issue for about If supply line and water tanks are considered, the 30 percent of level 3 and 4 public facilities and majority of facilities reported having access to water. 20 percent for church-run facilities. Most facilities However, less than 10 percent of level 3 and 4 facilities (90 percent or more) had ambulances and more than were connected with a supply line. What is consider- a third of facilities reported that the ambulances were ably alarming is the number of health facilities that do out of service at least for some time during 2014. 36 Service Delivery by Health Facilities in Papua New Guinea Table 3.12  Average Number of Personnel in Level 3 and 4 Facilities, by Personnel Type Human resource Average number at level 3 and 4 public Average number at level 3 and 4 church category Position Filled % Filled Position Filled % Filled Doctors 0.54 0.29 53.33 0.64 0.36 56.25 Residents 0.00 0.00 — 0.00 0.00 — HEOs 1.46 1.14 78.05 0.64 0.44 68.75 Nurses 5.64 4.18 74.05 9.72 7.76 79.84 CHWs 10.04 8.96 89.32 10.12 8.60 84.98 Other clinical staff 1.61 1.29 80.00 1.60 1.44 90.00 Clinical support staff 0.79 0.71 90.91 0.56 0.52 92.86 Administrative staff 0.96 0.96 100.00 1.80 1.48 82.22 Other nonclinical staff 1.86 1.68 90.38 2.24 2.12 94.64 Other maintenance staff 2.71 2.54 93.42 3.92 3.72 94.90 Total personnel per facility 25.61 21.75 84.94 31.24 26.44 84.64 Both level 3 and 4 public and church-run health The strength of any health system is reliant on a robust facilities show low readiness score for service deliv- health workforce. Tables 3.12 and 3.13 present a ery. Infrastructure and other resources, both required snapshot of the health workforce in the surveyed and available, determine a health facility’s capability facilities and the average number of personnel by to deliver services. A readiness index23 (table 3.11) type in different categories of health facilities (level 3 describes the scale to which health facilities are and 4 public, level 3 and 4 church, level 5 and 6, and capable of providing services. Overall, level 3 and level 7 public facilities). Annex C (table C.3) lists all 4 facilities, both public and church-run, score quite the personnel types or designations reported by the low. The readiness indexes were found to be 40.3 for facilities. Table 3.12 is a summary table derived from level 3 and 4 public sector facilities, 48.6 for level 3 the annex table by categorizing the personnel into a and 4 church-run facilities, 84.6 for level 5 and 6 facil- number of broad groups. ities, and 100 percent for the level 7 facility. Level 3 and 4 health facilities were predomi- PNG presents comparatively low in readiness nantly staffed by community health workers (CHWs) in providing health services. Comparison of health and nurses. To present the information on person- facility readiness across countries is often tricky nel availability in a summary form, all personnel types because of differences in the level and comprehensive- or designations were grouped into the following ness of facilities in different countries of the world. categories: doctors, residents (resident medical officer A report indicates that 98 percent of public hospi- and resident health extension officer), health exten- tals in Indonesia had electricity (World Bank, 2014). sion officers (HEOs), nurses, CHWs, (other) clinical In Zambia and Kenya, 57 percent and 34 percent of staff, clinical support staff, administrative staff, (other) health centers (not hospitals) reported having electric- nonclinical staff, (other) maintenance staff. Level 3 and ity. In Laos, the index of basic amenities was reported 4 health facilities were predominantly staffed by CHWs as 64 percent (World Bank, October 2013) and the and nurses. The CHWs represent about 41 percent of overall index of basic amenities for PNG was 58 per- personnel in level 3 and 4 public sector facilities and cent (weighted mean of the values in table 3.11). 33 percent in level 3 and 4 church-run facilities. Nurses represent another 19 percent and 29 percent of per- sonnel in public and church-run facilities, respectively. Human Resources The ratio of nurses to CHWs is higher in level 3 and 4 PNG is experiencing critical shortages in key church-run health facilities compared with public health health staff cadres, which is leading to an inequity facilities. There are more nurses and CHWs working in of care between facility level, type, and location. level 3 and 4 church-run health facilities; whereas level 3 and 4 public health facilities may employ a similar num- ber of CHWs, they employ fewer nurses. This may mean 23 The readiness index reflects the percent of all readiness aspects that the quality of service, in relation to diagnosis, treat- or variables satisfied by the health facilities in a specific category. ment, and patient management, may be poorer at level 3 For example, if the readiness index of a facility category is 32.5, it means that 32.5 percent of all the readiness indicators were satis- and 4 government-run health facilities compared with fied on average for the facility category. church-run health facilities. Input-Related Aspects of Health Facilities 37 Table 3.13  Average Number of Personnel in Level 5 to 7 Facilities, by Personnel Type Level 5 and 6 public facilities Level 7 facility Personnel type Position Filled % Filled Position Filled % Filled Doctors 25.50 18.78 73.64 176.00 133.00 75.57 Residents 1.39 1.67 120.00 58.00 40.00 68.97 HEOs 7.11 7.39 103.91 4.00 3.00 75.00 Nurses 97.39 81.72 83.91 633.00 435.00 68.72 CHWs 62.00 56.22 90.68 235.00 216.00 91.91 Other clinical staff 24.94 20.72 83.07 85.00 57.00 67.06 Clinical support staff 8.67 6.94 80.13 13.00 12.00 92.31 Administrative staff 44.44 39.22 88.25 144.00 115.00 79.86 Other nonclinical staff 28.11 25.78 91.70 66.00 66.00 100.00 Other maintenance staff 31.06 29.39 94.63 57.00 52.00 91.23 Average 330.61 287.83 87.06 1471.00 1129.00 76.75 There is a general shortage of doctors, and even need to provide more education opportunities for in higher-level facilities CHWs and nurses make women at the upper levels of medical training in up a large proportion of the clinical workforce. PNG. Only two female doctors were in level 3 and Table 3.13 shows the number of different types of 4 church-run facilities. Even in level 5 and 6 and personnel in level 5 and 6 and level 7 facilities. The level 7 facilities, the percentage of female medical average personnel in level 5 and 6 facilities and the officers is about 25 percent. The percentage of other level 7 facility were about 287 and 1,129, respectively. doctors (as a percent of total) was better in level 5 to Similar to level 3 and 4 health facilities, collectively 7 facilities. Even among residents, the percentage of CHWs and nurses make up a large proportion of the females was aboutone third, implying that the gender clinical workforce. Administrative staff numbers are ratio among doctors will not change significantly in also quite high in these facilities. the near future. The table indicates that 80 percent or Of the health facilities surveyed there were no more of nurses were female in all facility categories female medical officers, including dental officers except the level 7 facility (where it was 62 percent). and radioloists, employed at level 3 and 4 public Other characteristics of the health workforce are pre- health facilities (table 3.14). This implies the strong sented in Annex B of the report. Table 3.14  Total Number of Female Personnel and Percentage of Female Personnel, by Facility Category and Cadre Level 3 and 4 public Level 3 and 4 church Level 5 and 6 Level 7 facility facilities facilities facilities (tertiary) % Female % Female % Female % Female Personnel type Female of total Female of total Female of total Female of total Medical officers 0 0.00 2 40.00 74 26.52 33 25.58 Other doctors 0 0.00 0 0.00 18 32.73 2 50.00 Residents 0 — 0 — 10 33.33 14 35.00 HEOs and RHEOs 17 53.13 4 36.36 64 48.12 1 33.33 Nursing officers 94 80.34 151 79.47 1155 85.87 104 61.90 CHWs 160 63.75 143 68.75 759 79.06 127 58.80 Midwife 12 100.00 16 100.00 97 100.00 8 100.00 Dental personnel 3 42.86 1 33.33 20 44.44 9 60.00 Pharmacy and lab 2 11.76 5 23.81 69 42.86 19 61.29 personnel Office staff 15 65.22 15 62.50 176 63.77 20 58.82 Driver 1 3.33 0 0.00 0 0.00 0 0.00 Cleaner/handyman 16 39.02 26 52.00 118 44.19 4 33.33 Kitchen staff 2 25.00 5 62.50 68 61.26 2 28.57 Other staff 19 27.94 18 24.66 639 46.92 214 48.64 Total 341 55.54 386 60.60 3267 62.53 557 49.73 38 Service Delivery by Health Facilities in Papua New Guinea Table 3.15  Percent of Providers Reporting a Second the maximum at 20 (table 3.16). Doctors, nurses, Job Outside the Surveyed Facility, by Provider Type and HEOs in church-run health facilities reported and Facility Level higher satisfactory scores. CHWs reported very Level 3 and Level 3 and Level 5 to similar levels of satisfaction across all the facility Provider type 4 public 4 church 7 public types. Doctor 0 0 4.88 Most health workers reported that the availabil- HEO 9.09 11.11 31.25 ity of drugs, equipment, and physical condition of Nurse 21.15 29.17 14.29 CHW 22.58 24.53 14.71 the health facilities is the most important factor in affecting their satisfaction level. This result is con- sistent with other findings. The “enablers,” those Working Hours and Workload that are essential for health workers to do their job, are highly correlated with their job satisfaction level. All health workers interviewed reported spending Training opportunities are important, followed by 40 to 50 hours per week and indicated the workload availability of living accommodation and salaries is high or moderately high. These findings, to some (table 3.17). Such perceptions are quite consistent extent, are not consistent with anecdotal evidence. across facility types.25 Therefore, the report does not present the tables; Concerning pride, the health care providers in however, they are available on request. the church-run facilities show the highest level of Overall, the proportion of health workforce motivation. Table 3.18 shows the motivation indexes employees who maintained two jobs was very low. of the health care providers interviewed.26 The pride Only about 5 percent of doctors employed in level 5 to index was lowest for the level 5 to 7 public facilities 7 health facilities reported having a second job, whereas and highest for church-run facilities. Concerning the 31 percent of HEOs reported having a second job financial reward aspect, the doctors were quite happy (table 3.15). Among nurses and CHWs at level 3 and in the church-run facilities, but the other person- 4 health facilities, about one fourth reported hav- nel in church facilities reported the lowest levels of ing another job, but the number was lower at about motivation. 15 percent for nurses and CHWs employed at level 5 to 7 facilities. Maintaining two jobs creates potential conflict Provider Knowledge Index of interest situations and may affect the commitment of health care providers to their principal workplace and A high proportion of health care providers quality of services delivered in the principal health care reported being involved in the provision of prior- facilities. In general, however, it appears that the prob- ity health services. A well-trained health workforce lem may not be a serious concern for PNG, especially is important to improving population health out- because the proportion of health care providers, espe- comes. The survey questionnaire included a set of cially the doctors, reporting a second job was very low. questions to understand the knowledge of health care providers related to priority health services. Health Care Provider Satisfaction Percentage of providers providing child health ser- vices, maternal health services, and sexually trans- The average satisfaction scores24 were quite similar mitted infection (STI) services by provider type is across personnel types, ranging from 10 to 12 with reported in table 3.19. In general, providers of child health services received low scores related to important child 24 The satisfaction score or index was calculated by assigning a value of 1.0 if the provider indicated that she or he is generally satisfied with a specific situation, 0.5 for neutral, and 0 for dissatis- 25 fied (refer to question numbers 401 to 420 of Provider Interview Not reported here but available on request. 26 form). If the values for all the 20 aspects or situations are added The motivation indexes were constructed by assigning a value of together, the overall satisfaction score or index for the provider 1.0 for the highest level of motivation and a value of zero for the can be derived. Since there are 20 aspects or situations defined, the lowest level of motivation implied by the responses to the ques- maximum possible satisfaction score will be 20 and the minimum tions related to different aspects of motivation. The lowest level of will be zero (there are some missing values in the data set and so motivation happens when a health care provider disagrees with all this scoring system may create some bias. See Annex table A.8 for the statements in the questionnaire that represent different types estimates of satisfaction scores when missing values are taken into of “motivators” (for example, it is a source of pride to be able to account). work in the facility). Input-Related Aspects of Health Facilities 39 Table 3.16  Health Care Providers Satisfaction Scores, by Facility Type Average Standard Coefficient Provider satisfaction deviation of variation Minimum Maximum type Facility type score (M) (SD) (SD/M) value value Doctor Level 3 and 4, public 12.20 3.409 0.279 8.0 17.0 Level 3 and 4, church 14.25 4.052 0.284 10.5 18.0 Level 5 to 7, public 12.35 3.000 0.243 6.0 17.5 HEO Level 3 and 4, public 10.93 3.646 0.334 2.0 18.0 Level 3 and 4, church 12.22 3.784 0.310 7.5 20.0 Level 5 to 7, public 12.13 4.089 0.337 3.5 20.0 Nurse Level 3 and 4, public 10.84 3.146 0.290 3.5 16.5 Level 3 and 4, church 11.51 3.459 0.301 4.0 20.0 Level 5 to 7, public 12.16 3.700 0.304 3.0 20.0 CHW Level 3 and 4, public 11.44 3.809 0.333 3.0 18.5 Level 3 and 4, church 11.95 2.783 0.233 6.5 18.0 Level 5 to 7, public 11.69 3.550 0.304 5.0 18.0 Table 3.17  Three Most Important Factors Affecting Table 3.18  Motivation Index, by Dimension, Provider Satisfaction Facility Type, and Provider Type Aspects or factors Motivation most important index by Provider type (three most dimensions of important factors) MO HEO NO CHW motivation Doctor HEO Nurse CHW Availability of equipment 70.91 68.09 54.81 50.34 Pride index in the health facility Public facilities, 0.688 0.716 0.719 0.804 Availability of medicines 63.64 34.04 38.52 33.56 level 3 and 4 in the health facility Church facilities, 0.875 0.722 0.823 0.844 Your training opportunities 21.82 23.40 31.11 32.89 level 3 and 4 to upgrade your skills and Public facilities, 0.616 0.695 0.718 0.750 knowledge level 5 to 7 Living accommodations 10.91 29.79 29.63 31.54 for your family Financial reward index The physical condition of 23.64 21.28 25.19 30.87 Public facilities, 0.650 0.532 0.606 0.611 the health facility building level 3 and 4 Your salary 10.91 19.15 28.89 26.85 Church facilities, 0.850 0.456 0.517 0.462 level 3 and 4 Your ability to provide 14.55 14.89 9.63 12.08 high quality of care Public facilities, 0.541 0.575 0.623 0.579 level 5 to 7 Management of the 14.55 14.89 9.63 7.38 health facility Self-efficacy index Working relationships 9.09 8.51 14.07 14.07 Public facilities, 0.790 0.755 0.790 0.842 with upper-level staff level 3 and 4 Employment benefits (travel 3.64 14.89 11.85 13.42 Church facilities, 0.950 0.744 0.804 0.791 allowance, bonus, etc.) level 3 and 4 Note: MO = medical officer; NO = nursing officer. Public facilities, 0.751 0.738 0.769 0.735 level 5 to 7 Index of conscientiousness Public facilities, 0.867 0.845 0.878 0.876 level 3 and 4 Church facilities, 0.938 0.824 0.865 0.829 level 3 and 4 Public facilities, 0.839 0.818 0.867 0.850 level 5 to 7 40 Service Delivery by Health Facilities in Papua New Guinea Table 3.19  Percentage of Health Staff Providing Health workers have much better knowledge Priority Health Services level on STIs and the possible consequences. The Percentage of providers reported knowledge index related to STIs was derived by com- providing the following services bining knowledge of symptoms of STI, consequences Provider Child Maternal STI of STI on women, and consequences of STI on chil- type health health treatment dren. The overall index was found to be quite high for Doctor 65.45 43.64 62.96 all the provider types in both level 3 and 4 and level 5 HEO 95.56 61.70 72.34 Nurse 91.34 72.39 71.85 to 7 facilities. For doctors in level 3 and 4 and level 5 CHW 85.31 63.76 63.76 to 7 facilities, the average indexes were 0.89 and 0.91, All 85.68 63.64 67.53 respectively. For HEOs, nurses, and CHWs, the indexes Note: Among the providers interviewed during the survey. were 0.86, 0.81, and 0.77, respectively, irrespective of level of facility. health concerns. Two aspects of child health ser- Summary vice are reported in table 3.20: whether the provid- ers can correctly identify the danger signs among Facility Readiness children and whether the providers know what The overall health facility infrastructure condi- is meant by fast breathing in diagnosing children tion is quite poor. Most facilities reported the need with respiratory illnesses.27 Health care providers, for major building repairs and emphasized the lack including the medical officers, received relatively of adequate toilets, stable electrical supply, and consis- low knowledge scores. None of the health care pro- tent water supply. Most level 3 and 4 health facili- viders received higher than a 0.7 score. ties require better connectivity to electricity and clean The knowledge index related to maternal water supply. On average, the indexes are lower for health is quite low for all provider types in both level 3 and 4 public health facilities than level 3 and level 3 and 4 and level 5 to 7 facilities (table 3.21). 4 church-run facilities. Note that only the providers who reported that An assessment of family planning items, vaccines, they provide maternal health services were asked test kits, and other related items stocked in health these knowledge questions. When asked to define facilities revealed the need to improve the supply of maternal mortality, most could not provide the items to level 3 and 4 public health facilities. Level 3 correct definition. The inability to define maternal and 4 public facilities show lower availability for all the mortality may not be a significant concern if the different item types surveyed, as compared with level 3 providers know what to do in relevant situations. and 4 church-run facilities. A significant number of Even in these questions, the proportion of correct health facilities lack very basic maternal and neonatal responses was relatively low. equipment. The vaccine availability index varies from 73.7 for level 3 and 4 public sector facilities to 100 per- cent for the level 7 facility. The most limited vaccine Table 3.20  Knowledge Index Related to Child Health Services, supply, as reported by the facilities, is the BCG vaccine. by Provider Type and Facility Level Level 5 to 7 facilities are usually better in availability Index of knowledge Index of knowledge of items and equipment; level 7, in general, shows the Provider related to danger signs related to fast breathing best availability indexes. type Level 5 to 7 Level 3 and 4 Level 5 to 7 Level 3 and 4 Higher-level public health facilities can provide Doctor 0.458 0.564 0.440 0.600 more services than lower-level facilities, and church HEO 0.486 0.529 0.643 0.529 health facilities provide more services than pub- Nurse 0.469 0.482 0.496 0.540 lic health facilities at comparable levels. The over- CHW 0.538 0.474 0.570 0.521 all service availability index was 55.9 for level 3 and 4 public facilities, 67.4 for level 3 and 4 church-run 27 The indexes were calculated using the following scoring system: facilities, 82.5 for level 5 and 6 facilities, and 100.0 for Provider knowledge on child health services: (i) Danger signs: no the level 7 facility. Again, among level 3 and 4 facili- errors = 1, one error = 0.8, two errors = 0.6, three errors = 0.4, ties, church-run facilities appear to be better prepared four errors = 0.2, five or six errors = 0. (ii) Fast breathing index: no errors = 1 and after that a similar approach as in danger signs was for the provision of various types of medical care used to obtain the index of knowledge. services. Input-Related Aspects of Health Facilities 41 Table 3.21  Knowledge Index Related to Maternal Health Services, by Provider Type and Facility Level Doctor HEO Nurse CHW Knowledge of maternal health Level 3 Level 5 Level 3 Level 5 Level 3 Level 5 Level 3 Level 5 related to the following and 4 to 7 and 4 to 7 and 4 to 7 and 4 to 7 Danger signs of pregnancy 0.58 0.64 0.55 0.50 0.60 0.64 0.57 0.56 What to do: vaginal bleeding in late 0.46 0.43 0.63 0.56 0.64 0.45 0.72 0.67 pregnancy What to do: convulsions/unconscious 0.61 0.58 0.66 0.71 0.68 0.65 0.65 0.70 Danger signs during labor/delivery 0.71 0.71 0.69 0.64 0.68 0.65 0.69 0.70 Obstetric complications 0.83 0.84 0.69 0.69 0.55 0.48 0.43 0.53 Maternal mortality definition 0.33 0.36 0.28 0.75 0.27 0.36 0.30 0.00 Overall knowledge Index 0.59 0.59 0.58 0.64 0.57 0.54 0.56 0.53 More than 90 percent of level 5 to 7 health facili- reported to be supplied in insufficient quantities were ties reported the ability to do all the laboratory antibiotics, paracetamol or other pain medicines, and tests surveyed, but the ability of level 3 and 4 public antimalarial and antacids. facilities to perform basic tests was woefully lacking. Almost all level 3 and 4 facilities reported the abil- Human Resources for Health ity to conduct malaria testing, either by RDT or by PNG has an acute shortage of physicians, particularly microscopy, but were not able to conduct many other in lower-level facilities. Nursing officers and CHWs critical tests, such as pregnancy and anemia tests. make up the majority of the workforce employed Upper-level facilities scored significantly higher on at health facilities, excluding nonclinical staff. The the index for laboratory tests: 100 for the level 7 facil- number of nonclinical staff is quite high in all facility ity and 95 for level 5 and 6 facilities, implying that all types. Level 5 to 7 health facilities tend to complement the tests considered were available in the level 7 facili- doctors with trained clinical “specialist” staff, such as ties and 95 percent of the tests were available in level 5 radiographers, physiotherapists, and anesthetic techni- and 6 facilities. The indexes were 54.5 for level 3 and cal officers, but these types of specialist staff are often 4 church-run facilities and 29.7 for level 3 and 4 public not present at level 3 and 4 facilities. facilities. A significant number of budgeted facility posi- The quantity and quality of different types of tions remained vacant. At level 3 and 4 health facili- service provision crucially depends on the facility’s ties, only a small percentage of facilities had medical readiness to provide the clinical services. Overall, officer positions officially sanctioned, but more than level 3 and 4 facilities, both public and church-run, 40 percent of these positions remained unfilled. scored quite low on the readiness index, which Even for level 5 to 7 facilities, about one fourth of describes the extent to which facilities are capable of medical officer positions were vacant, implying over- providing services. The readiness indexes were found all shortage of doctors in the country. Overall, more to be 40.3 for level 3 and 4 public sector facilities, than 20 percent of nursing officer and HEO positions 48.6 for level 3 and 4 church-run facilities, 84.6 for were not filled. level 5 and 6 facilities, and 100 for the level 7 facility. Gender disparities are apparent in employment Drug availability indexes were quite low for patterns and personnel composition. Only 26 per- all types of health facility. Some basic drugs were cent of all medical officers are female. Among resi- reported to be not available on a continuous basis in dents, only about 34 percent were female, implying many health facilities, including the level 7 facility. that the male-to-female ratio of doctors will improve Poor availability of paracetamol elixir across different in the future but by a relatively small amount. Overall, levels of health facilities also points to the manage- about 60 percent of all health facility personnel were ment and supply issues of basic drugs. In general, drug female and most were nurses, CHWs, midwives, and stock-outs at the facilities surveyed indicate the mas- kitchen staff. At the time of data collection, there sive improvement required for supply chain and drug were only two female chief executive officers (either procurement. The top three types of drugs that were at district or provincial level). 