Guinea-Bissau: Service Delivery Indicators Report – Health June 2019 Health Nutrition Population Global Practice Africa Region Document of the World Bank 1 Standard Disclaimer: This volume is a product of the staff of the International Bank for Reconstruction and Development/ The World Bank. The findings, interpretations, and conclusions expressed in this paper 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. Copyright Statement: The material in this publication is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law. The International Bank for Reconstruction and Development/ The World Bank encourages dissemination of its work and will normally grant permission to reproduce portions of the work promptly. For permission to photocopy or reprint any part of this work, please send a request with complete information to the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA, telephone 978-750-8400, fax 978-750-4470, http://www.copyright.com/. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street NW, Washington, DC 20433, USA, fax 202-522-2422, e-mail pubrights@worldbank.org. 2 LIST OF ABBREVIATIONS CFA West African Franc (Communaute Financiere Africaine franc) CMNND Communicable, Maternal, Neonatal, and Nutritional Diseases DALYs Disability-Adjusted Life Years DHS Demographic and Health Survey DRS Regional Health Directorate (Direcao Regional de Saude) GBD Global Burden of Disease GDP Gross Domestic Product HIV Human Immunodeficiency Virus ILAP Poverty Evaluation Household Survey (Inquérito Ligeiro para a Avaliação da Pobreza) IMR Infant Mortality Rate INASA National Institute of Public Health (Instituto Nacional de Saude) LIC Low-Income Countries LLIN Long Lasting Insecticide Nets MDG Millennium Development Goals MICS Multi Indicators Cluster Survey MINSAP Ministry of Public Health (Ministerio da Saude Publica) MMR Maternal mortality ratio NCD Non-Communicable Diseases NCDI Non-Communicable Diseases and Injuries NGO Non-Governmental Organizations NMR Neonatal Mortality Rate NPHRH National Plan for Human Resources for Health OOP Out-of-Pocket PNDS National Health Development Plan (Plano Nacional de Desenvolvimento Sanitario) PPP Purchasing Power Parity SDG Sustainable Development Goals SDI Service Delivery Indicators SNP Strategic Nutrition Plan (Plano Strategico de Nutricao) THE Total Health Expenditures UNICEF The United Nations Children's Fund USAID United States Agency for International Development U5MR Under-Five Mortality Rate WDI World Development Indicators WFP World Food Program WHO World Health Organization 3 Executive Summary 1. Guinea-Bissau meets many if not all the criteria that characterize health systems in fragile states. The country’s health system faces persistent challenges related to low public spending, poor infrastructure, inadequate supply of health workers, inadequate clinical and managerial training systems, malfunctioning referral system, non-operational health-information systems, weak governance and inadequate management capacity and systems (such as budgeting, public financial management and human resources management). Public spending accounts for about 20 percent of total health spending and is mostly used to pay staff salaries, while donors finance nearly 90 percent of the recurrent costs of the sector, including medicines and other critical health inputs. 2. This report aims to provide a comprehensive diagnostic of the health service delivery system in Guinea-Bissau. The diagnostic complements previous World Bank report which used secondary data to analyze the main health system’s building blocks in Guinea-Bissau. The previous report identified areas to be further explored, among them quality of care, workload/productivity and absenteeism. To do so the team implemented a service delivery indicators (SDI) survey which covers these and other key aspects of the health system for which no data were available in Guinea- Bissau (such as public expenditure tracking survey). The report provides an extensive analysis of the SDI data and links it to other analytical and operational work under implementation in the country. 3. The SDI questionnaire in Guinea-Bissau includes additional sections and modules to those traditionally collected by the SDI survey. The SDI survey traditionally collects information in 5 modules that include (i) facility characteristics and inputs which include infrastructure, equipment, medicine and vaccine availability, (ii) health worker characteristics and absenteeism, (iii) health worker knowledge and (iv) facility management practices and financing. The survey was designed in 2009 and piloted in Senegal and Tanzania, underwent a redesign in 2012 that was implemented in Kenya and has been generally only adjusted to each country context since, keeping intact all variables necessary for the calculation of the core SDI indicators. 4. User fees and sale of medicines are by far the most important source of income for health facilities. These amount to an average of 41 percent and 38 percent of total funds received by facilities, respectively, meaning that together these make up close to 80 percent of facility funds, on average. Not only are the average amounts received from user fees large compared to amounts received from other sources, but also the coverage among facilities, in the sense that almost all facilities received funds from either or both sources. In contrast, financing received by the Ministry of Health only contributed 0.6 percent of total funds (this does not include salary payments of health workers, which the state government also pays). 4 5. Health workers randomly selected for a knowledge assessment and observations were asked about the frequency with which they deliver consultations. Details on consultation frequency are available from 238 health workers. Among the health workers interviewed, 87% conduct pediatrics and adult consultations at least once per week while 70% undertake obstetrics and gynecology consultations and 66% family planning and antenatal care consultations at least once per week. There is some variation in pediatrics consultation exposure for health workers across different regions, in Quinara and Tombali, all health workers conduct pediatrics consultations at least once per week while the proportion is lower, 70%, for health workers in SAB. 6. The average caseload for facilities in Guinea-Bissau is 2.25 patients per health worker per day. There is a large variation in caseload across the different regions in the country with Oio and Bafata reporting the greatest number of patients per health worker per day (3.59 and 3.55 respectively) while Quinara and Bolama-Bijagos have the lowest caseload among regions (0.91 and 0.94 respectively). Across all facilities, 605 health workers were randomly selected for an assessment of absenteeism. The average rate of absenteeism across all health facilities in Guinea- Bissau is 34%; meaning that over a third of health workers are absent from a health facility when they ought to be present. 5 Contents Executive Summary..................................................................................................................................................4 List of Tables ............................................................................................................................................................7 List of Figures ..........................................................................................................................................................7 1. Introduction ................................................................................................................................................... 10 2. Methodology................................................................................................................................................... 13 1.1 – Survey Instrument design ..................................................................................................................... 13 1.2 – Implementation ..................................................................................................................................... 14 1.3 – Analysis methodology .......................................................................................................................... 15 3. Results............................................................................................................................................................. 15 3.1 – Respondent Characteristics................................................................................................................... 15 3.2 – Facility Characteristics ......................................................................................................................... 17 2.2.1 – Health Centers ..................................................................................................................................... 17 2.2.2 – Hospitals .............................................................................................................................................. 18 3.3 – Basic Infrastructure .............................................................................................................................. 19 3.3.1– Basic Equipment ................................................................................................................................... 20 3.3.2 – Essential Medicines ............................................................................................................................. 20 3.3.3 – Essential Vaccines ............................................................................................................................... 21 3.4 – Infection Control and Waste Management ........................................................................................... 22 3.5 – Availability of Basic Inputs as compared with other SDI countries..................................................... 23 3.6 – Financial Resources .............................................................................................................................. 24 3.7 – Health Workforce Characteristics ........................................................................................................ 25 3.7.1 – Consultation Frequency ....................................................................................................................... 28 3.7.2 – Recent Health Worker Training .......................................................................................................