42 Service Delivery by Health Facilities in Papua New Guinea Health workers have similar weekly working aspects that affect their level of satisfaction, including hours with the exception of doctors at the level 7 the condition of the health facility infrastructure and facility; outreach varies by facility level. In rela- availability of medicine, equipment, and supplies. tion to outreach, more nurses and CHWs conduct Most health care providers performed poorly in outreach patrols from level 3 and 4 health facilities, the knowledge index, especially for maternal and whereas doctors at level 5 and 6 facilities reported child health issues. Health care providers were also greater involvement with outreach activities. The asked a series of questions to determine their level of level 7 health facility did not conduct outreach knowledge on priority health services of PNG. The in 2014, and it is not surprising given that it is a questions were focused on three areas: child health, tertiary-level facility located in the urban center. maternal health, and sexually transmitted diseases. Health workers consider their workload moder- Most health care providers, including the fully ately high or very high. This perception may not be trained doctors, performed poorly in the knowledge a direct reflection of hours worked, but perhaps the index, especially for maternal and child health issues. patient load. Despite the “high” workload, health work- Knowledge indexes were better for sexually transmit- ers are reasonably satisfied with their jobs. The deter- ted diseases, although the scores remained around minants of satisfaction are not the same across health 80 (80 out of 100). Strengthening in-service train- worker cadre or across health facility level. In the survey, ings should be considered to improve the level of health care providers mentioned a number of specific knowledge of health care providers. Chapter 4 MANAGEMENT, ACCESS, EQUITY, QUALITY, AND SAFETY OF HEALTH SERVICE DELIVERY This chapter first presents administrative and man- 28 percent of level 3 and 4 public facilities have agement issues and then discusses access and equity hospital or facility community advisory commit- issues. Management, leadership, and governance are tees and 53 percent for level 5 to 7 facilities. About crucial elements to determine how efficiently inputs 62 percent of level 3 and 4 public health facili- can be transformed to outputs. ties reported having a staff meeting in the previous 3 months. A higher proportion of church facilities had staff meetings, about 84 percent, while 74 percent of Leadership and Governance level 5 to 7 facilities reported having staff meetings. Effective and stable leadership and governance are Most of the facilities did not have any asset reg- critical components of a successful workforce. The istry. Only 10 percent of level 3 and 4 public health study revealed significant gaps in the day-to-day facilities had an asset registry compared with about leadership and governance structures within PNG’s 40 percent for other facilities. In level 3 and 4 facili- health system. ties, about 65 percent reported that they had enough Not all health care providers arrive on time for official registers, other stationery, and supplies. work or present in the facility during their assigned More than 90 percent of level 5 to 7 facilities had hours. The percentage of facilities reporting health care supplies and stationery. More than 80 percent of providers starting work on time (starting at 7:45 a.m.) facilities reported that they had enough admission was 52 percent for level 3 and 4 public health facili- and discharge forms. ties, 60 percent for level 3 and 4 church facilities, and While there appears to be theoretical acknowl- only 11 percent for level 5 to 7 facilities. As mentioned edgment that outside supervisory visits are impor- earlier, only five level 3 and 4 public facilities and four tant, in practice visits are rare. More than 50 percent level 3 and 4 church facilities had doctors, but at the of facilities reported that they have officially assigned time of the survey a doctor was present in three (out of external supervisors and the percentages were 79 per- five) for public and in all church-run health facilities. cent for level 3 and 4 public facilities, 60 percent for Although all the level 5 and 7 facilities employ doc- level 3 and 4 church facilities, and 53 percent for level tors, only 37 percent reported the presence of at least 5 to 7 facilities. Only a very small number of facilities one doctor at the time of the survey. Table 4.1 shows a were visited by supervisors in the previous 1 year: eight number of management- and governance-related indi- for level 3 and 4 public facilities, two for level 3 and cators. Some of the indicators are direct measures of 4 church facilities, and none for level 5 to 7 facilities. management, while others are related to outcomes of Facility authority levels are very limited at the better management. lower (3 and 4) levels and increase at the upper Seventy-six percent of level 3 and 4 church-run (5 to 7) levels. Level 5 to 7 facilities have greater health facilities have hospital or facility commu- authority to make expenditure decisions, while only nity advisory committees, which is considerably about one-fourth of level 3 and 4 public health more than public health facilities surveyed. Only facilities have a similar authority level. Table 4.2 43 44 Service Delivery by Health Facilities in Papua New Guinea Table 4.1  Health Facility Management and Governance Index, by Facility Level and Ownership Level 3 and 4 facilities Level 5 to 7 facilities Public facilities Church facilities (all public) % of % of % of Management- and governance-related indicators Number Public Number Church Number Facilities Health care providers arrived at the facility on time today 15 51.72% 15 60% 2 10.52% Doctors present at the time of survey 3 10.34% 4 16% 7 36.84% Meetings held of facility staff in past 3 months 18 62.06% 21 84% 14 73.68% Hospital/facility community advisory committee/ 8 27.58% 19 76% 10 52.63% village health committee exists Facility has written records of activities carried out by 6 20.68% 16 64% 9 47.36% the hospital/facility community advisory committee Facility has asset registry (equipment and furniture 3 10.34% 10 40% 8 42.10% inventory) There is an officially assigned external supervisor for 23 79.31% 15 60% 10 52.63% health facility Supervisors visited in 1 year 8 27.5% 2 8% 0 0% summarizes the level of authority in health facility that they reported much higher levels of satisfaction management. with the authority they have. Government-run level In relation to personnel decisions, most of level 3 3 and 4 health facilities are less satisfied with less and 4 facilities, both public and church-run, did authority or decision-making power compared with not have authority to hire new staff or set their church-run level 3 and 4 health facilities. salary levels. Most are also not allowed to dismiss About half of the providers interviewed reported personnel from health facilities or decide on their that they had a last supervisory interaction within promotions. Level 3 and 4 facilities are not given the a month of the survey. National level (National authority to decide the types of services to be deliv- Department of Health) is responsible for supervision ered or to provide new type of services. Most facilities of level 5 to 7 facilities, whereas provincial govern- at levels 3 and 4 do not have the authority to decide ments (and Provincial Health Authorities in some the charges or user fees for the services they deliver. cases), in collaboration with church health providers, When asked whether the managers were happy with are responsible for managing level 3 and 4 facilities. the level of authority they had on decision making Level 5 to 7 facilities reported the highest number related to funds use, human resources, and organiza- of supervisory interactions within the preceding tion of production, 55 percent of level 3 and 4 public month. About 43 percent of providers at level 3 and 4 facilities indicated that they were satisfied. The sat- public health facilities reported their last supervisory isfaction levels were significantly higher for level 3 interaction at more than 6 months or never. For level and 4 church-run facilities and level 5 to 7 facilities, 3 and 4 church-run facilities, 36 percent reported last with 84 percent and 89 percent of facilities report- interaction at more than 6 months or never. A signifi- ing satisfaction, respectively. Level 3 and 4 public cant proportion, 24.8 percent, of providers employed facilities show the lowest authority indexes on fund at the level 5 to 7 facilities reported that they have use, human resources, and organization of produc- never interacted with their supervisors. However, the tion. Using the levels of authority mentioned by the category “never” is quite nonspecific because a newly facilities, three indexes were constructed: authority appointed person may not have interacted with the index for use of funds, human resource, and orga- supervisor at all simply because of recent entry into nization of production. The indexes are also shown the workplace. Table 4.3 shows the days since the last in table 4.2. Level 3 and 4 public facilities show the interaction with the supervisor. lowest authority indexes among the three catego- Supervisory visits throughout the PNG health ries of health facilities presented. Even though the system could be strengthened to include supply indexes for church facilities were slightly higher than delivery, health consultation feedback, and facil- for public facilities at levels 3 and 4, it is interesting ity and financial records assessment and oversight. Management, Access, Equity, Quality, and Safety of Health Service Delivery 45 Table 4.2  Facility Authority Index Related to Decision Making on Expenditure, Personnel, and Service Delivery Level 3 and 4 facilities Level 5 to 7 facilities Church facilities Church facilities (all public) % of % of % of Management- and governance-related indicators Number Public Number Church Number Facilities Has authority to make expenditure decisions Total amount to be spent 8 27.58% 13 52% 16 84.21% Reallocation between categories 5 17.24% 9 36% 11 57.89% Can sign checks to pay suppliers and other vendors 7 24.13% 11 44% 16 84.21% Need preapprovals for exercising the authorities that 19 65.61% 15 60% 13 68.42% facility managers have related to fund allocation Index of authority related to use of funds 30.5 47.3 75.4 Has authority in decisions related to personnel Disciplinary action/reporting for poor performance 21 72.41% 16 64% 16 84.21% Identify needs for additional/new staff Hiring of new staff 21 72.41% 24 96% 17 89.47% Determine level of payment for staff 6 20.68% 4 16% 17 89.47% Allocating tasks/duties to existing staff 5 17.24% 6 24% 12 63.15% Commending staff for good performance 21 72.41% 23 92% 19 100% Promoting staff to higher position 23 79.31% 24 96% 18 94.73% Dismissal of staff 7 24.13% 9 36% 15 78.94% Index of authority related to personnel 47.0 54.5 82.9 Authority on type/organization of service provision Planning the schedule of services 24 82.75% 20 80% 17 89.47% Changing the content of services 9 31.03% 9 36% 16 84.21% Delivered Quality of services 19 65.51% 17 68% 18 94.73% Operating procedures 9 31.03% 12 48% 17 89.47% Procedures for accessing services 15 51.72% 11 44% 16 84.21% Charges for services 10 34.48% 9 36% 10 52.63% Addition of new services 5 17.24% 6 24% 13 68.42% Index of authority related to service delivery 44.8 48.0 80.5 Satisfied with the level of authority 16 55.17% 21 84% 17 89.47% Types of authority facility managers reported needing to improve performance of health facility Authority to generate revenue 5 17.24% 5 20% 0 0% Authority to allocate budget 12 41.37% 5 20% 3 15.78% Authority to hire and fire staff 5 17.24% 2 8% 1 5.26% Authority to give reward/punishment 5 17.24% 2 8% 1 5.26% Authority to plan 12 41.37% 5 20% 1 5.26% Authority of quality assurance 5 17.24% 2  8% 1 5.26% Table 4.3  Days Since Last Supervisory Interaction, by Facility Type Days since last interaction with supervisor Level 3 and 4 public Level 3 and 4 church Level 5 to 7 public Total <30 days 45.9 44.7 67.2 52.5 30 to 90 days 6.2 12.3 2.4 6.7 3 to 6 months 4.8 7.0 0.8 4.2 > 6 months 8.9 13.2 4.8 8.8 Never 34.2 22.8 24.8 27.8 All facilities 100.0 100.0 100.0 100.0 46 Service Delivery by Health Facilities in Papua New Guinea Table 4.4  Supervisor Activities at Most Recent Supervisory Visit Percent of providers reporting the activity Activities done by supervisors Level 3 and Level 3 and Level 5 to during last interaction 4 Public 4 church 7 public % of All N Brought supplies 8.22 7.89 3.17 6.48 25 Checked records 20.55 20.18 15.87 18.91 73 Checked finances 2.74 7.02 0.79 3.37 13 Observed consultations 8.9 8.77 21.43 12.95 50 Asked knowledge questions 7.53 20.18 17.46 14.51 56 Provided health-related instructions 32.19 46.49 39.68 38.86 150 Provided administrative instructions 34.25 57.02 30.95 39.90 154 Provided instructions on NHIS forms 2.05 6.14 3.97 3.89 15 Other activitiesa 26.03 21.05 26.98 24.87 96 Note: NHIS = National Health Information System. a. Other includes checking attendance of staff, help in clinical activities, staff performance monitoring, checking on infrastructure and construction of new building, incident-related visit. Table 4.4 shows the types of activities done by the of services based on the limited number of health supervisors during their last visit or interaction. The facilities surveyed. The number of facilities sur- most common activity reported by the facilities was veyed is 18 in Southern region, 21 in Highlands, “the supervisor provided administrative instruc- 17 in Momase, and 17 in New Guinea Islands tions”; about 40 percent of all facilities interviewed (NGI).28 Table 4.5 reports the availability of differ- reported this specific activity. About 39 percent ent types of 24/7 services in the surveyed facilities. reported that the supervisors provided health-related Most of the facilities surveyed in all the four regions instructions. Only a small proportion of supervi- were open 24/7 for the provision of at least some sory visits were associated with financial assessment limited services. Inpatient service, by definition, is a or supplies delivery. Notably, only about 19 percent 24/7 type of service and therefore 93 percent of sur- reported that the supervisors checked the facil- vey facilities reporting provision of inpatient services ity records during the last interaction. Only about remain open for the provision. Laboratory, diagnos- 13 percent reported that the supervisors observed tic, and pharmacy services were available in about consultations. Only about one-fourth of facilities one-third of all facilities surveyed, but the availability reported that the supervisors checked attendance of of these services 24/7 was higher in Highlands and staff or monitored staff performance. Momase regions than in Southern and NGI regions. People have lower accessibility to health care Access to Health Care Services facilities in Momase than in other regions of PNG (table 4.6). Regardless, travel times in all areas are Remoteness and other geographic variables within found to be very high, ranging from about 43 min- and across different provinces play an important utes in NGI to 88 minutes in Momase. Most of the role in determining access to health care services in patients sought care in the closest facilities from PNG. The population of PNG is distributed over a home. Walking remained as the main method of large geographic area; approximately 7 million peo- transportation from home to health facilities. ple are distributed over 462.8 thousand square kilo- Outreach activities are far below sufficient meters, and about 87 percent of the population lives in across regions; Momase shows the lowest rate of rural areas with limited access to transportation. The outreach activities. Table 4.7 shows the percentage need for health care, especially maternal-child health of facilities by region offering outreach patrols on a services and infectious disease control and treatment, has remained relatively high. This health facility sur- 28 vey was not designed to collect information on geo- Although the survey was not representative of the regions sur- graphic access to medical care services. Therefore, the veyed (sample was not selected by considering total number of facilities per region, small sample size, and not covering lower-level geographic information presented here is only to pro- facilities), the results are presented as indicative of access to services vide an indirect indication on geographic availability by the population living in the catchment areas of the facilities. Management, Access, Equity, Quality, and Safety of Health Service Delivery 47 Table 4.5  Types of 24/7 Services Available in Surveyed Health Facilities, by Region Regions in which the facilities are located Access measures Southern Highlands Momase NGI Total Facility open 24/7 94.44 100.00 82.35 100.00 94.52 Inpatient services 24/7 94.44 100.00 82.35 94.12 93.15 Emergency service 24/7 94.44 85.71 82.35 94.12 89.04 Outpatient clinic for adults 24/7 77.78 80.95 82.35 94.12 83.56 Outpatient clinic for children 24/7 77.78 76.19 76.47 88.24 79.45 Outpatient clinic for women 24/7 66.67 61.90 47.06 58.82 58.90 Lab and diagnostic services 24/7 5.56 47.62 47.06 29.41 32.88 Pharmacy services 24/7 5.56 61.90 41.18 29.41 35.62 Snakebite clinic services 24/7 27.78 9.52 41.18 52.94 31.51 Table 4.6  Patient Access to Health Care Facilities, by Region Indicators of access Southern Highlands Momase NGI Travel time to come to the facility in minutes 67.43 50.64 88.20 42.54 Percent of patients reporting that they came to this facility walking today 69.34 59.06 51.89 48.28 Percent of patients reporting that this is the closest facility from home 74.45 78.52 70.75 68.10 Travel cost incurred to come to facility in PNG kina 5.22 3.25 1.25 4.85 Total fee paid today for services received in PNG kina 74.87 1.19 2.05 87.65 Percent of patients reporting visiting this facility for the first time 20.44 22.15 17.92 12.07 Percent of patients reporting that there is no other facility nearby 44.03 45.69 58.62 44.11 Note: Patient exit interview information can be used to understand some of the access issues. For example, those who visited the health facilities were asked about travel distance and cost, waiting time, and their opinion about the accessibility of the facilities. It should be noted that the patients seeking care from a facility are likely to be located near the facility and, therefore, this will bias the results. However, if patient responses are presented by facility categories, they should provide some information on accessibility. Table 4.7  Health Service Coverage Provided in Catchment and Remote Areas, by Region Regions of PNG surveyed Coverage indicators Southern Highlands Momase NGI All areas Percent of facilities conducting outreach patrols 83.33 80.95 52.94 88.24 76.71 Number of people in catchment area 23,715 28,203 25,466 15,263 23,249 Outreach patrol planned 29.6 24.7 3.5 60.2 29.2 Outreach patrol conducted 15.1 14.4 1.8 45.2 18.9 Number of MCH patrols 1.7 1.7 1.3 0.2 1.3 Number of immunization patrols 1.2 2.4 0.8 2.3 1.7 Population in remote areas 11,044 7,449 5,432 6,355 7,679 Population in remote areas covered 5548 3,268 1,286 3,248 3,427 Percent of population living in remote areas 46.57 26.41 21.33 41.64 33.03 Percent of remote area population covered 50.24 43.87 23.67 51.10 44.63 Note: MCH = maternal-child health. 48 Service Delivery by Health Facilities in Papua New Guinea Table 4.8  Percentage of Health Facilities with Expired Drugs Present Safety-Related Indicators Level 3 and Level 3 and 4 public 4 church Level 5 to 7 Total Safety indicators are important in determining the Drug % (n) % (n) % (n) % (n) health outcomes of the population. For example, the Paracetamol/ 14.3 (4) 4.4 (1) 5.3 (1) 8.6 (6) presence of expired drugs and immunizations cou- panadol pled with the absence of specific cold-chain equip- Diclofenac 7.4 (2) 5.3 (1) 10.5 (2) 7.7 (5) ment or consumables is an important safety indicator. Artemisinin- 20.7 (6) 4.0 (1) 15.8 (3) 13.7 (10) Eight to 27 percent of all facilities were found to based have at least one expired drug present. Church-run combination facilities performed better than the public facilities. Fansidar 31.0 (9) 28.0 (7) 21.0 (4) 27.4 (20) Fansidar was the most common expired drug pres- TB Blister packs 7.1 (2) 13.0 (3) 6.3 (1) 9.0 (6) ent followed by ergometrine and artemisinin com- Oxytocin 7.1 (2) 12.0 (3) 5.3 (1) 8.3 (6) bination. In more than 10 percent of level 3 and 4 Ergometrine 10.3 (3) 28.0 (7) 15.8 (3) 17.8 (13) public facilities, seven out of the 10 drugs surveyed Ferrous sulfate 13.8 (4) 4.2 (1) 5.6 (1) 8.5 (6) had expired. Half of the drugs were expired in more Metronidazole 10.3 (3) 8.0 (2) 5.3 (1) 8.2 (6) than 10 percent of level 3 and 4 church-run and level Ciprofloxacin 10.7 (3) 9.1 (2) 11.1 (2) 10.3 (7) 5 to 7 public health facilities surveyed. In general, the Note: Selected drugs have been used in the table. The drugs are related to maternal conditions, malaria, and tuberculosis (TB). percentage of facilities with expired drugs present was relatively high for level 3 and 4 public facilities in comparison with level 3 and 4 church-run facili- regular basis. Overall, 77 percent of facilities provide ties. Table 4.8 shows the percentage of facilities with services through outreach. The facilities surveyed in expired drugs present. NGI and Highlands appear to be most active in offer- Another potential safety concern in PNG is the ing outreach antenatal care, while Momase region proper storage of vaccines and medicines that sup- facilities were least active. For the proportion of catch- port regular immunization services and the treat- ment population living in remote areas, about 47 per- ment of other chronic health conditions.29 Table 4.9 cent in Southern region and 42 percent in NGI lived in shows the percentage of facilities that offer immuni- remote areas. This proportion was lowest for Momase, zation services but did not have specific cold-chain which possibly can explain the relatively low participa- equipment. All level 5 to 7 facilities had vaccine tion in outreach patrols. Only about half of the remote refrigerators, while the unit was missing (or not func- area population in NGI and Southern regions was tional) for 7.4 percent of level 3 and 4 public facili- covered through outreach patrol activities. Clearly, a ties and 12 percent of level 3 and 4 church facilities. significant part of remote area populations was not A vaccine thermometer was not present in 11 per- covered by the health facilities in the area. cent of level 3 and 4 public facilities and 8 percent of level 3 and 4 church facilities. The presence of a Table 4.9  Percentage of Health Facilities Providing Immunization vaccine thermometer in the facility does not neces- Without Specific Cold-Chain Equipment sarily mean that temperature logs were kept. Almost Level 3 and Level 3 and all level 3 and 4 church facilities kept a temperature Cold-Chain 4 public 4 church Level 5 to 7 Total log (except one), but 17 percent of level 5 to 7 pub- Equipment % (n) % (n) % (n) % (n) lic facilities and 18 percent of level 3 and 4 public Vaccine refriger- 7.4 (2) 12.0 (3) 0.0 (0) 7.1 (5) facilities did not keep a temperature log. Most of the ator not present or functional cold boxes and vaccine carriers were not functional in Vaccine 11.1 (3) 8.0 (2) 0.0 (0) 7.1 (5) a small number of facilities surveyed. thermometer Facilities providing delivery services were not present assessed for readiness to provide adequate, safe No temperature 18.5 (5) 4.0 (1) 16.7 (3) 12.9 (9) log kept 29 Most cold boxes/ 3.7 (1) 12.0 (3) 0.0 (0) 5.7 (4) It is possible for a facility to provide safe immunization services vaccine carriers even without any cold-chain equipment if the vaccines are trans- not working ported from the cold storage facility located elsewhere on the vac- cination day. In the context of PNG, given the distances between Ice packs not 3.7 (1) 4.0 (1) 5.6 (1) 4.3 (3) health facilities, facilities should have basic cold-chain equipment available to ensure safety and efficacy of the immunization program. Management, Access, Equity, Quality, and Safety of Health Service Delivery 49 services and were found to generally present a Table 4.10  Percentage of Health Facilities Providing Obstetric serious safety risk (table 4.10). Depending on the First Aid, Basic Emergency Obstetric Care, and Comprehensive equipment, drugs, and personnel available, health Emergency Obstetric Care facilities providing maternal health and obstetric Level 3 Level 3 services can be categorized into three types: facilities Level of EmOC and and services provided 4 public 4 church Level 5 to 7 Total able to provide obstetric first aid, facilities able to by facilities % (n) % (n) % (n) % (n) provide basic emergency obstetric care (BEmOC), Obstetric first aid and facilities able to provide comprehensive emer- Have the capacity to 81.4 (22) 80.0 (20) 94.4 (17) 84.3 (59) gency obstetric care (CEmOC).30 About one-fifth of administer antibiotics, level 3 and 4 facilities were able to provide obstetric oxytocics, and anti­ services but were not equipped to provide even the convulsants basic first aid type of services, and about 40 percent BemOC of level 3 and 4 facilities were not equipped to pro- Have the capacity to 63.0 (17) 72.0 (18) 94.4 (17) 74.3 (52) vide BEmOC even though they provide obstetric do assisted vaginal delivery services. This situation is clearly a very significant Have the ability to use 63.0 (17) 60.0 (15) 88.9 (16) 68.6 (48) safety concern. Even among the level 5 to 7 facilities, three drugs plus do 10 percent of facilities lacked one or more of the vacuum and forceps necessary items to be considered a BEmOC provider. delivery When considering the standards of care, almost CEmOC all level 3 and 4 health facilities cannot be catego- Have the ability for 77.8 (21) 88.0 (22) 100.0 (18) 87.1 (61) rized as CEmOC, which underscores the need to manual removal of placenta strengthen EmOC services across PNG, given the Have the ability to 14.8 (4) 16.0 (4) 100.0 (18) 37.1 (26) relatively high level of fertility. Only about 15 per- manage caesarean cent of level 3 and 4 facilities had the capacity to man- section age a caesarean section or blood transfusion. Most of Have the capacity 14.8 (4) 20.0 (5) 94.4 (17) 37.1 (26) the level 5 to 7 facilities were able to perform these to do safe blood functions. The majority of the facilities at all levels transfusion reported the capacity to remove a placenta manually Have the capacity 11.1 (3) 12.0 (3) 83.3 (15) 30.0 (21) to use all drugs plus (78 percent for level 3 and 4 public health facilities, services listed above 88 percent for level 3 and 4 church-run facilities, and 100 percent of level 5 to 7 facilities). However, consid- ering all the functions required under CEmOC, only Another important safety aspect for health 11 to 12 percent of level 3 and 4 facilities are quali- facilities is the disposal of medical waste. Only a fied to safely provide this high-level obstetric service. small proportion of health facilities reported hav- Eighty-three percent of level 5 to 7 facilities may be ing a functional high-temperature incinerator, considered CEmOC providers. Since the level 5 to 7 one-chamber drum, or brick incinerator. Table 4.11 facilities are relatively high-level facilities, 17 percent reports the waste disposal equipment or system the of the facilities not able to provide CEmOC services facility had at the time of the survey. The majority should be considered an important safety concern. of level 3 and 4 facilities did not follow the system of segregating waste into colored waste bins. If the waste is not segregated, it is unlikely that it is disposed of 30 On the day of the survey, the interviewers asked the facility man- properly in accordance with waste disposal protocol. agers some specific questions to determine whether the facility was ready to provide obstetric first aid or BEmOC or CEmOC. If the The survey revealed the need for protocols and facility had the capacity to administer antibiotic injections, oxy- equipment to be used in the disposal of medical tocics, and anticonvulsants (availability of the required injections waste. Tables 4.12 and 4.13 report the waste dis- and presence of relevant personnel to administer the injections), the facility is considered as an obstetric first aid facility. BEmOC is posal approach followed by the facilities as reported defined by the ability to do all the functions required in an obstet- by facility managers. Table 4.12 shows the disposal ric first aid facility plus the capacity to do assisted vaginal delivery. practice for sharps. The standard practice for sharps CEmOC facilities are defined by all the functions required for disposal for level 3 to 7 facilities is burn and bury. being a BEmOC plus the ability of the facility for manual removal of placenta, ability to conduct caesarean section, and capacity to The recommended approach for sharps is to use an do safe blood transfusion. incinerator before placing the sharps in a pit. Burn 50 Service Delivery by Health Facilities in Papua New Guinea Table 4.11  Medical Waste Disposal Equipment and System Availability, by Level and Ownership Level 3 and 4 facilities Level 5 to 7 facilities Public facilities Church facilities (all public) Medical waste disposal equipment and system N % of Public N % of Church N % of Facilities High-temperature incinerator 3 10.34% 2 8% 9 47.36% One-chamber drum/brick incinerator 1 3.4% 6 24% 9 47.36% Has burial pit for use for disposal of wastes 27 93.10% 20 80% 8 42.10% Segregates medical waste into three colored waste bins 8 27.5% 9 36% 16 84.21% and bury is often used as an alternative for ensuring facilities, the most common approach is to dispose that biomedical waste is burned before burying. In of biomedical waste in covered or uncovered pits. some cases, facilities simply bury the sharps without The most common disposal method in level 3 and 4 burning them. Throwing the sharps in an open pit church-run health facilities is burn or bury, or both. (whether covered or not) without burning is not rec- Level 5 to 7 public facilities reported using an incin- ommended because of safety concerns. About 14 per- erator for biomedical waste disposal. cent of level 3 and 4 public facilities and 8 percent of level 3 and 4 church facilities reported disposing Patient Satisfaction sharps in covered or uncovered pits. The disposal of biomedical waste varies based Outpatient Services on health facility type and needs to be standardized In almost all categories of patient satisfaction in to remove safety risks. Table 4.13 reports the prac- outpatient services, level 3 and 4 church-run health tice of disposing biomedical waste by level of facility facilities performed better than public facilities. and administration. The modal approach of dispos- Church facilities performed better than level 5 and ing biomedical waste differs among the three facility level 6 facilities in all dimensions of overall satis- types considered here. In level 3 and 4 public health faction scores. The patients using outpatient clin- ics were interviewed on exiting the health facility to understand their personal evaluation on the quality Table 4.12  Health Facility Disposal Practice for Sharps of services received and their opinion about the facil- Level 3 Level 3 ity.31 According to users, all health facilities scored Waste disposal and and Level 5 to 7 practice for 4 public 4 church public Total poorly in relation to distance traveled to reach the sharps % (n) % (n) % (n % (n) health facility. Level 5 to 7 health facilities scored Incinerator 6.9 (2) 12.0 (3) 47.4 (9) 19.2 (14) relatively better than level 3 and 4 health facilities, Burn and/or bury 75.9 (22) 72.0 (18) 47.4 (9) 67.1 (49) scoring 65. All health facilities scored quite poorly Pit (covered or 13.8 (4) 8.0 (2) 0.0 (0) 8.2 (6) in health facility cleanliness. In reference to service uncovered) delivery and health care provider interaction, user Off-site 3.4 (1) 8.0 (2) 5.0 (1) 5.5 (4) satisfaction was greater than 0.5. The detailed table Total 100.0 (29) 100.0 (25) 100.0 (19) 100.0 (73) can be found in Annex B. The summary measures of satisfaction as well as a few other overall satisfaction Table 4.13  Health Facility Disposal Practice for Biomedical Waste measures are reported in table 4.14. Waste disposal Level 3 Level 3 practice for and and Level 5 to 7 Inpatient Services biomedical 4 public 4 church public Total waste % (n) % (n) % (n) % (n) Church facilities performed better than public Incinerator 0.0 (0) 4.0 (1) 42.1 (8) 12.3 (9) facilities in inpatient satisfaction scores. The survey Burn and/or bury 41.4 (12) 56.0 (14) 31.6 (6) 43.8 (32) asked patients to provide their experience on different Pit (covered or 44.8 (13) 24.0 (6) 5.3 (1) 27.4 (20) uncovered) Off-site 13.8 (4) 16.0 (4) 21.0 (4) 16.4 (12) 31 The scores are defined between 0 and 100; a value of 100 reflects Total 100.0 (29) 100.0 (25) 100.0 (19) 100.0 (73) the highest level of satisfaction. Management, Access, Equity, Quality, and Safety of Health Service Delivery 51 Table 4.14  Outpatient Clinic Satisfaction Scores Level 3 and Level 3 and Overall satisfaction measures 4 public 4 church Level 5 Level 6 Level 7 Average satisfaction score based on 14 specific indicatorsa 58.79 70.47 65.45 65.87 76.64 Overall satisfaction of the visit 56.66 74.63 66.99 53.70 83.50 Would prefer to return to the facility if medical care needed in 67.70 78.09 77.19 68.43 83.50 the future Will recommend the facility to friends and relatives 62.96 76.03 73.76 68.43 83.50 Note: Highest possible client satisfaction score = 100. a. Patient satisfaction questionnaire included 14 aspects of satisfaction and the numbers are average values of these 14 aspects.[AQ1] aspects of inpatient services, ranging from the admis- are very similar when the patient was asked if the sion process to services received at the health facility doctor explained the treatment that the patient was (table 4.15). In general, patients in level 5 to 7 facilities receiving. Privacy concerns also appear to be impor- thought that the admission process was difficult and tant for patients. The satisfaction scores with doctors the wait time was not reasonable. Again, the average respecting patient privacy were 62.1, 58.5, 67.8, 60.0, satisfaction scores related to admission or front office and 80.1 for level 3 and 4 public, level 3 and 4 church, interactions was higher for level 3 and 4 church-run level 5, level 6, and level 7, respectively. The satisfaction facilities than for other facility types. scores for doctors’ services are reported in table 4.16. Patients expressed lukewarm satisfaction with The average satisfaction scores on services pro- doctors’ nighttime availability, quality of inter- vided by doctors were not very high. Even though action, and respect for privacy, although scores questions related to behavior and skills of doctors were higher at higher-level facilities. The patients show higher satisfaction scores than others, the scores were asked about the doctors’ availability during are still quite low, implying that patients in inpatient the day and at night. In all the facility types, patient units were relatively unhappy with the quality of ser- satisfaction with the availability of doctors at night vices received from doctors. They clearly expected declined significantly compared with their availabil- better interactions with health care providers. Some ity during the day. With regard to availability of doc- of the satisfaction scores, such as maintaining patient tors, level 6 and 7 facilities were better than level 3 privacy, can be improved significantly through and 4 or level 5 health facilities. It appears that patients changes in few basic aspects of practice when physi- were not very satisfied with the quality of inter­ cally examining patients or when asking sensitive actions they had with doctors, especially in level 3 questions. The problem appears to be significantly and 4 public health facilities. For example, the aver- higher in level 3 and 4 public health facilities. Some age satisfaction score with a doctor’s explanation of of the satisfaction scores may have been affected by medical condition cause was only 52 for level 3 and 4 low availability of doctors in level 3 and 4 facilities, public health facilities compared with 58.5 for level 3 but the scores indicate that availability of doctors and 4 church-run health facilities and 68, 63, and 83 is not the only problem; the quality of interactions for levels 5, 6, and 7 facilities, respectively. The scores between doctors and patients also needs improving. Table 4.15  Inpatient Admission Process Satisfaction Scores Level 3 and 4 facilities Level 5 Level 6 Level 7 Satisfaction related to admission office/front office Public Church public public public Very easy to get admitted in the hospital 54.19 70.40 57.16 65.57 66.80 Waiting time for admission reasonable 45.75 72.42 53.84 54.37 63.50 Staff members in the admission office helpful and courteous 56.25 83.21 80.99 81.20 76.90 Admission office explained the admission process and rules/regulations 45.81 67.98 67.21 66.80 73.50 of the hospital clearly Overall mean score: admission related 50.50 73.50 64.80 66.98 70.18 Note: Highest possible client satisfaction score = 100. 52 Service Delivery by Health Facilities in Papua New Guinea Table 4.16  Inpatient Satisfaction Scores for Doctors’ Services Level 3 and 4 Satisfaction with different aspects of services facilities Level 5 Level 6 Level 7 provided by doctors Public Church public public public Satisfaction with doctor’s availability during daytime 62.1 68.1 75.8 65.7 83.5 Satisfaction with doctor’s availability at night 47.6 47.6 62.3 52.9 76.8 Satisfaction with care and concern shown by doctors 62.1 64.5 74.6 64.6 80.2 Satisfaction with doctor’s explanation of the cause of your medical condition 52.4 58.5 67.7 63.4 83.5 Satisfaction with the doctor’s explanation of your treatment 57.3 58.4 65.5 57.7 83.5 Satisfaction with doctors respecting your privacy (using curtains for 62.1 58.5 67.8 60.0 80.1 physical checkups, asked sensitive questions discretely, etc.) Satisfaction with the time spent by doctors treating you 52.4 63.2 66.3 57.6 80.1 Satisfaction with the number of times a doctor checks on you 47.6 60.9 67.0 56.5 76.8 Satisfaction with the politeness and respectfulness of the doctors 67.0 68.1 78.1 68.0 83.5 Satisfaction with the skills and abilities of the doctors and other health 67.0 65.6 75.8 65.7 83.5 care providers in this health facility Overall mean score: services provided by doctors 57.8 61.3 70.1 61.2 81.2 Note: Best possible score = 100. Patients are generally less satisfied with the during the day and night shifts. This change will availability of nurses and other clinical personnel. improve the quality of interactions between clinical Patient satisfaction was also evaluated on the avail- staff and patients and ensure that they remain sensi- ability of other clinical personnel and the service qual- tive to the needs of patients. ity rendered in the hospital. The satisfaction scores In general, satisfaction levels associated with on the availability of nurses and other personnel at various aspects of housekeeping services and food night were quite similar to their availability during services were very low. The indicator that scored the the day for all facility types shown in table 4.17. One worst in patient satisfaction was associated with clean- patient satisfaction indicator, timeliness of bringing liness of bathroom and toilets. Most of level 3 and 4 medication to patients, placed particular focus on the health facilities did not provide food service and so the expected role of nurses and support personnel during averages at level 3 and 4 facilities are based on three inpatient stay. Even for this specific function, patient to four facilities only. Only four of the level 3 and 4 satisfaction was quite low for all types of health facil- church-run health facilities provided food services, ity. It appears that the shortfall in patient satisfaction, and patients were relatively more satisfied with their caused by the limited availability of doctors, was not quality than in other facility types. Satisfaction scores compensated by the availability and quality of ser- for level 5 and 6 health facilities were also quite low; vices provided by the nurses and other clinical sup- the satisfaction rating for food quality was slightly port staff. To improve service quality, it is important higher at the level 7 facility. to strengthen the role of nurses and other clinical The overall index of patient quality perception personnel involved in the provision of services, espe- for inpatient and outpatient services was found cially for level 3 and 4 health facilities. Overall patient to be quite low, about 57.1 for level 3 and 4 public satisfaction in all types of health facilities can be facilities, 64.7 for level 3 and 4 church-run facilities, improved by increasing the availability of clinicians and 69.9 for the level 7 facility. Patients were asked Table 4.17  Inpatient Satisfaction Scores for Nurse and Support Staff, Housekeeping, and Food Services Services provided by nurses and Level 3 and 4 facilities Level 5 Level 6 Level 7 other personnel Public Church public public public Service provided by nurse/support staff 50.7 62.0 69.2 62.7 72.6 Satisfaction with housekeeping services 33.1 46.0 58.0 52.0 79.5 Satisfaction with food service 30.0 44.1 43.9 35.6 59.3 Note: Best possible score = 100. Management, Access, Equity, Quality, and Safety of Health Service Delivery 53 Table 4.18  Inpatient Perceptions of Health Facility Quality and Security Level 3 and 4 Patient’s range of agreement with various aspects of services facilities Level 5 Level 6 Level 7 at health care facility (inpatients) Public Church public public public Fighting and trouble in this area does not make it difficult for people to 58.5 62.6 64.4 63.5 70.2 use available health services. Health care providers are extremely thorough and careful. 60.2 65.5 66.6 64.6 76.8 Patient trusts in the skills and abilities of health workers. 71.0 68.1 66.6 62.4 73.5 Patients completely trust health care provider’s decisions about medical 49.0 58.5 63.5 63.5 70.2 treatments. Health care providers are very friendly and approachable. 45.8 54.6 67.7 62.3 66.8 Health care providers are available 24 hours a day. 50.1 59.4 65.5 64.6 66.9 Health care providers care about health just as much or more than the 79.4 75.6 74.0 66.8 66.9 patients do. Patient always try to follow the advice of health care providers. 45.9 50.4 56.7 51.2 63.6 Health care provider offers highest quality of medical care. 39.6 69.1 60.3 56.8 66.8 Doctors treat poor and rich patients the same. 79.3 79.9 76.2 71.3 70.2 Patients should always follow the advice given to them at health facilities. 47.9 67.6 64.3 70.1 76.9 Overall mean score: perception of inpatients 57.1 64.7 65.9 63.3 69.9 a series of questions to understand the contextual church-run health facilities were 0.142 and –0.073, and clinical factors that affect overall perception respectively. At level 4, the wealth scores for both with the quality of inpatient and outpatient services church and public facilities were almost identical (0.00 (table 4.18). The answers are subjective, but to some and –0.04). The average wealth score of households degree reflect patients’ confidence of the services tends to increase with the level of facility. The average delivered in various health facilities. scores were 0.899, 1.487, and 2.312 for levels 5, 6, and 7 health facilities, respectively. It appears that relatively better-off households are more likely to use upper- Equity-Related Indicators level facilities than level 3 and 4 facilities. One of the outcome measures proposed by the World Health Organization building block approach is Summary equity in access and fairness in financing. A patient exit survey, containing several income and asset own- Health Facility Autonomy: Expenses, ership variables, was used to construct a proxy socio- Personnel, and Services economic index. Level 5 to 7 health facilities have relatively more Access to health facilities is skewed by income and autonomy than level 3 and 4 facilities. In general, wealth status. In the survey, patients at exit and at the facility (inpatients) were asked to indicate their income church facilities reported a higher degree of auton- status compared with others in the locality. The omy than level 3 and 4 public facilities. Health facility respondents could choose one of the three options: autonomy is determined by the independent control lower income, middle income, or high income. The lower-income segment of the population has less Figure 4.19  Percentage of Clients from Low-Income Households, access to higher-level facilities (figure 4.19). Only by Facility Level 10 percent of users of level 7 facilities are from the low- Level 7 10 income group, while more than 70 percent of users of level 3 health facilities are from the low-income group. Level 6 30 Church health facilities serve more poor people Level 5 42 than their public counterparts. Patients seeking care Level 4 56 from church-run facilities were, on average, poorer Level 3 73 than the patients seeking care from level 3 public facil- ities. The average wealth scores for level 3 public and 0 10 20 30 40 50 60 70 80 54 Service Delivery by Health Facilities in Papua New Guinea that the management has over health facility expenses, Patient Satisfaction staff, and the services offered. The survey also asked the facility managers whether they were happy with Patient satisfaction was similar across the facility the level of autonomy that they had and 89 percent of types, although level 3 and 4 church-run facilities level 5 to 7 health facilities, 84 percent of level 3 and 4 and the level 7 facility performed better than others. church-run facilities, and 55 percent of level 3 and 4 Patient satisfaction was evaluated using a wide range public facilities reported their satisfaction. of factors, such as ease of access, cost, cleanliness, and interactions with the health facility staff. Level 3 and 4 public facilities show the lowest average satisfaction Health Facility Safety rates, followed by level 5 and 6 facilities. In one spe- Health facility safety remains a major concern in cific dimension, the cost of obtaining services, level 3 PNG. This study measured a number of safety indica- and 4 public facilities performed better than others. tors using the facility survey data. Most of level 3 and 4 facilities did not have a blood bank. The presence Equity of expired drugs in more than 10 percent of facili- ties may be considered a significant problem. Few The survey found that poorer households are more facilities did not have all the equipment and supplies likely to seek medical care from level 3 and 4 health needed for the child vaccination program. The avail- facilities than other facility types. The fraction of ability of EmOC was relatively low in level 3 and 4 treatment seekers that considered themselves poor health facilities. drastically declines with the increasing level of the Nonavailability of CEmOC in most level 3 and 4 health care facility. The percentage of treatment seek- facilities is especially concerning. Given that PNG ers who are “poor” was 73 percent at level 3 health has a high fertility rate, emergency obstetric service facilities, whereas it was 10 percent at the level 7 facil- is essential for improving maternal health services. ity. If wealth scores are used (based on the ownership For safety indicators, level 5 to 7 health facilities of assets and access to various amenities of life), aver- were found to be much better than other facility age wealth score was lowest for the users of level 3 types. Policy makers should carefully evaluate how to and 4 church-run facilities and the wealth score strengthen the availability and provision of EmOC. increased with increasing facility level. Chapter 5 COST AND OUTPUT OF FACILITIES This chapter presents an analysis of the outputs survey asked the facility respondents to report the produced by health facilities and the cost of produc- output produced in different months in the previ- tion. In a health care system, output is often repre- ous year. Information from 4 months, distributed sented by the final health outcomes of the population, over 1 year, was collected to allow for seasonal vari- but health facility surveys do not provide information ability of output. Each month is assumed to be the on health outcomes. Health facilities are production representative of the output produced in the specific units, which produce a number of services that are quarter, and annual output levels were estimated important in improving the health and well-being from the information on the output produced in 4 of the population. As with other production units, months. Tables 5.1 to 5.3 report health facility output health facilities produce various health and medi- by categories. The tables represent annual output esti- cal care services by using labor and capital. Analyses mates by multiplying the monthly output by 12. The of cost and output provide information not only on quantities of services delivered are reported by facility how the inputs are converted into outputs but also category levels 3 and 4, levels 5 and 6, and level 7. on the efficiency of the health facilities in converting On average, the church health facilities receive inputs into outputs. more antenatal (ANC) visits and well-care visits. Table 5.1 illustrates the annual estimated output for maternal and child health services. Note that the num- Outputs of Health Facilities: Quantities ber of deliveries conducted by the level 7 health facil- of Various Services Produced ity was not reported. The child vaccination numbers The most common output of a health facility is the refer to number of doses of vaccines administered general outpatient consultations or patient visits. rather than number of children immunized. In some In the facility survey, information was collected on instances, level 3 and 4 public sector facilities treated quantities of outputs produced through two differ- more patients compared with level 3 and 4 church-run ent approaches. The first approach is recording the facilities. For example, level 3 and 4 public health facili- information from the most recent monthly report. ties administered more tetanus toxoid (TT), measles Health facilities prepare and submit monthly reports vaccines, and vitamin A supplements, and distributed to the National Department of Health (NDOH).32 more family planning (FP) pills, intrauterine devices The second source of output data is the information (IUDs) and loops, and condoms. It is interesting to note collected directly from health facility managers. The that the number of bacillus Calmette-Guérin (BCG) vaccines administered in all the facility types exceeds the number of deliveries by a wide margin. Since BCG 32 Since the survey obtained information on the output produced vaccine is the first dose administered, the expectation in 1 month, annual estimates of output can be derived by assum- is that the number of BCG doses administered is close ing that the monthly report from which data were obtained is rep- resentative of other months and can be used to estimate annual to the number of baby deliveries. However, this gap outputs simply by multiplying monthly outputs by 12. may be explained by the number of babies who are 55 56 Service Delivery by Health Facilities in Papua New Guinea Table 5.1  Annual Average Maternal and Child Health Services Delivered by Facilities Prenatal care, delivery, vaccination, Level 3 and 4 Level 5 and 6 Level 7 and family planning Public Church Public Public Number of ANC visits 897.9 1214.9 1,485.2 3,960.0 Number of TT doses delivered 596.3 297.6 583.1 4,044.0 Number of facility deliveries 146.1 224.6 1,968.7 NA Number of complicated deliveries 14.9 19.7 259.1 NA Number of deliveries in community 5.0 6.7 8.5 0.0 Number of well-baby visits 2,073.1 3347.5 3,667.1 20,280.0 Total BCG vaccines delivered 311.2 377.3 656.5 12,276.0 Total hepatitis B delivered 164.7 194.4 1,217.6 13,152.0 Total DTP/Hib doses delivered 885.5 1123.7 1,635.5 8,256.0 Total polio doses delivered 1,025.8 1225.4 1,665.9 8,652.0 Total measles doses delivered 999.3 865.9 799.1 3,756.0 Total vitamin A delivered 677.0 581.8 555.5 2,952.0 Number of breast-feeding counseling sessions 59.2 63.8 139.3 180.0 Number of cases receiving FP pills 149.0 125.3 194.7 1,260.0 Number of cases receiving FP injection 385.7 529.0 398.7 6,600.0 Number receiving condoms 50.5 19.2 130.0 24.0 Visits for IUD or loops 25.7 0.0 20.7 180.0 Permanent family planning methods 24.4 1.0 150.0 0.0 delivered outside health facilities and are subsequently also common at level 3 to 6 facilities. The pattern of brought to a facility for the BCG vaccine. outpatient visits in level 3 and 4 church-run facili- The most common reasons for seeking treat- ties was similar to level 3 and 4 public facilities. For ment were respiratory illness, diarrhea, malaria, level 5 and 6 facilities, most patients sought treat­ skin diseases, and accidents and injuries (table 5.2). ment for respiratory illness, malaria, diarrhea, Respiratory infections, both upper and lower respi- accident and injury, skin disease, ear infection, and ratory infections, appear to be the most common sexually transmitted infections (STIs). The distribu- condition for which health care is sought (table 5.2). tion of outpatient visits by facility type implies that The number of yearly visits for respiratory reasons patients suffering from malnutrition and anemia are var­ied from 2,971 for level 3 and 4 public sector facil- treated at level 5 to 7 facilities rather than at level 3 ities to 8,805 for level 5 and 6 facilities. Malaria was and 4 facilities. Table 5.2  Annual Average Outpatient Visits by Facility Type Level 3 and 4 facilities Level 5 and 6 Level 7 Outpatient visits by visit type and category Public Church Public Public Outpatient measles cases 43.4 26.4 138.4 24.0 Outpatient respiratory infection cases 2,970.6 3,475.2 8,805.2 5,796.0 Outpatient diarrhea cases 706.8 510.7 3,728.5 2,100.0 Outpatient malaria cases 1,089.1 804.5 4,550.8 900.0 Outpatient fever cases 194.5 97.4 379.8 492.0 Outpatient anemia cases 51.7 37.9 222.4 780.0 Outpatient malnutrition cases 6.2 13.9 94.6 444.0 Outpatient accident and injury cases 407.6 476.2 1,885.4 2,460.0 Outpatient STI cases 193.2 109.9 762.4 1,044.0 Outpatient tuberculosis cases 51.3 78.7 467.3 1,368.0 Outpatient leprosy cases 0.0 8.2 4.2 24.0 Outpatient yaws cases 86.1 152.2 93.9 0.0 Other skin disease cases 837.9 966.7 1,162.6 576.0 Ear infection cases 137.4 282.2 1,000.9 288.0 Eye infection cases 296.7 252.0 285.2 276.0 Other cases and visits 4,315.9 3,804.0 8,697.9 13,560.0 Total outpatient visits for diseases 11,388.4 11,096.2 32,279.3 30,132.0 Note: Based on monthly report data and estimated from most recent 1-month information. Cost and Output of Facilities 57 Table 5.3  Annual Average Discharges by Facility Category Level 3 and 4 Level 5 and 6 Level 7 Inpatient discharges on the average per year Public Church Public Public Discharges related to vaccine-preventable childhood illnesses 78.6 36.5 224.7 60.0 Respiratory illnesses of adults 237.1 87.8 390.7 60.0 Diarrhea among children 6.6 13.9 105.3 264.0 Malaria cases of children 3.3 2.9 70.7 492.0 Malaria cases of adults 237.1 87.8 390.7 60.0 Cases of anemia 6.6 13.9 105.3 264.0 Child malnutrition cases 3.3 2.9 70.7 492.0 Cases related to accidents and injury 27.7 35.5 660.7 768.0 TB cases 3.7 15.8 192.0 1,200.0 Typhoid cases 3.7 5.3 76.7 300.0 Leprosy cases 1.2 0.0 12.7 0.0 Child meningitis 0.8 0.5 48.0 204.0 Snakebite cases 2.5 2.4 16.7 180.0 Skin disease cases 11.6 14.9 184.7 0.0 HIV/AIDS children 0.0 0.0 25.3 120.0 HIV/AIDS adults 0.4 5.3 306.7 144.0 Heart disease 0.0 1.0 22.7 144.0 Cancer 0.8 2.9 97.3 336.0 Hypertension 0.8 3.4 36.0 372.0 Diabetes 0.0 0.5 32.7 264.0 Other discharges 245.8 263.5 2856.7 6912.0 Total discharges 871.9 596.6 5,926.7 12,636.0 Note: Estimated from most recent monthly report of hospitals. The patient mix in hospitals becomes increas- and 6 facilities and 23 percent for the level 7 facility. ingly more severe and complex with the level of Conversely, noncommunicable disease-related con- health facility. At level 3 and 4 facilities, the patient ditions were relatively small in lower-level facilities. mix is dominated by infectious diseases, while the If heart disease, cancer, hypertension, and diabe- importance of noncommunicable diseases increases tes are considered as representative of chronic and significantly at the higher-level health facilities. noncommunicable diseases, these represent close to Table 5.3 reports the number of inpatient discharges 0 percent among all discharges in level 3 and 4 pub- in a year, estimated from the most recent month’s data lic facilities, 2 percent for level 3 and 4 church-run on hospital discharges.33 Table 5.3 also indicates that facilities, 6 percent for level 5 and 6 facilities, and discharges related to infectious and poverty-related 19 percent for the level 7 facility. diseases, such as vaccine-preventable childhood ill- The general outpatient consultations or visits nesses, respiratory illness, diarrhea, malaria, tubercu- are similar for level 3 and 4 public facilities and losis (TB), child malnutrition, typhoid, and leprosy, church-run facilities. Table 5.4 reports the average account for about 66 percent of cases in level 3 and annual output derived from information on facility 4 public sector health facilities and 42 percent of output produced in January, April, July, and October level 3 and 4 church-run facilities. The percentage of 2014 (or closest month).34 Note that the annual con- total discharges belonging to infectious and poverty- sultation estimates taken from monthly reports were related medical conditions was 26 percent for level 5 lower than these numbers by about 30 percent. Other 33 Although using monthly data to obtain estimates for annual num- 34 ber of discharges may be considered problematic, mainly due to In the costing questionnaire, facility respondents were asked to seasonality of medical care utilization, the information is useful provide information on facility outputs in 4 different months of in better understanding the medical conditions for which patients 2014. In the analysis of output-based efficiency, the output reported get admitted to hospitals in PNG. Since the data are presented by by facility managers and respondents was used. The monthly report facility category, it is expected that the infectious diseases and com- is a part of the National Health Information System (NHIS) and mon medical conditions present quite commonly at level 3 and may be biased because of the inclusion of outputs produced in 4 facilities compared with level 5 to 7 facilities. Higher-level facili- aide posts. Moreover, it was considered important to collect the ties are more likely to see chronic conditions and other non­ information directly from the health facility rather than using the communicable conditions than lower-level facilities. information from the official reporting form. 58 Service Delivery by Health Facilities in Papua New Guinea Table 5.4  Average Annual Health Facility Outputs, by Facility Type Level 3 and 4 facilities Level 5 Level 6 Level 5 to 7 Health facility output per facility Public Church Public Public Public Annual number of outpatient visits and consultations per facility (average) General outpatient consultations 14,993 14,047 48,133 38,059 48,261 Dental visits 9 10 1,363 2,306 2,018 Ophthalmology outpatient visits 0 1 690 117 1,204 Psychiatry outpatient visits 0 0 24 191 53 Number of users of laboratory services 1,648 1,826 32,412 56,148 52,877 Other outpatient visits 12,119 4,107 17,059 38,368 20,394 Total outpatient (excluding lab test) 28,769 19,991 99,681 135,189 124,807 Average annual number of inpatient days by ward Medicine ward 411 612 4,735 9,508 8,492 Surgical ward 73 124 4,458 8,557 6,630 Ophthalmology ward 0 0 450 1,774 635 Obstetrics and gynecology ward 260 598 4,686 7,411 5,683 Pediatric ward 246 479 4,497 6,690 6,017 Other inpatients 880 1,997 6,266 12,142 11,020 Total inpatient days 1,871 3,809 25,091 46,082 38,477 Note: Based on the output of 4 months in 2014, as reported by facility managers. outpatient visit categories were also higher when 2014.35 The survey collected information on child using the numbers based on the 4-month average immunization activities. Table 5.5 shows the num- rather than using the most recent 1-month report. ber of children immunized by the facilities, either The total number of inpatient days in church- from the facility or through outreach activities. Note run health facilities was almost double that of pub- that the numbers reflect children immunized, not lic health facilities (3,800 vs. 1,800). This number is the number of vaccine doses delivered. The percent- possibly due to a greater number of beds at surveyed age of facilities providing services through outreach church-run facilities. The average number of beds in patrols was 83 percent for level 3 and 4 public sec- level 3 and 4 public facilities was 21.7 as compared tor facilities, 92 percent for level 3 and 4 church-run with 37.7 in level 3 and 4 church-run facilities. For all facilities, and about 50 percent for level 5 to 7 facili- facility types, the medical ward was the most impor- ties. In 2014, the average number of outreach patrols tant ward for inpatients, followed by the obstetrics and planned was 18, 51, and 18 for level 3 and 4 public, gynecology (OBGYN) ward (excluding the category level 3 and 4 church-run, and level 5 and 6 facilities, “other inpatients”). However, the number of OBGYN respectively. The actual numbers of patrols organized inpatient days in public facilities (levels 3 and 4) in the year were lower than planned for all facility was relatively small compared with the number of types. In 2014, the actual number of outreach patrols OBGYN days in church-run facilities. In church-run organized was 4.9 for level 3 and 4 public facilities, facilities (levels 3 and 4), the OBGYN ward was 38.7 for level 3 and 4 church-run facilities, and 14.1 as important as the medical ward for inpatient days. for level 5 and 6 facilities. The level 7 facility did not The pediatric ward was also very important in organize any outreach patrols. Lack of funding was level 3 and 4 facilities for outputs produced. Similarly, reported as the main reason that facilities were not in level 5 to 7 facilities, the medical ward was the able to conduct the planned number of outreach most important inpatient category followed by the activities (76 percent). Other reasons included lack surgical and pediatric wards. The OBGYN ward was also important in relation to inpatient days produced 35 In general, outreach activities, on their own, may not be con- at level 5 to 7 facilities. Ophthalmology inpatients, as sidered the output of the health facility. The number of outreach expected, were not present in level 3 and 4 facilities. patrols organized helps increase health facility coverage. In some Figure 5.1 shows the proportions of inpatients in dif- instances, outreach patrols in remote, low-density areas do not sig- nificantly increase the production of services, but these patrols may ferent categories of facilities. serve an important social purpose. In that sense, the number of out- Church-run facilities organized twice as many reach patrols organized can be considered as output, in addition to outreach patrols as the public health facilities in the specific services the facility produces through outreach. Cost and Output of Facilities 59 Figure 5.1  Percent of Inpatients in Health Facility Wards, by Facility Type Level 3-4 Public Level 3-4 Church Medicine Surgical Medicine ward ward ward 16% 3% 22% Other Surgical inpatients ward OBGYN Other 47% 4% ward inpatients OBGYN 52% 16% ward 14% Pediatric Pediatric ward ward 13% 13% Level 5 Level 5-7 Medicine Medicine Other Other ward ward inpatients inpatients 19% 22% 25% 29% Surgical ward Surgical Pediatric 18% ward ward Pediatric 17% 18% OBGYN ward OBGYN ward 15% ward 18% 15% Ophthalmology Ophthalmology ward 2% ward 2% of fuel (9 percent) and staff constraints to organize contractors and deliver health care where the gov- outreach. ernment cannot adequately provide health services. For the purposes of this study, the only aspect of Budget, Expenditure, and Financing facility funding that was considered relevant was of Health Facilities the collection of user fees at the facility level. The estimation shows that the cost of running a The facilities in PNG are funded, to a large extent, level 3 and 4 public sector facility is about K 3.2 mil- by the public sector. This is also true for church- lion per year, while the expenditure per year was run facilities, which can be viewed as government K 2.3 million for a level 3 and 4 church-run Table 5.5  Estimated Child Vaccinations Delivered, by Facility and Outreach Activities Level 3 and 4 Number of children immunized in a year facilities Level 5 Level 6 Level 5 to 7 (based on 3 month’s data) Public Church Public Public Public Total number of children vaccinated in 1 year (3 months’ coverage × 4) 1,352 1,774 3,392 3,152 3,458 Total vaccinated through outreach activities in 1 year  345  790  742 1,024  792 Percent of vaccinations given through outreach 26% 45% 22% 33% 23% Note: Data estimated from previous 3 months’ information. 60 Service Delivery by Health Facilities in Papua New Guinea Figure 5.2  Average Cost of Running a Facility in PNG A comparison of expenditures for level 3 and 4 (in Millions of PNG Kina [PGK]), 2014 facilities highlights some significantly higher costs for the church-run facilities compared with public facili- 90.00 88.51 ties. For example, transport and fuel expenditure was 80.00 reported at about K 349,300 for church-run facilities, 70.00 but it was only K 11,500 for public facilities. Similarly, 60.00 Millions of PGK other operational expenses, rentals of property, consul- 50.00 tancy fees, and capital expenses were also significantly 40.00 higher for church-run facilities. Conversely, travel and 30.00 subsistence benefits are significantly higher for public 16.51 20.00 facilities compared with church-run facilities (K 83,000 10.00 3.17 2.25 compared with K 11,000). An analysis of cost function 0.00 relating to various activities and services provided may Level 3–4 Level 3–4 Level 5–6 Level 7 Public Church be able to reveal underlying reasons for the cost differ- ences across facilities. Anecdotally, one of the reasons appears to be the significantly higher level of outreach facility, K 16.5 million for a level 5 or 6 facility, and patrols organized by the church-run facilities compared K 88.5 million for the level 7 facility (table 5.6 ).36 with public sector health facilities at levels 3 and 4. Excluding the capital expenditures, total expendi- In all facility categories, personnel costs account ture per facility remains, on average, K 3.2 million for about 62 percent to 67 percent of total recurrent for level 3 and 4 public health facilities (capital expen- diture is considerably lower compared with other expenses (not including capital items) (figure 5.2). health facilities). In comparison, the total expenditure Operational expenses were also very similar in rela- excluding capital expenditure, per health facility is tive terms across different health facility categories. K 1.6 million for level 3 and 4 church-run health facili- Drugs supplied through the push system appear to ties, and K 15.8 million and K 57.7 million, on aver- represent only a small proportion of total recurrent age, for level 5 and 6 facilities and the level 7 facility, expenses. In church-run facilities, drugs account for respectively. Table 5.6 presents only a few aggregated 4 percent of recurrent expenses, while it was in between expenditure categories. For details on each of the 1 and 2 percent in other facility categories. Better avail- categories, the full expenditure table is reproduced in ability of drugs in church facilities may have allowed Annex D (table D.1).37 it to offer better quality services to patients. 36 These figures are reported actual expenditures, not the budgeted Budget and Expenditure Comparisons amount. Although the facility managers reported the annual expen- Level 3 and 4 health facilities are severely under- ditures of health facilities surveyed, it is not clear if the expenditures actually include all the cost items. For example, salary and fringe ben- funded. This section provides information on health efits are paid by the Department of Health and few other resources facility budgets and their actual expenditures. Table 5.7 used at the facility level are not procured or purchased by the facilities. compares the budget and expenditure numbers for If the inputs are not purchased by the facilities, managers are unlikely to be able to report the expenses incurred. Drug delivery, by the push level 3 and 4 facilities. Only the aggregated budget system in which boxes of drugs are procured at a central location and expenditure categories are used here. Actual and delivered to the facilities, may not be reflected as an expenditure expenditure in level 3 and 4 public facilities was only category in a facility manager’s budget. Since the budget and expendi- ture information includes salary and fringe benefits, capital expenses, 28 percent of the reported budget amount, while utilities, and so forth, it is likely that many of the resources used at actual expenditure was about 52 percent for level 3 the facility level are included in the budget and expenditure numbers. and 4 church-run facilities. The gap between budgeted The budget and expenditure numbers are often finalized at the higher level of administration and therefore the value of most of the items amount and actual expenditure does not necessarily should be in included in the budget-expenditure data. imply that the facilities could not use the budget allo- 37 Since the budget and expenditure data do not show drug expen- cated. When the facilities lack the authority of deciding ditures as a separate category, it is assumed that the drug costs are not included in the budget or expenditure numbers. Using the on expenditures, it is more likely that they have little number of boxes of drugs received by facilities in a year, cost of or no control either on the budget or on the expen- drugs supplied to each and every facility was derived. The deri- diture levels. From the table, however, it is not clear vation of drug costs required information on contents of differ- ent types of boxes and then valuing the boxes using procurement why the actual fringe benefit payments were so low prices of the drugs and supplies. compared to the budget for this category, less than 9% Cost and Output of Facilities 61 Table 5.6  Average Facility-Level Costs, by Cost Categories (PNG Kina) Aggregate expenditure Level 3 and Level 3 and Level 5 and categories 4 public 4 church 6 public Level 7 Total salary expenditure 2,024,717 999,458 9,857,372 35,318,243 Total benefit expenditure 83,376 15,607 801,160 2,311,596 Utility, rent, and maintenance 17,699 99,037 1,786,750 16,124,275 Operational and office expenses 962,809 436,765 1,844,845 3,140,609 Capital expenditure 4,080 602,939 705,379 30,788,303 Training 5,000 2,604 156,604 186,420 Drug expenses 51,472 71,041 134,250 641,615 Other expenditure 23,043 22,660 1,221,250 0 Total 3,172,196 2,250,111 16,507,610 88,511,061 of the budgeted amount. Other items that show low User Fees and Revenue Generation use of budget were capital expenditures and training expenses for level 3 and 4 public sector facilities. Despite the introduction of the free primary health Upper-level facility (levels 5 to 7) budgets and care and subsidized secondary health care policy in expenditure numbers are more closely aligned 2014, 28.6 percent of level 3 and 4 public facilities, than for lower-level facility budgets (table 5.8). For 60 percent of level 3 and 4 church-run facilities, and example, salary expenditures were 96 percent of the 79 percent of level 5 to 7 facilities reported charging budget for level 5 and 6 facilities and 100 percent for user fees. Determinants of user fees differ by health the level 7 facility. Training expenditures were higher facility level. Most level 3 and 4 facilities reported that the facility decided the level of user fees. For than the budgeted numbers for both of these facility level 5 to 7 facilities, most indicated that the NDOH types. Administrative expenses were also higher than determines the fee. the amounts budgeted. The total user fees received at level 5 to 7 health There is no statistically significant relationship facilities far exceed those at level 3 and 4 health of total drug costs with the number of inpatient and facilities, driven by the volume of inpatient and outpatients seen in the facility for level 3 and 4 public outpatient visits. When comparing total user fees and church-run facilities. The relationship was weak and patient numbers, level 3 and 4 government-run for level 5 to 7 facilities with inpatient and outpatient health facilities collect K 1,696 on average per year, see numbers explaining about 26 percent variability of an average 28,769 outpatients, and admit an average drug expenses. The lack of relationship between out- 1,871 inpatients per year. In comparison, level 3 and puts produced in the facility and drug expenses prob- 4 church-run facilities collect K 19, 084, on average ably reflects issues with drug supply and availability. per year, see 19,991 outpatients, and admit an average 3,809 inpatients per year. The bulk of the total user Figure 5.2  Percent Distribution of Recurrent Health fees from level 3 and 4 health facilities come from reg- Facility Expenses, by Facility Type istration or consultation fees. On average, clinicians at level 5 to 7 health facilities see 119,892 outpatients 100 1 1 1 2 4 9 and admit 36,550 inpatients per year. Further, average 1 total user fees are K 183,169 per year. Although the 80 31 23 33 33 collection for consultation fees is considerably higher at level 5 to 7 health facilities compared with level 3 60 and 4 health facilities, the bulk of the user fees are collected for laboratory tests rather than consulta- 40 67 67 65 tion fees. 62 The average registration fee in the facilities were 20 K 1.60, K 2.70, and K 6.00 for level 3 and 4 pub- lic, level 3 and 4 church-run, and level 5 to 7 public 0 Levels 3-4 Levels 3-4 Levels 5-6 Level 7 facilities, respectively. Public facilities at levels 3 and Public Church Public 4 did not charge a separate user fee to see a gen- Salary and benefits Drugs eral health care provider. Only two level 3 and 4 Operational expenses Other church-run facilities charged K 10 per visit. Out 62 Service Delivery by Health Facilities in Papua New Guinea Table 5.7  Budget and Expenditure Comparison for Level 3 and 4 Health Facilities, PNG Kina Budget and Level 3 and 4 public facilities Level 3 and 4 church facilities expenditure categories Budget (B) Expenditure (E) E * 100/B (%) Budget (B) Expenditure (E) E * 100/B (%) Salary 3,071,113 2,024,716 65.93 1,041,852 999,459 95.93 Fringe benefits 960,193 83,736 8.72 181,998 15,607 8.58 Administration 5,940,749 980,508 16.50 1,562,950 535,802 34.28 Capital expenditure 1,054,693 4,081 0.39 1,445,108 602,940 41.72 Training expenses 263,537 5,000 1.90 2,958 2,604 88.03 Other expenses 27,002 23,043 85.34 123,760 22,660 18.31 Drug cost 0 51,472 0 71,041 Total 11,317,287 3,172,556 28.03 4,358,626 2,250,113 51.62 Table 5.8  Budget and Expenditure Comparison for Level 5 and 6 and Level 7 Health Facilities, PNG Kina Budget and Level 5 and 6 public facilities Level 7 public facility expenditure categories Budget (B) Expenditure (E) E * 100/B (%) Budget (B) Expenditure (E) E * 100/B (%) Salary 10,312,403 9,857,371 95.59 35,251,000 35,318,243 100.19 Fringe benefits 1,087,834 801,160 73.65 2,074,000 2,311,596 111.46 Administration 3,404,201 3,631,595 106.68 17,817,000 19,264,884 108.13 Capital expenditure 992,376 705,379 71.08 31,240,000 30,788,303 98.55 Training expenses 86,725 156,604 180.58 100,000 186,420 186.42 Other expenses 2,133,314 1,221,250 57.25 1,000,000 0 0.00 Drug cost 0 134,250 0 641,615 Total 18,016,853 16,507,609 91.62 87,482,000 88,511,061 101.18 of nine level 5 to 7 facilities that charged user fees use the user fee revenue to improve infrastructure for general consultation, five facilities charged K 20 and to buy drugs. The funds are also reportedly used per consultation and the remaining charged about to buy nonmedical supplies and pay for things such K 5 to K 6. When asked about user fees for seeing a as casual labor, facility administrative expenses, and specialist, 10 level 5 to 7 facilities reported charging food services. user fees and almost all charged K 20 per visit. Only Only four church-run facilities reported that two level 3 and 4 church-run facilities charged user they received voluntary contributions or donations fees for specialist consultation and the average fee was K 20. Table 5.9 shows estimates of average yearly user fees collected by the facilities. Note that user fees are not Table 5.9  Average Annual User Fees Collected, that significant in overall expenditures of the facilities. by Facility Type, PNG Kina The user fees reported by the facilities recovered Level 3 and 4 Level 5 to 7 Fee collection facilities public only about 0.25 percent of expenditures for level 3 and 4 public facilities, 1.92 percent for level 3 and 4 by facilities Public Church Facilities church-run facilities, and 1.14 percent for level 5 to Number of facilities 28 25 19 7 public facilities. Yearly collection 857 12,684 44,226 of registration and Almost all of the facilities charging user fees fol- consultation fees lowed some guidelines for fee exemption. Only two Yearly collection for 86 164 8,107 church-run facilities (out of 15 charging fees) and one drug dispensation level 3 and 4 public facility (out of eight) reported not Yearly collection for 437 2,880 84,265 having any written guidelines and another reported laboratory tests not allowing any fee exemptions. Guidelines were pre- Yearly collection for 316 3,356 46,570 admission fees/rents pared by only one level 3 and 4 public health facility, Yearly collection for 1,696 19,084 183,169 but six church-run facilities and five level 5 to 7 pub- total user fees lic facilities prepared their own user fee policies. For Percent of total facility 0.25 1.92 1.14 the most part, facilities retain any collected user fees. expenditure Thirty-nine percent of facilities reported that they Note: Figures based on monthly fee collection reported by facilities surveyed. Cost and Output of Facilities 63 from individuals in the community. If the facility level 7 facility. The range of services provided by arranged outreach activities, most received support health facilities clearly indicate that even the level 3 from local communities with regard to arranging and 4 facilities are quite complex production units, a place for carrying out the outreach. One-third of therefore it is difficult to collect comprehensive facilities received support in the form of voluntary information on production and costs. contributions of time and money. Salaries, operations, and utilities, rent, and maintenance make up a large component of health facility expenses. Level 3 and 4 church-run health Summary facilities show lower average expenditure on salary and Outpatient Consultation and Inpatient Admissions compensation than the public sector facilities. Further, training is significantly lower in level 3 and 4 facilities Level 5 to 7 health facilities had significantly compared with that for level 5 to 7 facilities, implying higher inpatient days than level 3 and 4 facilities. greater emphasis on training at upper-level facilities. Average inpatient days in level 6 facilities were quite Drug expenses are relatively low in all facility types, high compared with level 5 facilities. However, one especially in level 5 to 7 facilities. The capital expendi- should keep in mind that there are only three facili- ture (without annualizing) was quite high for level 3 ties in level 6 and 15 facilities in level 5. The length of and 4 church-run facilities. Church facilities show stay also varies by the inpatient ward of admission. highest operational expenses in relative terms, exclud- This variation is probably more of a reflection of num- ing capital cost items. This cost is probably to the result ber of beds in different wards than the length of stay of the significantly higher number of outreach patrols in wards. Among all the different inpatient wards, they have organized compared with other facility types. medical and surgery wards experienced the highest level of inpatient stays. Pediatric and OBGYN wards Budget and Expenditure were also important in relation to inpatient days. The most important cause of seeking care was The percentage of budgets expended differs quite infectious disease. Respiratory illnesses were the drastically between health facility levels. Level 3 leading cause of seeking treatment from health facili- and 4 public health facilities spent 28 percent of their ties. For all health facility levels, aside from level 7, the budgets, whereas level 3 and 4 church health facilities spent 52 percent of their budgets. In comparison, level second leading cause of seeking treatment was malaria. 5 and 6 health facilities spent, on average, 92 percent At the tertiary hospital (level 7 facility), malaria was of their budgets, and the level 7 health facility over- the sixth leading cause of seeking care. Other common spent its budget by about one-fifth. It appears that causes of seeking treatment are diarrhea and acci- health facilities have little or no control over the bud- dents and injuries. geted resources. It is important to develop the budget in close collaboration with the facilities and the facili- Outreach Activities ties should have access to the budgeted resources to improve their effectiveness. Level 3 and 4 public health facilities performed User fees are still charged by some lower-level poorest in conducting outreach activities. Level 3 facilities, despite the policy of free primary health and 4 public sector facilities planned to conduct a care and subsidized secondary care. Level 5 to 7 relatively high number of outreach patrols but cited health facilities collected significantly higher user fees lack of funds as the primary reason they were ulti- than level 3 and 4 facilities. In general, however, user mately unable to conduct the outreach. Outreach fees collected were very low compared to the cost of patrols improve accessibility to health care services running a health facility. Even for level 5 to 7 facilities, for populations in remote areas, but in PNG they are total user fees collected were only K 183,000. User fees quite expensive and hard to sustain. are mostly retained by health facilities and funneled back into the health facility to either meet shortfalls or Cost of Running a Health Facility to cover unplanned expenses. Health facilities noted that the user fees generally went to improving infra- Excluding capital expenses, the average cost of facil- structure; buying drugs or other medical supplies; ity operation varied from K 1.6 million for a level 3 paying for casual labor, administration, and other and 4 church-run facility to K 57.7 million for the office expenses; and purchasing food for patients. Chapter 6 EFFICIENCY OF HEALTH FACILITIES SURVEYED This chapter focuses on the measurement of effi- the bed occupancy rates in the facilities surveyed. ciency scores, which compares levels of inputs used In both the approaches, the length of stay reported in the health facility in relation to the output pro- by the facilities, in particular level 3 and 4 health duced to measure the health facility efficiency. This facilities, was quite low. Table 6.1 reports the num- measurement is not straightforward because mul- ber of occupied beds on the day of the survey tiple inputs and outputs are involved in the health (observed by the survey team, method 1), estimate facility production. Two methods are used. The first of annual hospital inpatient days by using 1 month method is to use the variables, which directly mea- of administrative data (method 2A), and estimate of sure some aspects of efficiency, for example, hospi- annual hospital inpatient days by using 4 months of tal bed occupancy rate as a measure of efficiency for data collected from the facility by the survey team health facilities offering inpatient care; cost per out- (1 month per quarter in 2014, method 2B). Using put is another set of measures. The second method the inpatient day estimates and given the bed size uses statistical approaches to measure efficiency, in of the hospitals or facilities, bed occupancy rates can which efficiency scores based on production func- be calculated. tion or cost functions were constructed.38 Bed occupancy rates, using method 1, vary roughly from 40 percent for level 3 and 4 public sector facilities to 81 percent for the level 7 facility. Specific Ratios and Proportions Note the differences in the estimates of bed occupancy as Efficiency Measures rates by these three approaches and methods. The bed occupancy rates were the highest for method 1, Bed Occupancy Rate denoted by occupancy rate 1, and were the lowest Two approaches were used (based on actual data for method 2B, denoted by occupancy rate 2B. On collection from the facility or observed occupied the basis of different measures of bed occupancy rate, it is assumed that method 1, denoted by occu- beds on the day of the survey and administra- pancy rate 1, provides the most unbiased estimates. tively reported hospital days of stay) to calculate Bed occupancy rates vary by the size of the hospital and location, but not by the ownership of facility. The regression analyses show that a greater 38 Different approaches of measuring hospital or health facility number of beds in a facility corresponds to a higher efficiency are discussed in a book by Barnum and Kutzin (Barnum bed occupancy rate. The bed occupancy rate increases and Kutzin, 1993). One of the concerns in the measurement of efficiency is the problem of controlling for patient-mix. It is likely by 0.056 with one additional bed. Facilities located that the patients in upper-level hospitals have higher severity of in urban areas show significantly higher bed occu- medical conditions than the patients in lower-level facilities. In pancy rates than hospitals located in other areas. the data set, there is not have enough information for controlling the variability of severity of medical conditions of patients across Ownership of the facility shows no systematic effect health facilities. on bed occupancy. 65 66 Service Delivery by Health Facilities in Papua New Guinea Table 6.1  PNG Hospital Inpatient Days and Bed Occupancy Rates, 2014 Level 3 and 4 Level 3 and 4 Inpatient days and bed occupancy rate public church Level 5 and 6 Level 7 Number of beds occupied   Method 1: Average number of beds occupied 8.7 14.0 112.1 720.0 on the day of survey (observed)    Average number of beds in facilities 21.7 37.2 158.8 886.0 Number of inpatient days in a year based on 1 month of 4 months of data   Method 2A: Average hospital days of stay in 1,802.1 3,759.5 24,638.7 25,8852.0 2014 based on inpatient days in December 2014 (1 month administrative data × 12)   Method 2B: Average hospital days of stay 1,632.3 3,372.8 24,186.5 175,353.0 in 2014 based on 4 months’ information (4 months of data collected by the survey × 3) Max days of stay that can be created in a year 7,925.7 13,565.8 57,953.9 323,390.0    Method 1: Bed occupancy rate 1 39.9 37.6 70.6 81.3    Method 2A: Bed occupancy rate 2A 22.7 27.7 42.5 80.0    Method 2B: Bed occupancy rate 2B 20.6 24.9 41.7 54.2 Personnel-Based Performance Measures of personnel. The mean number of doctors per bed varied from 0.00 to 0.16. In level 3 and 4 public facili- Personnel-based performance measures are pre- ties, on average, there are 28 beds per doctor. In level sented in Table 6.2. The measures are based on the 3 and 4 church-run facilities, there are 54 beds per number of beds in the facility and the quantities of doctor on average. Average number of beds per output produced in comparison with the number doctor was 9.5 and 7.3 in level 5 and 6 facilities, Table 6.2  Select Performance Measures Based on Inputs and Output Level 3 and 4 public Level 3 and 4 church Level 5 public Level 6 public Level 5 to 7 public Indicators Mean SD Mean SD Mean SD Mean SD Mean SD Number of doctors 0.02 0.05 0.00 0.01 0.13 0.07 0.16 0.08 0.14 0.07 per bed Number of personnel 1.21 1.29 0.86 0.68 1.86 0.81 2.47 0.59 1.93 0.79 per bed Number of nurses 7.30 4.94 7.25 7.50 4.19 2.57 4.90 1.19 4.20 2.37 per doctora Number of annual 2,836 3,186 3,910 2,227 1,480 615 1,332 501 1,445 572 inpatient days per doctora Number of annual 28,759 18,101 16,633 14,753 6,870. 4,338 5,755 5,577 6,475 4,370 outpatient visits per doctora Number of annual 388 377 806 732 1,129 2,759 304 200 970 2,455 inpatient days per nurse Number of annual 8,704 18,934 4,957 4,856 3,045 3,853 1,047 942 2,636 3,509 outpatient visits per nursea Note: SD = standard deviation. a. The category “doctors” includes all medical officers, specialist medical officers, senior specialist medical officers, dentists, and radiologists. Efficiency of Health Facilities Surveyed 67 respectively. Average number of beds per doctor in considered. Since many of the level 3 and 4 facili- level 5 to 7 facilities was 9. Average total number of ties did not have doctors, measuring productivity personnel per bed was 1.21, 0.86, 1.86, 2.47, and 1.93 of doctors using total utilization of the facilities will for level 3 and 4 public, level 3 and 4 church, level 5, overestimate true productivity of doctors. At level 3 level 6, and level 5 to 7 facilities, respectively. The and 4, the principal health service providers are the mean nurse-to-doctor ratios were 7.30, 7.25, 4.19, health extension officers (HEOs), nurses, community 4.90, and 4.20 for level 3 and 4 public, level 3 and 4 health workers (CHWs) as well as doctors. For level 3 church, level 5, level 6, and level 5 to 7 facilities, and 4 facilities, productivity has been measured by respectively. considering all the health care providers. For level 5 Output produced by health facilities are mea- to 7 facilities, productivity was measured for doctors sured by two broad categories: outpatient visits as well as for nurses separately to indicate the quan- and inpatient days. The average number of annual tities of services delivered by the facility personnel. inpatient days created per doctor was 2,836, 3,910, The productivity levels in Table 6.3 are expressed 1,480, 1,332, and 1,445 in level 3 and 4 public, level 3 as number of outputs per health care provider (for and 4 church, level 5, level 6, and level 5 to 7 facilities, level 3 and 4 facilities) and per doctor (for level 5 to 7 respectively. With 3,910 inpatient days per doctor, facilities) per day. In the final three rows, all the level 3 and 4 church-run facilities produced sig- ratios are expressed as number of output per health nificantly higher inpatient days than public sector care provider per day (doctors, nurses, HEOs, and facilities. Number of annual inpatient days per nurse CHWs) for comparative purposes. It is assumed that varied from 388 for level 3 and 4 public facilities to the average number of working days in a year is 250 970 days for level 5 to 7 public facilities. Number (i.e., a full-time equivalent will work 250 days in a of outpatient visits per nurse for level 3 and 4 year). In deriving the total output, inpatient days public facilities was 8,704, while it was 4,957 for are considered twice as intensive as an outpatient level 3 and 4 church-run facilities. This ratio was visit. Therefore, total output is simply the number 3,045 for level 5, 1,047 for level 6, and 2,636 for of outpatient visits plus two times the number of level 5 to 7 facilities. Since nurses are one of the inpatient visits. This type of average productivity val- important health care providers in PNG, inpatient ues is widely used at the sectoral level for comparative or outpatient visits per nurse may be more relevant purposes in many countries of the world, including than output per doctor. Note that the ratios do not the Organisation for Economic Co-operation and represent the number of patients seen by nurses Development (OECD) countries. or doctors per year; these only reflect the general Productivity per day of health care providers in workload of doctors and nurses assuming that both PNG can be increased further at the lower health doctors’ and nurses’ services are required at both facilities. At level 3 and 4 facilities, outpatient visits inpatient and outpatient areas. are predominant and health care providers should be Productivity levels of health facility person- able to produce 15 to 20 visits in a day if the demand nel are easier to understand if per-day outputs are at the local level is present. Compared with the level 7 Table 6.3  Productivity of Health Care Providers in Surveyed Facilities by Facility Level (Output per Day), 2014 Average level of output per health care provider Level 3 and 4 Level 3 and 4 Level 5 Level 6 Level 7 Productivity of health facility personnel public church public public public Number of inpatient days per health care provider 0.5 0.9  6.1  6.4  6.5 (level 3 and 4) or per doctor (level 5 to 7) per day Number of outpatient visits per health care provider 8.5 5.1 32.1 26.4 24.7 (level 3 and 4) or per doctor (level 5 to 7) per day Number of inpatient days per health care provider 0.5 0.9  0.8  0.7  1.7 Number of outpatient visits per health care provider 8.5 5.1  4.1  2.9  6.3 Total output per health care provider per day 9.5 6.9  5.6  4.3  9.6 68 Service Delivery by Health Facilities in Papua New Guinea Figure 6.1  Cost of Inputs and Outputs per Unit, by Facility Type 120,000 120 106 100,000 103 100 Total cost to total output ratio Average cost in PNG Kina 88 80,000 80 60,000 59 60 54 40,000 40 20,000 20 0 0 Level 3-4 Public Level 3-4 Church Level 5 Level 6 Level 5-7 Total cost per bed Total cost to total personnel Total cost to total clinical staff Total cost to total output facility, level 5 and 6 facilities also show relatively low per bed, average cost per personnel, average cost per output per health care provider. Level 5 and 6 facili- clinical staff, and average cost per output (defined ties should be able to produce at least at the same in footnote 39).39 level as the level 7 facility, implying that the level 5 Large variation exists in cost per hospital bed, and 6 facilities can improve the health care provider ranging from K 34,012 to K 110,417 per year. The productivity by about 55% on average. In general, average cost per bed was the highest for level 5 facili- level 5 to 7 facilities in PNG produce about 31 out- ties and was the lowest for level 3 and 4 church-run patient visits per doctor per day on average. This health facilities (figure 6.1). Similar patterns are number is much higher than the outpatient visits found in other indicators, reiterating the finding that per day produced by physicians in several coun- church-run facilities are generally more efficient com- tries. For example, OECD data indicate that number pared with level 3 and 4 public health facilities. As of outpatient visits per physician per day was 6.2 expected, level 5 to 7 health facilities generally have in Brazil, 12.4 in China, and 12.5 in South Africa higher costs because the case mix is generally more (for the years 2014 or 2015). However, these num- complicated in the secondary and tertiary levels of bers are sectoral (i.e., for the whole health sector hospitals. but the PNG ratios are for health facilities provid- ing health care services). A significant proportion 39 The output was defined by calculating total outpatient visits, of doctors at the sectoral level do not provide direct inpatient days, and the vaccinations delivered through outreach patient services (work as educators, administrative programs. The outpatient visits also include the number of indi- officials, non-patient-related medical services, etc.) viduals using the laboratory services, if the service is offered by a facility. The data on inpatient and outpatient facility use and, therefore, the number of patients seen per day were collected for 4 months, 1 month to represent a quarter in will be significantly higher for these countries if the the previous year, and total visits and inpatient days per year were doctors working only in health care facilities are estimated from the information for these 4 months. The aggregate output was defined by adding two times of inpatient days with considered. outpatient visits. In PNG, average cost per day of hospital stay has been reported as twice as much as the cost of seeing a doctor in the outpatient department (in the private sector based on information Cost-Based Performance Indicators collected from few facilities by PNG Institute of Medical Research researchers after the completion of the survey). Once the cost per This section presents the cost-based efficiency unit of output is derived, all facilities were ranked by the value of measures. Four indicators were used: average cost the ratio. Efficiency of Health Facilities Surveyed 69 Efficiency Measure Based on Multiple Ratios total output is outpatient visit equivalents. Costs per output produced are ranked from the lowest to the Cost per Unit of Output Produced highest values by ownership type (church-run and One measure of efficiency could be based on cost- public facilities). The figure also shows the aver- to-output ratio (i.e., cost of producing one unit of age cost and the first quartile value of the cost per output). In this measure, cost is considered a com- unit of output for level 3 and 4 facilities. In general, posite input where all inputs are combined into one church facilities produce output at a lower cost per number using prices of inputs. Combining all inputs unit compared with public facilities. For example, using prices is a valid approach of defining the com- only two out of 22 church-run level 3 and 4 facilities posite index if the Hicksian aggregation principle is (for which we have relevant data) had average cost satisfied. Hicksian principle says that commodities per output higher than the overall average, but seven can be aggregated into one combined measure if public facilities out of 24 facilities show higher than all prices are the same or change by a fixed propor- the average cost. If the first quartile value is used, a tion. In the case of health facilities, prices of inputs higher proportion of level 3 and 4 church-run facili- are likely to be very similar across all public facili- ties (7 out of 22, i.e., 32% of facilities) were found to ties. As mentioned earlier, salary and other benefits be efficient than level 3 and 4 public facilities (6 out paid to facility personnel are significantly lower for of 24, i.e., 25%). church-run facilities compared with public facili- Figure 6.3 shows the cost per output graph for ties. Therefore, church-run facilities do not satisfy level 5 to 7 facilities. These facilities provide more the Hicksian aggregation principle and comparison comprehensive and other specialized services. The of cost per output between public and church-run figure also shows the average cost and first quartile facilities will be misleading. value of the cost per output. If the average cost num- Figure 6.2 shows cost per unit of output pro- ber of these higher-level facilities is used, three out duced by level 3 and 4 church-run and public sector of 19 were above the overall average cost (16% of facilities. Total output has been defined by combin- facilities). If the first quartile value is used to define ing the outpatient visits and inpatient days of stay efficient facilities, five out of 19 (26%) may be con- (one inpatient day is considered equivalent to two sidered efficient within this upper-level facility type. outpatient visits) and the unit of measurement of Among these facilities (levels 5 to 7), the average Figure 6.2  Comparison of Cost per Unit Output for Church-Run and Public Level 3 and 4 Health Facilities in PNG (2014 Costs) 300 250 Cost per unit output 200 150 100 50 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 Level 3 and 4 Church-run facilities Level 3 and 4 public sector Facilities Cost per output Average 1st Quartile 70 Service Delivery by Health Facilities in Papua New Guinea Figure 6.3  Cost per Output for Level 5 to 7 Facilities in PNG (Cost for the Year 2014) 600 500 400 Cost per output 300 200 100 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 Level 5-7 facilities Cost per output Mean 1st Quartile costs per unit of output were significantly higher that have low occupancy and low turnover; zone 2 than others for two specific facilities. Excluding contains facilities that have low occupancy and high these two facilities would have significantly affected turnover; zone 3 contains facilities that have high the average cost per output but would have very little occupancy and high turnover; and zone 4 contains effect on the first quartile value. facilities that have high occupancy and low turnover. In this classification, facilities that fall in zone 3 are Lasso Efficiency Measure deemed “efficient.” Figure 6.4 shows the facilities Lasso (Lasso 1986) developed a hospital performance in the survey and their location by the four zones assessment tool that simultaneously incorporates defined above. Note that a total of 18 facilities are bed occupancy rate, bed turnover rate, and length classified as zone 3, the efficient zone. of stay; bed turnover rate is plotted against the Size matters for health facility efficiency; in occupancy rate. The chart is divided into quadrants PNG smaller health facilities are more likely to based on the mean values of bed occupancy rate and be less efficient compared with the larger health bed turnover rate. Quadrants are categorized into facilities. In the survey, 49 facilities had 0 to efficiency zones (figure 6.4). Zone 1 contains facilities 50 beds and another 10 were in the size category of 51 to 100 beds. Among the smaller health facili- ties (50 beds or lower), 14 percent of facilities fell Figure 6.4  Plot of PNG Health Facilities, by Bed Occupancy within the efficient zone, while for 51- to 100-bed and Bed Turnover Rates hospitals, 40 percent were in the efficient zone. Lasso Efficiency of Facilities Fifty percent of larger health facilities (i.e., facili- 150 ties with more than 100 beds) were in the efficient zones. Table 6.4 shows the distribution of health facilities in the survey by bed size and percentage Bed turnover 100 Zone 2 Table 6.4  Distribution of Health Facilities, Zone 3 by Bed Size and Efficiency Number Number of Percent of 50 of facilities facilities facilities Bed size in the Lasso- efficient in categories survey efficient the category Zone 1 Zone 4 0–50 49  7 14.29% 0 51–100 10  4 40.00% 0 20 40 60 80 100 101+ 14  7 50.00% Occupancy rate Total 73 18 24.66% Efficiency of Health Facilities Surveyed 71 of efficient facilities based on Lasso efficiency Table 6.5  Distribution of PNG Health Facilities, by Location, criteria. Ownership, and Efficiency Variation in health facility efficiency is signifi- Number of Percent of cant across the regions. Table 6.5 shows that in the Number of Lasso-efficient facilities Region/facility type facilities facilities efficient Highlands about 14 percent of facilities surveyed belonged in the Lasso-efficient zone compared with By region   Highlands 21  3 14.29% 24 percent in Momase, 41 percent in New Guinea   Momase 17  4 23.53% Islands (NGI), and 22 percent in Southern.   NGI 17  7 41.18% Ownership is also significant in health facil-   Southern 18  4 22.22% ity efficiency. Only about 7 percent of level 3 and 4 By level and ownership    Level 3 and 4 public 29  4 13.79% public facilities were in the efficient zone as compared    Level 3 and 4 church 25  2  8.00% with 16 percent of church-run level 3 and 4 facilities    Level 5 to 7 19 12 63.16% and 63 percent of level 5 to 7 facilities. Production and Cost Function-Based A simple regression shows that a 1 percent increase in the number of beds in the facility is Efficiency Measures associated with an 0.41 percent increase in total out- The report also examines the overall production, puts produced, after controlling for other factors. which is defined as total outpatient visits, in patient log ( output ) = 8.79 + 0.41log ( BED ) visits and outreach patrols. As they are different in ( 23.4 ) (3.9) nature, a subjective weight was applied and in this − 0.06 CHU3_4 + 1.05 PUB5 case40 inpatient days were weighted by 2.0. ( −0.33) (3.81) The inputs of the production process included + 1.34 PUB6 + 2.12 PUB7 the number of beds, number of clinical personnel, ( 4.29) ( 4.96) number of nonclinical personnel, floor space of facility, and number of outreach patrols conducted. R 2 = 0.69, N = 68 In the estimation process, it became clear that many (Note: t-statistics are shown below the estimated inputs cannot be used directly because of the multi­ coefficients in parentheses.) collinearity problem (i.e., facilities with a higher number of clinical personnel are also larger in size The regression analysis does not show a significant and use more drugs and other supplies). For example, efficiency difference between level 3 and 4 public number of beds and the number of personnel in the health facilities versus church health facilities, after facilities are likely to be highly correlated. The cor­ controlling for number of beds. relation coefficient between total number of beds and Similarly, a positive relationship was observed total personnel is 0.92, between total beds and floor between health facility personnel and output, space is 0.87, and between total personnel and total implying that health facilities in PNG can ben- floor space is 0.74. efit from more health workers to use the capital A positive relationship is observed between bed and other resources to better serve patients and size of the facility and the output produced. The improve the overall efficiency of health service relationship also appears to be quite strong, given that delivery. Figure 6.6 shows another scatter plot with log of total number of employees in the facilities as most of the observations (points in the graph) are the explanatory variable. The relationship between clustered close to the fitted line shown in figure 6.5. these two log values was found to be quite strong However, there are three relatively small health facili- with all points close to the fitted line. However, there ties that produce relatively high levels of output. Also, are some facilities that employ a relatively high num- there is one facility that is clearly an outlier with a ber of employees even though their outputs are quite relatively high bed size. low and vice versa. The regression result shows that with a 1 per- cent increase in the number of personnel across 40 Inpatient days of stay was weighted by 2.0. For sensitivity the facilities, output of the facilities increased by analysis, other weights were used as well (e.g., 5.0 and 10.0). 0.73 percent. 72 Service Delivery by Health Facilities in Papua New Guinea Figure 6.5  Relationship Between Total Production and Total Number of Beds 14 Relatively small facilities with high output Relatively large 12 facility with high output Log of output 10 Relatively large facility with low output 8 0 2 4 6 8 Log of beds Southern Momase Highlands Islands Fitted values Figure 6.6  Relationship Between Total Production and Total Personnel 14 12 Log of output 10 Relatively high employment but low output 8 0 2 4 6 8 Log of labor Southern Momase Highlands Islands Fitted values Efficiency of Health Facilities Surveyed 73 Figure 6.7  Relationship Between Total Production and Total Number of Annual Patrols Conducted 14 12 Log of output 10 8 0 2 4 6 Log of beds Southern Momase Highlands Islands Fitted values log ( output ) = 7.89 + 0.731log ( PERSONNEL ) a proxy for regional characteristics, such as popula- ( 49.70) ( 21.34 ) tion density and access barriers. However, significant effects of the level ownership category on output R 2 = 0.76, F stat = 455.57 were found when used with beds and floor space as (Note: t-statistics are in parentheses.) covariates in the production estimation. Regional categories have been found to be consistently insig- Outreach efforts have a weak impact on outputs. nificant in all of the estimations. There appears to be no relationship between the Constrained regression analysis was used to get number of outreach patrols conducted and outputs a rough measure of efficiency scores where the final produced. It is possible that outreach allows facili- production function is estimated using a two-step ties to reach a higher number of users of the facilities approach.41 In the first stage, all the data points and, therefore, outreach programs may improve uti- are used to estimate the production function and lization. On the other contrary, it may also be possi- the observations that show negative “error terms” ble that facilities conduct more outreach in relatively (i.e., the observations located below the production low population density areas. Figure 6.7 shows the function are dropped). The remaining observations relationship between log value of outreach patrols in a year and the log of outputs produced. In the data 41 set, a significant number of facilities did not organize The stochastic frontier model was estimated using the two input variables and one output variable but the test statistics any outreach. For the facilities that organized out- indicate that the decomposition of error term to identify sys- reach, the relationship between log of output and log tematic variations due to differences in efficiency does not exit. outreach visits is weak (significant only at 10 percent In other words, the data do not show any systematic efficiency differences among the facilities in the survey. The results do not significance level with an R2 of 0.05). change if the estimations are controlled for level and ownership Since the level and ownership of the facilities and regions of the facilities. may indicate differences in management practice To obtain efficiency scores for the facilities, a number of alter- or complexity of patients’ medical condition, these native approaches can be used, including the nonparametric approaches such as data envelopment analysis and full disposal variables were used as control variables in the sta- hull. The data envelopment analysis using the inputs and outputs tistical model. Also, the facility regions can serve as described above also did not produce any efficiency scores. 74 Service Delivery by Health Facilities in Papua New Guinea Table 6.6  Distribution of Health Facility Efficiency Scores, by Level and Ownership Efficiency score of health facilities using Level 3 and 4 facilities order-alpha partial Number Percent Number of Percent of Level 5 facilities Level 5 to 7 facilities frontier estimation of public of public church-run church-run Number Percent Number Percent Less than or equal to 1 12  48 13  54 14  93 18  95 1 to 2 11  44  7  29  1   7  1   5 Greater than 2  2   8  4  17  0   0  0   0 Total 25 100 24 100 15 100 19 100 are used to estimate the second production function, differences in efficiency scores across the facilities and using this production function as the “frontier,” were quite small, even though church-run facilities efficiency scores were estimated for each of the appear to be slightly better than other categories if facilities. In the calculation of efficiency scores, if the least efficient percentages are considered (i.e., the error term is positive, the facilities are considered percentage of facilities not belonging to least efficient “fully efficient.” category). The final production function estimated through The order-alpha approach42 was used to confirm the second stage is reported below: the initial findings: level 3 and 4 health facilities, on average, are more efficient than level 5 to 7 facilities. The distribution of efficiency scores log ( output ) = 10.23 + 0.016 BEDS obtained through the order-alpha approach is pre- ( 49.67 ) ( 2.51) sented in table 6.6. Because of the presence of miss- − 0.000037 BEDS2 − 0.594 CHU3_4 ing values, efficiency scores could be calculated for ( −1.94 ) ( −0.54 ) 67 facilities. The efficiency scores of 43 facilities + .55 PUB5 + .670 PUB6 + 18.55 PUB7 (out of 67) were found to be 1.0 or less than 1.0. In (1.44 ) (0.89) (1.83) the table, this category represents the least efficient group of health facilities. Among level 5 to 7 facilities, Using this production function, 16 out of 70 88 percent belonged to the least efficient category, (22.8 percent) health facilities were classified as while the percentages were 44 percent for level 3 fully efficient. Among these fully efficient health and 4 public facilities and 20 percent for level 3 and 4 facilities, five (31.2 percent) were level 3 and 4 public church-run facilities. Twenty-four percent of level 3 health facilities, five (31.2 percent) were level 3 and 4 and 4 public facilities and 52 percent of level 3 and 4 church-run facilities, five (31.2 percent) were level 5 church-run facilities fall into the middle efficiency public facilities, and one was a level 7 facility. category. The efficiency score of more than 2.0 has After controlling for the bed size or removing been defined here as super-efficient health facilities the scale effects, level 5 to 7 hospitals, on average, and about one-third of level 3 and 4 public health are less efficient compared with level 3 and 4 health facilities belonged to this category. facilities. Among level 3 and 4 public health facilities, Figure 6.8 is the scatter plot of all the facilities 30 percent were in the least efficient facility category showing the output produced and the cost incurred (lowest one-third of all facilities), 22 percent were by facilities to produce the output. Since some of the in the middle efficiency category, and the remain- facilities did not report costs, inpatient outputs, or ing 48 percent were in the most efficient category. outpatient outputs, not all facilities are plotted on For level 3 and 4 church-run facilities, 25 percent the scatter plot. As can be seen in the plot, the scatter were in the least efficient category, 38 percent were 42 Since the constrained regression model uses the first regression in the middle efficiency category, and the remaining model to arbitrarily drop observations, it may lead to a biased 38 percent were in the most efficient category. For estimate of efficiency scores. To avoid this problem and to com- level 5 facilities, these percentages were 53 percent, pare consistency of efficiency scores by using a more systematic 40 percent, and 7 percent, respectively; for level 5 to 7 approach, a partial frontier estimation method of estimating effi- ciency scores, known as order-alpha, has been adopted. The order- facilities, these percentages were 53 percent, 44 per- alpha approach allows the outliers to become “super-efficient” so that cent, and 5 percent, respectively. In any case, the sensitivity to outliers is significantly reduced (Tauchmann, 2011). Efficiency of Health Facilities Surveyed 75 Figure 6.8  Free-hand Production Frontier (Output vs. Cost) and the Relative Inefficiency of Health Facilities in PNG 13 12.5 12 11.5 11 Log of output 10.5 10 9.5 9 8.5 8 1 2 3 4 5 6 7 Log of cost shows a high degree of variability and clearly shows The number of beds per doctor at level 5 to 7 facil­ why a frontier is difficult to define. A simple hand- ities was about six to seven, and for level 3 and 4 drawn frontier is shown by connecting the highest facilities the ratios were much higher because of points, the highest level of output produced, given a shortage of doctors at lower-level facilities. The the costs incurred by the health facilities in produc- number of personnel per bed was highest in level 6 ing the output. Since lower output with increasing facilities followed by the level 7 facility. The low- cost of running facilities implies inefficiency, the est personnel-to-bed ratio was found in church- production frontier cannot show a declining seg- run facilities. The total cost per bed was the lowest ment with increasing cost. Using the straight line for level 3 and 4 church-run health facilities, and segments of the production frontier (with no declin- the second lowest ratio was for level 3 and 4 public ing section with increasing cost), there are only sector facilities. The cost per output was also the two facilities that fall on the frontier and all other lowest for church-run facilities and the highest for points are below the frontier. This result implies that level 5 facilities. The cost per clinical personnel was almost all facilities in the survey were not as efficient the lowest for level 6 facilities followed by church- as they could be. run facilities. In relation to the cost per unit of output, the most efficient facility category appears to be church-run Summary facilities. The highest cost facilities were level 5 to In general, upper-level facilities (i.e., facilities at 7 facilities. Since these upper-level facilities provide levels 5 to 7) are able to use hospital inputs more services to high-severity patients, it is not surprising efficiently than other facilities. For example, the to see relatively high cost per output in these facilities. bed occupancy rate improved with the increase in It was not possible to correct output values because bed size of the facility and whether or not the facility of the severity of medical conditions of the patient is located in urban areas. Facilities located in urban population. Therefore, this ratio may not represent areas had improved bed occupancy rates. relative efficiency of the facilities. A number of ratios were used to estimate the For Lasso efficiency categorization, a signifi- relative efficiency of health facilities. These ratios cantly higher proportion of health facilities in the are either input-output based or cost-output based. NGI region were found to be efficient compared 76 Service Delivery by Health Facilities in Papua New Guinea with facilities in the other three regions. Lasso 50 to 100 beds, and 50 percent for more than 100-bed efficiency was measured using two ratios: bed occu- facilities. pancy and turnover rates. For Lasso efficiency cate­ Production function-based efficiency mea- gorization, a significantly higher proportion of surements provided mixed results. The standard health facilities in the NGI region were found to be production function analysis does not support the efficient compared with facilities in the other three hypothesis that health facilities in PNG vary signifi- regions. Level 5 to 7 facilities were also found to be cantly in efficiency in resource utilization. Although more efficient than other facility categories in Lasso other approaches show some variability in efficiency efficiency. Larger facility sizes were more efficient, scores, the scores are usually quite similar across which is consistent with the result that level 5 to 7 surveyed facilities. Order-alpha produces wider vari- facilities were relatively more Lasso-efficient. Only 14 ability of efficiency scores; however, no clear under- percent of facilities with bed size less than 50 beds lying factors can be found to explain the relative were Lasso-efficient, while it was 40 percent for size efficiency levels. Chapter 7 POLICY IMPLICATIONS The facility survey intended to generate information who ultimately need to spend it. The World Bank on the provision of health services, readiness of the report “Below the Glass Floor: the analytical review facilities in service provision, human and material of expenditure by provincial administrations on resources used in health care service delivery, and the front line rural health” (2013) and “Financing the quality, safety, and efficiency of the surveyed facili- frontline: an analytical review of provincial admin- ties. The recommendations for PNG policy makers istrations in Papua New Guinea’s rural health expen- are listed below. diture 2006–2012 (2015)” provides detailed analyses The policy recommendations are grouped into four and policy recommendations. major areas: key enablers for service delivery, key prior- Stable water and electricity supply needs to be ity service areas, efficiency, and technical assistance to addressed for all facilities. Although most of the level 5 the National Department of Health (NDOH). to 7 facilities were connected with an electric supply line, most of the level 3 and 4 facilities surveyed were not. About 15 to 20 percent of level 3 and 4 facilities Policy Recommendations to Improve did not have water for health care providers to use in the Key Enablers for Service Delivery treating patients. A well-functioning health facility should be well The availability of essential drugs at all health maintained, and adequate resources should be allo- facility levels needs to be improved. Using the list of cated for facility maintenance, especially for lower- basic essential drugs, the survey found that drug avail- level facilities. Infrastructural conditions of the ability on a continuous basis is lacking in all levels of surveyed facilities indicate that about 30 to 40 percent health facilities, including the level 7 facility. The non- of level 3 to 6 facilities require major repairs. More than availability of drugs affects the quality of health ser- 60 percent of level 3 and 4 facilities did not have vices as well as satisfaction and motivation levels of an adequate number of toilets (for level 5 and 6, the providers. Therefore, nonavailability of drugs has two percentage was 40 percent). Therefore, PNG health distinct adverse effects: one on clients and the other sector decision makers will have to deal with two on health care providers. Although PNG allows facili- aspects of infrastructural issues: one is related to main- ties to supplement drugs obtained through the “push tenance and the other is related to the overall design of system,” a direct requisition and purchase by health health facilities. facilities, it still appears inadequate to ensure a stable Getting the ‘right amount’ to the ‘right place’ supply of essential drugs. and ‘on time’ needs to be a guiding priority. Budgets The drug monitoring and accounting system and funding streams need to be appropriately needs to be strengthened. In the survey, facilities aligned to ensure the right amount gets to the right reported the number of drug and supply boxes level. Getting the money right is not only a matter of they received in the previous year as well as the providing the right amount, it’s also a matter of get- drugs procured outside of the push system. The ting the money to the right place and to the people amount of drugs obtained or procured, as reported, 77 78 Service Delivery by Health Facilities in Papua New Guinea was significantly lower than expected. In most cases, provider types imply that any of these personnel will drug costs were less than 3 percent of the total facil- be equally effective (or ineffective) in the provision ity operation cost. It is likely that the facilities do not of the basic priority services. The in-service trainings keep detailed information on drugs obtained or pro- should be strengthened on these specific areas; and cured. The system of monitoring drug supplies needs with quality training, nurses or CHWs should be able strengthening. Moreover, the push system of drug to provide certain services at the same effectiveness supply may create a significant misallocation of scarce level as doctors. This is not to imply that policy makers drugs among health facilities. should not consider appointing doctors at lower-level The number of human resources in health facili- facilities. It is important to have doctors in referral ties needs to be improved to match the budgeted facilities among level 3 and 4 facilities, but given the positions. Human resource availability at health facil- overall shortages of doctors, it will be appropriate to ities is lower than the number of positions assigned consider appointing other health personnel to provide or budgeted. At level 3 and 4 facilities, about 40 per- maternal, child health, and STI-related services. cent of assigned doctor positions were not filled. Institute the user fee waiver policy for patients The percentage of positions not filled was also high at lower-level health facilities. Users of level 3 and 4 for nurses and community health workers (CHWs) facilities are predominantly poor, as indicated by the (about 25 percent). Even at higher levels (level 5 to client surveys from the health facilities. The facilities 7 facilities), more than 20 percent of physician and in the survey reported that they collect only a small nurse positions were vacant. PNG started the health amount of money in user fees. User fees collected, as extension officer cadre to deal with the shortage of reported by facilities, represent less than 2 percent of clinical personnel, but the emphasis on this alter- facility expenditure. Level 3 and 4 church facilities native may have declined in recent years. It appears collected about 2 percent of facility expenses in user unlikely that PNG will be able to meet the need for fees. Level 3 and 4 public facilities collected less than doctors at lower-level facilities (levels 3 and 4) within 0.3 percent in user fees. It is possible that facilities a reasonable period of time, and policy makers should underreport user fee collection. If the user fees are carefully consider how to ensure access to care in the that low, especially for level 3 and 4 public facilities, it absence of trained doctors. Policy makers will also may not be worth collecting the fees because they do have to decide how to increase the supply of nurses not represent a substantial stream of operating rev- and other allied health workers. Over the years, the enue for the facility but may represent a substantial supply of these personnel has consistently declined, barrier to care for the poor population. implying that the human resource situation will continue to worsen, especially at lower-level facili- Policy Recommendations ties. The fact that facilities consistently have unfilled on Key Priority Service Areas positions, even for unskilled or semiskilled support staff, indicates a systemwide problem of personnel The provision of antenatal care in all health facilities management. needs to be prioritized. Maternal and neonatal health In-service training on priority health care ser- is one of the priority health areas for the health system vices needs to be strengthened and expanded for all of PNG because of the high fertility rate and relatively health cadres. Evaluation of the knowledge of health high infant and maternal mortality. It appears that care providers about different types of priority health PNG has been successful in ensuring supply of basic care services showed an interesting pattern. The sur- instruments and supplies needed for maternity care. vey asked a number of knowledge questions related The items needed for antenatal care (ANC) should be to child health care, maternity care, and sexually available in all the level 3 to 7 facilities, but the index of transmitted infections (STIs). Maternal care-related ANC items was 71 percent, 82 percent, and 81 percent knowledge was low for all provider types, including for level 3 and 4 public facilities, level 3 and 4 church- physicians. On average, the correct responses were run facilities, and level 5 and 6 public facilities, respec- only slightly above 50 percent. Similar results were tively. Some very basic ANC instruments and supplies found for child care services. However, the knowl- were missing or not functional, including weighing edge indexes were found to be relatively high for STIs, scales and tape measures. Inability of a health facility, between 80 and 90 percent. Given that these are pri- even a primary health center, to provide ANC is an ority health services, the knowledge indexes of health important concern for any health care system. Health Policy Implications 79 policy makers will have to ensure the availability of the operation theaters for conducting surgeries on a ANC in all facilities. regular basis. Nonuse of the facilities highlights the Child immunization services must be strength- need to ensure that resources, such as skilled person- ened. All the required childhood vaccination supplies nel and other instruments and supplies, are available. (vaccines) were not available in most of the facilities. Another similar issue is the availability of shelving for The most critical shortage appears to be with the storing drugs. Most facilities reported having adequate bacillus Calmette-Guérin vaccine. The child immu- space for storing drugs but some of the drug boxes nization program is highly cost-effective and so the were placed on the floor because of the lack of shelving. system should be able to improve its efficiency and Policy makers should ensure that all complementary effectiveness by ensuring the timely vaccination of inputs are supplied to use available resources effectively. children. The supply chain needs to be strengthened From an allocative efficiency perspective, the and the immunization program should develop an range of important medical care services in level 3 information system to identify locations with short- and 4 facilities needs to be expanded. Several impor- ages and excess supplies. tant medical care services are not available in most All facilities at level 3 or above should be able level 3 and 4 facilities, including laboratory and diag- to provide obstetric first aid and basic emergency nostic services and pharmacy. Availability of such ser- obstetric care (EmOC). Most facilities should also be vices at lower levels can improve the overall efficiency able to provide comprehensive EmOC. The survey of and quality of service delivery. Health facilities, espe- facilities indicated that 20 percent of facilities at levels cially the ones at levels 3 and 4, should be equipped 3 and 4 were not prepared to provide even obstetric with diagnostic services and essential pharmaceuticals first aid, and 40 percent were not ready to provide basic to improve the quality and safety of services provided. EmOC. Given the high fertility rate in PNG, the lack The range of health function tests at level 3 and 4 of obstetric first aid and basic EmOC at level 3 and 4 facilities could be broadened to include simple urine, facilities is a major concern. The health sector of PNG stool, and blood glucose tests. More than half of level 3 should carefully evaluate the gaps in EmOC at level 3 and 4 church-run facilities could perform malaria, to 7 facilities so that the system can be strengthened. human immunodeficiency virus (HIV), and tubercu- The PNG health system should better integrate losis (TB) tests, but less than half could conduct sim- outreach patrol planning into the health sector pler tests such as urine, stool, and blood glucose tests. planning process to consistently reach segments This situation probably indicates a national emphasis of the population in remote locations. The popu- on malaria, HIV, and TB, but the emphasis may be at lation of PNG is scattered over many islands, many the expense of other very basic but important tests. of which are not easily accessible. Making health The government needs to identify and resolve care available to the entire population will require a impediments to the provision of health care ser- well-structured outreach patrol or activities. Some vices at level 3 and 4 facilities. Bed occupancy rates, facilities are able to reach remote population groups a measure of health facility use, are very low in level 3 through outreach patrols. Facilities mentioned lack and 4 facilities, both public and church-run. Low uti- of funding, lack of fuel for transportation, and lack lization rates are likely a reflection of nonavailability of personnel as important factors that limited the of complementary inputs. Since level 3 and 4 facili- number of outreach patrols conducted. If national ties should provide more comprehensive health care and subnational goals on the number and types of services, policy makers should identify the resources patrols are determined, it will be possible to estimate in short supply. Without basic amenities such as water the resource requirements and improve access to and electricity, it is difficult to make health facilities health care for populations in remote locations. more effective in the provision of inpatient services. In addition to these basic amenities, facilities also lack skilled human resources, instruments, and supplies. Policy Recommendations The health information system needs to be to Improve Efficiency strengthened across all facility levels. Better data can Health facilities should use existing resources more lead to better decision making and better outcomes. efficiently. In the survey, it was found that although Many of the registries used for data collection by the about 30 percent of level 3 and 4 facilities had opera- health information system were not available at facil- tion theaters, less than half of the facilities actually use ity levels 3 to 6. Policy makers need to strengthen the 80 Service Delivery by Health Facilities in Papua New Guinea information system and ensure that different reports, well as strengthening health sector budget submis- forms, and registries are available in all facilities. sions. In addition, the TA offers a good opportunity to Data should be routinely used to generate evidence follow-up on the recommendations of the analytical for better decision making. work. There are four areas that have clear implications The government needs to improve the autonomy for PFM TA at the NDOH: of public health facilities to improve their efficiency. After examining the various indicators of health • Costing of health services. The report offers a facility performance and service delivery, the results detailed description of the outputs produced by indicate that the church-run facilities perform better the different facility levels. Combined with cost- than other types of health facility categories (level 3 ing data, this information can help assess the ade- and 4 public facilities and level 5 to 7 facilities). It is quacy of facility budgets and increase knowledge interesting that church-run facilities are not necessar- about funding needs in the front lines. ily the better-endowed facilities in relation to human • Efficiency analysis. Given the health sector’s finan- resource availability or availability of drugs and sup- cial constraints, an important component of the plies. Church-run facilities are also very active in try- TA focuses on achieving efficiency gains in health ing to provide coverage to population groups living in expenditure. The report offers a set of efficiency remote areas. The administrative structure of church- indicators that can potentially help the NDOH advance in the discussions around the facility’s run facilities and the degree of autonomy they have performance benchmarking. may improve the efficiency and effectiveness of these • Service delivery readiness. The report identifies key facilities compared with other facility types. NDOH service delivery elements that need to be available is exploring the facility-based budgeting (FBB) and for the effective provision of quality health ser- direct facility funding (DFF) as instruments to poten- vices, including outreach activities. This informa- tially improve the autonomy of and funding flow to tion can help the NDOH to develop a checklist for the public health facilities. supervision of the facility’s readiness. Readiness gaps can be monitored and incorporated into the Implications of Technical planning process. Assistance to NDOH • Health facilities’ autonomy. In the context of decen- tralization, the NDOH is exploring the potential for A junior public financial management (PFM) special- FBB and DFF. Understanding the facility’s desired ist provides imbedded technical assistance (TA) to level of autonomy, and identifying the areas where the NDOH. The main goals of this TA are to improve facilities would like to have more autonomy, will planning and prioritization of health expenditure, as help steer this process. REFERENCES Ascroft, J., R. Sweeney, M. Samei, I. Semos, and C. Morgan ———. 2016. 2015 Sector Performance Annual Review. Papua December 2011. “Strengthening Church and Government New Guinea: NDOH Government of Papua New Guinea. Partnerships for Primary Health Care Delivery in Papua ———. 2017. 2016 Annual Management Report. Papua New New Guinea: Lessons from the International Experience. Guinea: NDOH, Government of Papua New Guinea. Working Paper 16, Nossal Institute for Global Health, Pabón Lasso, H. 1986. “Evaluating hospital performance University of Melbourne, Australia. through simultaneous application of several indica- Barnum, H., and J. Kutzin. 1993. Public Hospitals in tors.” Bulletin of the Pan American Health Organization Developing Countries: Resource Use, Cost and Financing. (PAHO) 20 (4): 341–57. Baltimore and London: The Johns Hopkins University Tauchmann, H. 2012. “Partial Frontier Efficiency Analysis.” Press. The Stata Journal 12 (3): 461–78. Cairns, A., and X. Hou. 2015. Financing the Frontline in WHO (World Health Organization). 2007. Everybody’s Papua New Guinea: An Analytical Review of Provincial Business: Strengthening Health Systems to Improve Health Administrations’ Rural Health Expenditure 2006–2012. Outcomes: WHO’s framework for action. Geneva: WHO. Health, Nutrition and Population Discussion Paper. WHO (World Health Organization) and NDOH (National Department of Health). 2012. Health Service Delivery World Bank, Washington, DC. © World Bank. https:// Profile: Papua New Guinea. Geneva: WHO. openknowledge.worldbank.org/handle/10986/24075 World Bank. 2013. Below the Glass Floor: Analytical License: CC BY 3.0 IGO. Review of Expenditure by Provincial Administrations Government of Papua New Guinea. June 2010. National on Rural Health from Health Function Grants and Health Plan 2011–2020, Volume 1: Policies and Strategies. Provincial Internal Revenue. Washington, DC. © World Howes, S. M. 2014. A Lost Decade? Service Delivery and Bank. https://openknowledge.worldbank.org/handle/ Reforms in Papua New Guinea 2002–2012. Port Morseby 10986/17567 License: CC BY 3.0 IGO. and Canberra: National Research Institute of Papua New World Bank. (October 2013). Maternal Health Out-of- Guinea and Australian National University. Pocket Expenditure and Service Readiness in Lao PDR. Irava, W., K. Barker, A. Somanathan, and X. Hou. 2015. A Washington, DC: World Bank. Snapshot of Health Equity in Papua New Guinea: An ———. 2014. Universal Maternal Health Coverage? Assessing Analysis of the 2010 Household Income and Expenditure the Readiness of Public Health Facilities to Provide Survey. Washington, DC: World Bank. Maternal Health Care in Indonesia. Washington, DC: Janovsky and Travis. 2007. Non-state providers of health World Bank. care in Papua New Guinea: Governance, stewardship ———. August 2014. Assessment of Health Financing and international support. Geneva: World Health Options—Papua New Guinea. Washington, DC: World Organization. Bank. NDOH (National Department of Health). 2012. Papua New Yip, W., and R. Hafez. 2015. Reforms for Improving the Guinea Medical and Dental Catalogue, 10th Edition. Port Efficiency of Health Systems: Lessons from 10 Country Morseby: NDOH. Cases. Geneva: WHO. 81 Annex A WORLD HEALTH ORGANIZATION BUILDING BLOCK FRAMEWORK World Health Organization Building Block as well as organization and management approaches Framework for Performance Measurement used in the process of producing health care services and products can be observed and measured and, The goals, objectives, and strategies defined by the therefore, building block-based indicators should be PNG health plan are fully consistent with the World relatively easy to measure and report. Although the Health Organization’s (WHO’s) health system objec- services produced by health facilities or a subsystem, tives and strategies. Therefore, the health system goals such as the hospital sector, contribute toward over- defined by the WHO can be linked to the specific all improvements in health system outcomes, the objectives and strategies that PNG has developed final health-related outcomes cannot be measured in its 2011–2020 National Health Plan (NHP). The when the information on subsystems is collected WHO health system framework defines four overall from specific institutions and entities. For example, goals and outcomes of any health system. These goals improved health outcome of the population resulting are: (1) improved health, (2) responsiveness of the from improved hospital service provision cannot be system, (3) social and financial risk protection, and directly measured through hospital surveys. At the (4) improved efficiency. In a recent book, WHO facility level, the outcomes observed are not the final emphasized the importance of improving health sys- health outcomes; they are simply the outputs pro- tem efficiency to achieve universal coverage consistent duced by the health service production process and with the building block approach (Yip and Hafez, so the facility level outcomes are basically outputs or 2015). Considering the importance of access, cover- services produced (quantities and qualities of services age, quality, and safety, the framework outlines a produced or offered). number of building blocks for any health system. These building blocks are defined as the components Indicators of Performance of Health of health system strengthening, as described in the NHP 2011–2020. The WHO framework is reproduced Facilities Based on WHO Framework in figure A.1 (World Health Organization, 2007) . The framework used in this analysis identifies spe- The WHO’s building blocks of health systems cific indictors for each of the building blocks of the was developed to evaluate the performance of the health system (primary facilities and hospitals in this whole system rather than a component of the system. analysis) as well as the facility-level outputs and out- Because the purpose of this study was to examine the comes. For each of the six building blocks and health functionality of PNG’s upper-level primary and sec- facility outcomes, specific and reliable indicators or ondary health care system, the health system building measures were identified and used. Therefore, the blocks and outcomes were redefined to focus on the “output” block of the WHO building blocks in this target health care facilities. Building block-based eval- analysis focused on the outputs produced by the uation, which focuses on system inputs, can be car- health facilities. All other building blocks relate to dif- ried out both at subsectoral and facility levels. Inputs ferent types of inputs and resources used, including 83 84 Service Delivery by Health Facilities in Papua New Guinea Figure A.1  WHO System Building Blocks and the Overall Health System Goals and Outcomes SYSTEM BUILDING BLOCKS OVERALL GOALS / OUTCOMES SERVICE DELIVERY HEALTH WORKFORCE ACCESS IMPROVED HEALTH (LEVEL AND EQUITY) COVERAGE INFORMATION RESPONSIVENESS MEDICAL PRODUCTS, VACCINES & TECHNOLOGIES SOCIAL AND FINANCIAL RISK PROTECTION QUALITY SAFETY IMPROVED EFFICIENCY FINANCING LEADERSHIP / GOVERNANCE management resources, in the process of producing measures refer to the financial risks associated with hospital outputs. the health system in general and, therefore, health Five of the WHO building blocks were considered facility data, by itself, will not indicate fairness in relevant for evaluation of facility-level performance. financing. It is possible, however, to estimate the level The service delivery component was considered “the of payments that patients make to obtain health care outcome” of the system. The WHO framework attri- services at different levels. Finally, improved efficiency butes (access, coverage, quality, and safety) were mea- can be measured for any subsector of the health sys- sured using the information collected through health tem. In the WHO framework, the approach used for facility surveys. Again, the measurement of these efficiency measurement was to estimate the efficiency attributes was indirect, providing some information scores by defining aggregate health expenditures as the on how the hospital sector organization may affect measure of health sector inputs and population health these attributes rather than indicating system-level as health outcome. At the facility level, information on results. At the health facility level, some of the out- specific inputs used in the production process can be come measures defined by the WHO framework can collected and it should be possible to estimate quanti- be observed. For example, information on responsive- ties of various health services produced by using the ness of health facilities to the nonmedical needs of the specific inputs used in the production process. This clients can be obtained by examining patient satisfac- type of empirical modeling will be able to indicate rel- tion or other aspects of patient experiences with the ative technical efficiency of health facilities rather than health facilities. Social and financial risk-protection allocative efficiency. Annex B PROVIDER INTERVIEW RESULTS The results of the provider interviews are summa- Demographic Characteristics of Health rized in the main report. In this annex, the results Care Providers Surveyed are reproduced with detailed explanations so that the readers will be able to understand how different indi- Table B.2 reports the gender of health care providers cators (provider knowledge and satisfaction) were interviewed by health care provider type and level of measured. health facility. In level 3 and 4 facilities, only 7 percent of female MOs were interviewed, whereas in levels 5 to 7, 29 percent of MOs interviewed were female. Methodology of Selecting At level 3 and 4 facilities, 75 percent of NOs and Health Care Providers 65 percent of CHWs interviewed were female, but The health care provider survey was intended to these percentages were 91 percent and 74 percent interview a minimum of 400 providers from level 3 in level 5 to 7 facilities. The percentage of female to 7 facilities. The provider interviews included health care providers interviewed in level 5 to 7 only the health care providers directly involved in facilities was higher than the percentage in level 3 the provision of clinical services. The survey design and 4 facilities. required interviewing up to three medical officers Table B.3 shows the age distribution of health (MOs) per facility (since MOs were unlikely to be care providers interviewed by facility level. More present in most of the level 3 and 4 facilities, less than 60 percent of MOs interviewed in level 3 and 4 than three were interviewed if there were less than facilities were 40 years old or less, while 90 percent of three MOs in a facility), one health extension offi- the MOs interviewed in level 5 to 7 facilities were in cer (HEO), two nursing officers (NOs), and two the same age group. More than half of the NOs and community health workers (CHWs). Participant CHWs interviewed were 41 years of age or older selection within each cadre was purposive with in level 3 and 4 facilities. About 54 percent of NOs the overarching aim of achieving a minimum of interviewed in level 5 to 7 facilities were 40 years 50 interviews with MOs, 50 with HEOs, 150 with or less. NOs, and 150 with CHWs. Eighty-seven percent of the health care providers Table B.1 reports the number of providers sur- interviewed were married, 89 percent had children, veyed by provider-type and level of health care and 50 percent of the providers interviewed men- facilities. The number of interviews planned is also tioned that their place of work was their home area reported in the table to indicate how the surveyed or home district (53 percent for level 3 and 4 facili- numbers deviated from the numbers interviewed. ties and 42 percent for level 5 to 7 facilities). About The lower number of providers interviewed is mainly 83 percent of providers in level 3 and 4 facilities due to inability to conduct the surveys in some facili- reported living on facility campus or within 20 to ties as well as non-availability of clinical personnel 30 minutes from the facility. For level 5 to 7 facilities, during the visit days. the percentage of providers interviewed living on 85 86 Service Delivery by Health Facilities in Papua New Guinea Table B.1  PNG Health Care Providers Surveyed, by Provider Type and Facility Level Providers working Providers working in levels 3 and 4 in levels 5 to 7 Total Number Number Number Number number of Providertype (cadre) planned surveyed planned surveyed providers Medical officer 14 41 55 Health extension officer 31 16 47 Nursing officer 100 35 135 Community health worker 115 34 149 Total 250 260 150 126 386 Note: Number of facilities surveyed was 73. The plan was to survey 79 facilities but not all selected facilities could be surveyed. Table B.2  Distribution of PNG Health Care Providers Interviewed, by Gender and Provider Type Level 3 and 4 facilities Level 5 to 7 facilities Provider Total No. of Percent Total No. of Percent type interviewed females female interviewed females female MO 14  1  7.14 41 12 29.27 HEO 31 12 38.71 16 10 62.50 NO 100 75 75.00 35 32 91.43 CHW 115 75 65.22 34 25 73.53 Table B.3  Age Distribution of Interviewed PNG Health Care Providers, by Provider Type Level 3 and 4 facilities Level 5 to 7 facilities Provider Age Age Age Age Age Age Age Age type Ä30 yr 31–40 yr 41–50 yr ê51 yr Ä30 yr 31–40 yr 41–50 yr ê51 yr MO 28.57 35.71 21.43 14.29 43.90 46.34 9.76  0.00 HEO 25.81 38.71 16.13 19.35 12.50 56.25 18.75 12.50 NO 12.00 33.00 34.00 21.00 28.57 25.71 28.57 17.14 CHW 11.30 20.00 47.83 20.87 8.82 35.29 35.29 20.59 campus or nearby was only 48 percent. Thirty-two per- level 3 and 4 facilities were consultation services pro- cent of providers interviewed lived more than 1 hour vided to adults and inpatient services. Out of a total of away from the facility in the level 5 to 7 facilities, 14 MOs interviewed, 12 reported providing consulta- wheras only 9.6 percent lived 1 hour away in level 3 tion to adults and inpatients. If we consider allocation and 4 facilities. of more than 6 hours, on average, on various activities, In terms of years of experience of the health care the modal service type for the MOs was inpatient ser- providers interviewed, about 39 percent reported vices. More than half of the MOs interviewed reported working for 25 years or more. Another 39 percent of working on the following activities for more than the health care providers interviewed had been work- 6 hours a week, on average, over the past 3 months: ing in health facilities for less than 5 years. About inpatient care, adult patient consultation, tuberculosis two-thirds reported that their work experience was (TB) treatment and diagnosis, and malaria treatment. less than 15 years. The principal activities of HEOs at level 3 and 4 facili- ties appear to be consultation of children, consultation of adults, malaria treatment, and inpatient care. For Type of Services Provided NOs, the most frequent activities are consultation of by Health Care Providers adults and children, inpatient care, and malaria treat- Table B.4 reports the number of providers providing ment. Finally, for CHWs at this level, the important different types of services within the past 3 months activities are consultation of children and adults, and whether or not the provider spent more than malaria treatment, vaccination, and inpatient care. 6 hours per week on the specific activity in level 3 and Table B.5 reports the number of providers active 4 facilities. The modal service types for physicians in in the provision of services at the facilities in levels Provider Interview Results 87 Table B.4  Types of Primary Health Care Provider Services in Level 3 and 4 Facilities, by Provider Type MOs HEOs NOs CHWs Provide Provide Provide Provide this this this this Service type service >6 hr/wk service >6 hr/wk service >6 hr/wk service >6 hr/wk Supervise nurse, CHW, etc. 8 7 26 19 72 50 25 13 Consultation: children 10 7 30 25 90 80 103 88 Consultation: adults 12 8 30 23 95 87 107 91 Family planning 8 2 23 13 72 48 78 50 Antenatal care 5 1 25 16 70 50 70 54 Postnatal care 6 1 25 17 69 49 72 52 Surgery 7 7 21 17 33 14 32 22 Inpatient care 12 11 28 24 86 80 89 78 Delivery: in facility 5 4 26 16 72 58 82 63 Delivery: home 1 1 7 6 26 13 29 14 TB treatment and diagnosis 11 8 24 16 71 43 77 47 Vaccination 4 1 22 15 78 59 94 72 Malaria treatment 11 8 30 18 88 70 100 77 Nutrition 10 6 19 13 66 44 65 37 Other outreach active 6 4 15 9 56 34 68 36 CHW training 3 2 18 10 49 34 35 17 Other training 2 1 5 3 8 6 6 3 Other activitiesa 1 1 3 3 4 3 3 2 Total number surveyed 14 14 31 31 100 100 115 115 a. Other activities include administrative work, teaching or training other staff, supervising nursing students, health education, attending or offering in-service courses. Table B.5  Types of Primary Health Care Provider Services in Level 5 to 7 Facilities, by Provider Type MOs HEOs Nos CHWs Provide Provide Provide Provide this this this this >6 hr/ Service type service >6 hr/wk service >6 hr/wk service >6 hr/wk service wk Supervise nurse, CHW, etc 30 24 14 9 24 21 4 2 Consultation: children 29 28 12 12 23 22 23 20 Consultation: adults 33 31 15 13 19 19 16 13 Family planning 13 7 5 4 9 7 12 8 Antenatal care 11 6 4 3 4 1 5 4 Postnatal care 11 9 3 2 4 1 5 5 Surgery 22 19 7 6 6 5 8 8 Inpatient care 35 34 14 12 28 25 28 26 Delivery: in facility 8 6 2 2 11 7 8 5 Delivery: home 2 2 0 0 2 0 2 2 TB treatment/ diagnosis 27 19 15 13 23 14 18 12 Vaccination 9 5 5 3 14 7 15 8 Malaria treatment 25 19 16 15 20 15 23 18 Nutrition 19 15 9 6 16 13 17 12 Other outreach active 14 8 2 2 4 1 5 3 CHW training 9 7 5 3 12 8 9 5 Other training 3 3 2 2 2 1 1 0 Other activitiesa 4 4 2 2 2 1 1 1 Total number surveyed 41 41 16 16 35 35 34 34 a. Other activities include administrative work, teaching or training other staff, supervising nursing students, health education, attending or offering in-service courses. 88 Service Delivery by Health Facilities in Papua New Guinea 5 to 7. The principal activities of MOs at levels 5 to Table B.6  Percentage of PNG Health Care 7 are adult consultations, inpatient care, supervision Providers Reporting No Training in Previous 3 Years, of nurses, consultation service for children, and TB by Facility Level treatment and diagnosis. For the HEOs, the main Provider Levels 3 and 4, Levels 3 and 4 activities are malaria treatment, TB treatment and type public church Levels 5 to 7 diagnosis, inpatient care, adult and child consulta- MO 50.00 25.00 43.90 HEO 50.00 11.11 37.50 tions, and supervision of nurses or CHWs. NOs are NO 26.92 22.92 37.14 mainly involved in inpatient care, TB treatment and CHW 37.10 35.85 67.65 diagnosis, supervision of nurses and CHWs, con- sultation of children, and malaria treatment. The CHWs are involved in activities such as inpatient training was on infectious diseases followed by care, malaria treatment, consultation for children, child health services. TB treatment and diagnosis, nutrition-related activi- ties, and vaccinations. Health Care Provider Satisfaction Staff Trainings Using the questions 401 to 420 in the questionnaire, health care provider’s satisfaction scores were calcu- The percentage of providers who reported having lated. The satisfaction score or index was calculated not received any training in the previous 3 years is by assigning a value of 1.0 if the provider indicated reported in table B.6. Note that half of the doctors that she or he is generally satisfied with a specific and HEOs in level 3 and 4 public facilities did not situation, 0.5 for neutral, and 0 for dissatisfied. If the receive any training at all in the previous 3 years. For values for all the 20 aspects or situations are added church-run facilities, a significantly higher propor- together, the overall satisfaction score or index for tion received training. the provider can be derived. Since there are some Those who reported receiving training were asked missing values and not all 20 aspects may have valid about the types of training received. Table B.7 reports responses, satisfaction score was defined as total the types of training mentioned by the health satisfaction score divided by the maximum satis- care providers during the interview. For all clini- faction score possible for all valid responses (i.e., cal personnel listed in the table, the most common excluding the missing values). Table B.7  PNG Health Care Provider Training Categories, by Subject Matter Community Provider type and Child Infectious Family Maternal General health Other facility level health diseases planning health training training traininga MO Level 3 and 4, public 20.00 50.00 0 0 20.00 0 0 Level 3 and 4, church 25.00 50.00 0 25.00 0 0 50.00 Level 5 to 7 9.76 19.51 0 7.32 9.76 2.44 31.71 HEO Level 3 and 4, public 31.82 45.45 4.55 18.18 18.18 0 13.64 Level 3 and 4, church 22.22 55.56 33.33 22.22 55.56 0 66.67 Level 5 to 7 18.75 31.25 0 6.25 0 0 25.00 NO Level 3 and 4, public 26.92 51.92 21.15 19.23 15.38 1.92 15.38 Level 3 and 4, church 20.83 58.33 20.83 14.58 18.75 4.17 18.75 Level 5 to 7 20.00 20.00 5.71 5.71 14.29 8.57 25.71 CHW Level 3 and 4, public 22.58 54.84 12.90 14.52 8.06 9.68 6.45 Level 3 and 4, church 18.87 50.94 9.43 11.32 5.66 1.89 11.32 Level 5 to 7 14.71 20.59 2.94 2.94 2.94 2.94 14.71 a. Other training includes cold-chain management, accident and emergency, burn management, drug management, mental health, lab management, snakebite management, surgical, and ultrasound. Provider Interview Results 89 Table B.8  PNG Health Care Providers’ Satisfaction Scores, by Facility Type and Health Care Provider Type Provider type and Average satisfaction Standard deviation facility level score (M) (SD) Minimum value Maximum value MO   Level 3 and 4, public 0.61 0.17 0.40 0.85   Level 3 and 4, church 0.74 0.18 0.55 0.90   Level 5 to 7, public 0.62 0.15 0.30 0.88 HEO   Level 3 and 4, public 0.55 0.18 0.12 0.90   Level 3 and 4, church 0.61 0.19 0.38 1.00   Level 5 to 7, public 0.61 0.21 0.18 1.00 NO   Level 3 and 4, public 0.55 0.16 0.18 0.83   Level 3 and 4, church 0.57 0.17 0.20 1.00   Level 5 to 7, public 0.61 0.18 0.15 1.00 CHW   Level 3 and 4, public 0.57 0.19 0.15 0.93   Level 3 and 4, church 0.60 0.14 0.33 0.90   Level 5 to 7, public 0.58 0.18 0.25 0.90 Average satisfaction scores by facility type and and provider-type. The responses vary across facility- provider type are shown in table B.8. The maximum type as well as provider-type. For example, 40 per- possible value of satisfaction score, by construction, cent of MOs in level 3 and 4 public health facilities is 1.0 and the lowest possible value is 0. mentioned availability of medicines in the facility as During the interviews with health care providers important while none of the MOs in level 3 and 4 they were asked to report the factors they feel are church-run facilities mentioned this as an impor- most important in affecting their level of satisfac- tant factor. Among the MOs in church-run facili- tion. A number of possible factors are listed in the ties, the most important factor mentioned was the questionnaire but the providers were encouraged physical condition of the health facilities and avail- to mention any factor or situation they consider ability of living quarters. important. Table B.9 reports the responses of the During the provider interviews, the health care pro- health care providers interviewed by facility-type viders were asked to report the three most important Table B.9  Factors Affecting PNG Health Care Provider Satisfaction, by Provider Type and Situation and Aspect Type Physical Provider type and Availability Availability condition Training Living facility level of medicine of equipment of facility opportunities Salary quarters Others MO   Level 3 and 4, public 40.00 30.00 0 0 0 0 30.00   Level 3 and 4, church 0 0 50 0 0 25 25.00   Level 5 to 7 22.50 27.50 5.00 10.00 5.00 0 30.00 HEO   Level 3 and 4, public 13.64 31.82 4.55 0 0 13.64 36.35   Level 3 and 4, church 11.11 33.33 0 11.11 0 11.11 33.34   Level 5 to 7 37.50 0 6.25 12.5 6.25 12.50 25.00 NO   Level 3 and 4, public 15.38 19.23 3.85 13.46 3.85 9.62 34.61   Level 3 and 4, church 12.50 14.58 6.25 6.25 18.75 8.33 33.34   Level 5 to 7 25.71 8.57 5.71 8.57 8.57 14.29 28.58 CHW   Level 3 and 4, public 16.13 19.35 11.29 4.84 9.68 12.90 25.81   Level 3 and 4, church 7.69 11.54 23.08 9.62 21.15 9.62 17.30   Level 5 to 7 23.53 5.88 0 2.94 14.71 17.65 35.29 90 Service Delivery by Health Facilities in Papua New Guinea factors affecting their satisfaction at work with the “availability of medicines and equipment to do your facilities. The three most important factors mentioned job” followed by “salary/income.” by the provider types were ranked using the frequen- cies of responses received. Table B.10 reports the top Motivation of Health Care Providers 10 aspects as reported by the health care providers with percent of providers mentioning the aspect as Health care providers were asked a number of ques- one of the top three factors. tions to understand their degree of motivation to Average index for health care provider priorities work in the facilities. The motivation questions can by priority category was calculated by assigning the be grouped into a number of subcategories, such as response “important” as 1, “neutral’ as 0.5, and “unim- index of pride and index of financial reward. The portant” as 0. Almost all providers interviewed con- motivation indexes were calculated by assigning the sidered all the seven aspects mentioned as important value of 1 to the highest level of motivation and 0 to (salary/income, respect and collaborations received the lowest possible level for motivation. The average from colleagues, respect shown by community mem- indexes are reported in table B.11 by provider type bers, availability of medicines and equipment to do and health facility type. your job, living arrangement for your family, oppor- tunities for a better paying job, and access to schools for your children). Since there are almost no varia- Salary and Benefits of Providers tions across the providers by provider type and facil- Salary and benefits are an important aspect of any ity type, provider priority questions did not identify employment. The health care providers were asked any specific pattern useful for policy making. about their salary and allowance levels as well as Concerning the most important priority among whether they are involved in income-generating the seven priorities mentioned, most providers chose activities outside the health facility employment. Agriculture and trade and business are the most important sources of income other than facility Table B.10  Top Three Factor Categories Affecting income. Of the respondents who reported secondary PNG Health Care Provider Satisfaction incomes, 79 percent reported receiving income from Aspects or factors these two sources. most important (three most important factors) MO HEO NO CHW 406 Availability of 70.91 68.09 54.81 50.34 equipment in the health Table B.11  Index of Motivation, by Provider facility and Facility Types 405 Availability of 63.64 34.04 38.52 33.56 medicines in the health Provider type facility Facility level MO HEO NO CHW 410 Your training 21.82 23.40 31.11 32.89 Pride index opportunities to upgrade your skills and knowledge Public, level 3 and 4 0.688 0.716 0.719 0.804 Church, level 3 and 4 0.875 0.722 0.823 0.844 416 Living accommodations 10.91 29.79 29.63 31.54 Public, level 5 to 7 0.616 0.695 0.718 0.750 for your family Financial reward index 407 The physical condition 23.64 21.28 25.19 30.87 of the health facility Public, level 3 and 4 0.650 0.532 0.606 0.611 building Church, level 3 and 4 0.850 0.456 0.517 0.462 Public, level 5 to 7 0.541 0.575 0.623 0.579 412 Your salary 10.91 19.15 28.89 26.85 408 Your ability to provide 14.55 14.89 9.63 12.08 Self-efficacy index high quality of care Public, level 3 and 4 0.790 0.755 0.790 0.842 403 Management of the 14.55 14.89 9.63 7.38 Church, level 3 0.950 0.744 0.804 0.791 health facility and 4 Public, level 5 to 7 0.751 0.738 0.769 0.735 402 Working relationships 9.09 8.51 14.07 14.07 with upper-level staff Index of conscientiousness 413 Employment benefits 3.64 14.89 11.85 13.42 Public, level 3 and 4 0.867 0.845 0.878 0.876 (travel allowance, Church, level 3 and 4 0.938 0.824 0.865 0.829 bonus, etc.) Public, level 5 to 7 0.839 0.818 0.867 0.850 Provider Interview Results 91 Table B.12  Percent of PNG Health Care Providers Table B.13  Knowledge Index Related to Child Health Involved in the Provision of Priority Health Services Services, by Provider Type Percent of providers reported providing Index of Index of the following services: knowledge related knowledge related Provider Child Maternal STD to danger signs to fast breathing type health health treatment Provider Level Level 3 Level 5 Level 3 MO 65.45 43.64 62.96 type 5 to 7 and 4 to 7 and 4 HEO 95.56 61.70 72.34 MO 0.458 0.564 0.440 0.600 NO 91.34 72.39 71.85 HEO 0.486 0.529 0.643 0.529 CHW 85.31 63.76 63.76 NO 0.469 0.482 0.496 0.540 All 85.68 63.64 67.53 CHW 0.538 0.474 0.570 0.521 Note: Among the providers interviewed during the survey. Provider Knowledge Index two errors = 0.6, three errors = 0.4, four errors = 0.2, five or six errors = 0. Percent of providers providing child health services, On maternal health services, the questionnaire maternal health services, and sexually transmitted included a number of questions to assess the knowl- disease (STD) services by provider type is reported in edge of health care providers on maternal health table B.12. Note that a high proportion of health care services. Only health care providers who indicated providers reported being involved in the provision of that they provide maternal health services on a regu- these services. lar basis were asked these questions. Therefore, the The providers involved with the provision of child questions assess the knowledge of health care provid- health services were asked a series of questions to ers who were involved in the provision of maternal evaluate their knowledge related to important child health services. health concerns. Two aspects of child health service Below, all the questions related to maternal health are reported in table B.13. The indexes were calcu- service are reproduced with the system of scoring lated using the following scoring system for danger used for obtaining the specific and overall knowledge signs and fast breathing: no errors = 1, one error = 0.8, indexes. Responses used to calculate the knowledge index 609 Which of the following are danger-signs in pregnancy? Please read the choices and circle the answer Given: Correct response as indicated below gets 1.0, wrong response is assigned a value of zero. a. Swelling of hands, feet and face No b. Vaginal bleeding Yes c. Pain in Joints No d. Severe abdominal pain Yes 610 A woman in her late pregnancy showed up in the outpatient clinic or emergency room of a provincial hospital. Her principal complaint is vaginal bleeding. Which of the following actions would you take as the healthcare provider? [Please read the choices] . Correct response as indicated below is assigned a value of 1.0, incorrect response gets zero. a.  Conduct a quick sterile vaginal examination (with speculum) Yes b.  Suspect ectopic pregnancy and prepare for surgery No c.  Insert an IV line Yes d.  Give 4.3% Dextrose/Saline as a bolus fluid dose Yes 92 Service Delivery by Health Facilities in Papua New Guinea 611 A pregnant woman in her third trimester is having convulsions or has become unconscious. Which of the following actions would you take as the health care provider? [Please read the choices] a.  Transfer the woman to labor ward Yes b.  When convulsion ends, help the woman onto her right side No c.  Give magnesium sulfate Yes d.  Manage airways Yes e.   Give Diazepam IV Yes f.  Insert IV line and give fluids rapidly Yes 612 What are the danger signs at delivery meaning that a woman should be given immediate medical attention at a provincial hospital? [Please read the choices] a.  The woman loses consciousness (faints) or has fits (convulsions) Yes b.  The bag of water breaks, and labor starts within 12 hours No c. Strong labor (contractions) last for 12 hours but the baby Yes has not been delivered d.  A foot or arm can be seen coming out of the cervix Yes e. Meconium +++ is seen after the water breaks and the head is visible Yes f.  The women loses 500 milliliters of blood during birthing Yes g.  The placenta does not come out within 30 minutes of delivery Yes 613 What are some examples of direct maternal obstetric complications? [Please DO NOT read the choices; circle “1” if it is mentioned or “2” if it is NOT mentioned] a.  Antepartum and postpartum hemorrhage Mentioned b. Pre-eclampsia Mentioned c. Eclampsia Mentioned d.  Obstructed labor Mentioned e.  Other, specify 614 What does “Maternal Death” mean? [Please DO NOT read the choices. Record answer] Death of a woman within 6 months of delivery Not correct Death of a woman while pregnant or within 42 days of delivery Correct   caused or made worse by the pregnancy Death of a pregnant woman from any cause Not correct Death of a woman within 80 days of delivery Not correct Other, specify Review if correct Don’t know. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Not correct The results of the scoring are presented below. The summary table is included in the text of the report. Average score for q609 items Mean SE [95% Conf. interval] Doctor Level 3 and 4 facilities 0.583 0.053 0.479 0.687 Doctor: Level 5 to 7 facilities 0.643 0.034 0.575 0.710 HEO: Level 3 and 4 facilities 0.550 0.035 0.480 0.620 HEO: Level 5 to 7 facilities 0.500 0.000 . . Nurse: Level 3 and 4 facilities 0.601 0.017 0.567 0.635 Nurse: Level 5 to 7 facilities 0.636 0.039 0.559 0.714 CHW: Level 3 and 4 facilities 0.573 0.019 0.535 0.611 CHW: Level 5 to 7 facilities 0.556 0.037 0.483 0.628 Provider Interview Results 93 Average score for q610 items Mean SE [95% Conf. interval] Doctor Level 3 and 4 facilities 0.458 0.100 0.261 0.656 Doctor: Level 5 to 7 facilities 0.429 0.049 0.333 0.524 HEO: Level 3 and 4 facilities 0.630 0.041 0.549 0.711 HEO: Level 5 to 7 facilities 0.563 0.063 0.439 0.686 Nurse: Level 3 and 4 facilities 0.640 0.022 0.597 0.683 Nurse: Level 5 to 7 facilities 0.455 0.045 0.365 0.544 CHW: Level 3 and 4 facilities 0.725 0.023 0.680 0.770 CHW: Level 5 to 7 facilities 0.667 0.042 0.585 0.749 Average scores of q611 items Mean SE [95% Conf. interval] Doctor Level 3 and 4 facilities 0.611 0.035 0.542 0.680 Doctor: Level 5 to 7 facilities 0.583 0.052 0.482 0.685 HEO: Level 3 and 4 facilities 0.660 0.033 0.596 0.724 HEO: Level 5 to 7 facilities 0.708 0.042 0.626 0.790 Nurse: Level 3 and 4 facilities 0.677 0.019 0.639 0.714 Nurse: Level 5 to 7 facilities 0.652 0.047 0.558 0.745 CHW: Level 3 and 4 facilities 0.650 0.021 0.609 0.691 CHW: Level 5 to 7 facilities 0.704 0.072 0.561 0.846 Average score for q612 items Mean SE [95% Conf. interval] Doctor Level 3 and 4 facilities 0.714 0.037 0.642 0.787 Doctor: Level 5 to 7 facilities 0.714 0.026 0.663 0.765 HEO: Level 3 and 4 facilities 0.686 0.026 0.634 0.737 HEO: Level 5 to 7 facilities 0.643 0.041 0.562 0.724 Nurse: Level 3 and 4 facilities 0.683 0.013 0.658 0.708 Nurse: Level 5 to 7 facilities 0.649 0.030 0.591 0.708 CHW: Level 3 and 4 facilities 0.685 0.011 0.664 0.706 CHW: Level 5 to 7 facilities 0.698 0.016 0.667 0.730 Average scores for q613 items Mean SE [95% Conf. interval] Doctor Level 3 and 4 facilities 0.833 0.083 0.669 0.998 Doctor: Level 5 to 7 facilities 0.839 0.072 0.697 0.982 HEO: Level 3 and 4 facilities 0.690 0.042 0.608 0.772 HEO: Level 5 to 7 facilities 0.688 0.188 0.318 1.057 Nurse: Level 3 and 4 facilities 0.555 0.035 0.486 0.624 Nurse: Level 5 to 7 facilities 0.477 0.092 0.296 0.659 CHW: Level 3 and 4 facilities 0.430 0.034 0.363 0.498 CHW: Level 5 to 7 facilities 0.528 0.077 0.375 0.680 Defining Maternal Mortality Mean SE [95% Conf. interval] Doctor Level 3 and 4 facilities 0.333 0.211 –0.082 0.749 Doctor: Level 5 to 7 facilities 0.357 0.133 0.095 0.619 HEO: Level 3 and 4 facilities 0.280 0.092 0.099 0.461 HEO: Level 5 to 7 facilities 0.750 0.250 0.257 1.243 Nurse: Level 3 and 4 facilities 0.268 0.049 0.171 0.365 Nurse: Level 5 to 7 facilities 0.364 0.152 0.064 0.663 CHW: Level 3 and 4 facilities 0.304 0.052 0.201 0.406 CHW: Level 5 to 7 facilities 0.000 Overall knowledge: Maternal H. Mean SE [95% Conf. interval] Doctor Level 3 and 4 facilities 0.589 0.054 0.482 0.695 Doctor: Level 5 to 7 facilities 0.594 0.029 0.537 0.651 HEO: Level 3 and 4 facilities 0.583 0.021 0.542 0.623 HEO: Level 5 to 7 facilities 0.642 0.047 0.550 0.734 Nurse: Level 3 and 4 facilities 0.571 0.013 0.545 0.597 Nurse: Level 5 to 7 facilities 0.539 0.043 0.454 0.624 CHW: Level 3 and 4 facilities 0.561 0.014 0.533 0.589 CHW: Level 5 to 7 facilities 0.525 0.028 0.470 0.580 94 Service Delivery by Health Facilities in Papua New Guinea Hospital Bed Occupancy Rate days by 3. Thus, the number of inpatient days is esti- The first indicator used is the hospital bed occu- mated as pancy rates. It is assumed that hospitals showing a lower rate of occupancy are less efficient than occupancy rate 2B = ( number of health facility days ‫ ء‬3) hospitals showing a higher level of bed occupancy. total number of beds This is because the “hospital bed” is one of the 100 × most important cost items in the health care sector 365 and hospital service provision requires the utiliza- All approaches discussed above provide different tion of beds. bed occupancy rates. Method 1, denoted above as Occupancy rates are found using the basic prin- occupancy rate 1, is the most direct method of deter- ciple: the rate at which beds are occupied. Using mining occupation rates, because the survey enu- the information collected from health facilities, bed merators collected the information on the day of the occupancy rates can be calculated in several different survey by noting the number of beds occupied. This ways. The first method, denoted as occupancy rate 1, information is likely to show lowest level of errors or is estimated by the number of occupied beds relative bias in reporting. The second approach, denoted as to total beds, at a certain point in time. Since the sur- occupancy rate 2A, days of hospital stay in December vey enumerators could actually observe the number 2014, which is close to the survey date depending on of beds occupied on the day of the survey, this is the the month in which the survey was actually done, most independent measure of bed occupancy rate was used. Misreporting or miscoding in the ward among all the different measures presented here. books (admission and discharge books) is always a Occupancy rate method 1: possibility as well as errors in transferring the infor- mation from the books to the questionnaires. number of occupied beds The third approach, denoted as occupancy rate 2B, occupancy rate1 = total number of beds calculated the occupancy rates using bed occupancy days in 4 different months distributed over 2014. The second estimate of bed occupancy, denoted as This approach should be able to take into account the occupancy rate 2, can be estimated from the total num- seasonality in the use of hospital beds. At the same ber of days per year that health facility beds were occu- time, the possibility of errors in the earlier method pied. That is, the number of bed days generated over also exists for this method. the year divided by the total number of beds per facility. A simple regression model was run to see if the bed Occupancy rate method 2: size of the hospital has any effect on the occupancy rate. In the regression model, the following variables number of health facility days were included: bed size (BED), whether the facility is occupancy rate 2 = total number of beds public (PUBLIC), and whether the facility is located in 100 an urban area (URBAN). The estimated regression is × 365 reported below. The numbers in the parentheses are the t-statistics associated with the coefficients estimated: From the data there are two approaches to deter- mine the number of bed days. Approach 2A estimates occupancy rate = 30.27 + 0.598 BED the number of days per year by multiplying the num- ( 2.18) ber of inpatient days in 1 month (in all instances, + 3.77 PUBLIC + 19.25 URBAN, R 2 = 0.302 ( 0.60) ( 2.65) December2014) by 12, whereby an estimate for the number of days is determined: Essentially, one wants to know if the number of beds in a health facility (BED) affects the occupancy occupancy rate 2A = rate of the health facility. In addition, the possible ( number of health facility days × 12) × 100 effects of health facility administration type (i.e., total number of beds 365 whether bed occupancy is affected by whether the facility is government-run [PUBLIC] or church-run Similarly, approach 2B estimates the number of and whether the health facility is located in an urban days per year by multiplying 4 months of inpatient area [URBAN] were explored. Annex C EXIT INTERVIEW OF PATIENTS OR CAREGIVERS Table C.1  Frequencies, by Level and Ownership Level 3 and 4 Level 3 and 4 Item public church Level 5 Level 6 Level 7 Total Female as the main health care 53.21% 63.13% 72.78% 86.96% 40% 63.71% provider who provided service Laboratory/diagnostic test being 12.82% 18.13% 30.38% 34.78% 40% 21.50% recommended/done today Medication prescribed today 95.51% 96.25% 99.37% 100% 100% 97.24% Medication or prescription received 53.85% 49.38% 50.94% 52.17% 60% 51.57% today Respondents reduced food 16.67% 23.75% 24.53% 17.39% 20% 21.46% consumed due to medical cost Respondents borrowed money to pay 10.90% 9.38% 10.06%  4.35% 10% 9.84% for medical care cost Respondents with low-income status 66.03% 72.50% 42.14% 30.43% 10% 57.87% Table C.2  Mean Number of Medicines Listed Table C.4  Mean Satisfaction Score, by Level in the Prescription, by Level and Ownership and Ownership [95% Conf. [95% Conf. Level/ownership Mean SE interval] Level/ownership Mean SE interval] Level 3 and 4 public 1.611 0.099 1.417 1.805 Level 3 and 4 public 58.8 1.2 56.4 61.1 Level 3 and 4 church 1.786 0.091 1.607 1.964 Level 3 and 4 church 70.5 1.1 68.3 72.6 Level 5 2.192 0.089 2.017 2.367 Level 5 65.4 1.2 63.1 67.8 Level 6 2.043 0.213 1.625 2.462 Level 6 65.9 2.1 61.8 70.0 Level 7 2.200 0.554 1.112 3.288 Level 7 76.6 3.0 70.7 82.6 Table C.3  Mean Medication Dispensed from Health Table C.5  Mean Overall Satisfaction, by Level Facilities, by Level and Ownership and Ownership [95% Conf. [95% Conf. Level/ownership Mean SE interval] Level/ownership Mean SE interval] Level 3 and 4 public 1.987 0.074 1.840 2.133 Level 3 and 4 public 62.7 1.7 59.3 66.0 Level 3 and 4 church 2.019 0.076 1.871 2.168 Level 3 and 4 church 76.2 1.5 73.3 79.1 Level 5 2.212 0.074 2.065 2.358 Level 5 72.6 1.4 69.8 75.3 Level 6 1.826 0.215 1.404 2.248 Level 6 63.5 2.1 59.3 67.7 Level 7 2.000 0.596 0.828 3.172 Level 7 83.5 4.4 74.8 92.2 95 96 Service Delivery by Health Facilities in Papua New Guinea Table C.6  Reasons for Seeking Treatment for Children, by Level and Ownership Level 3 and 4 Level 3 and 4 Item public church Level 5 Level 6 Level 7 Total Illness/symptoms 88.06% 91.03% 85.1% 81.8% 75% 87.6% Regular checkups 1.49% 0 0 0 0 0.43% Emergency/accident related 1.49% 1.28% 0 0 0 0.85% Immunization 1.49% 1.28% 5.41% 0 0 2.56% Other 7.47% 6.41% 9.46% 18.1% 25.0% 8.54%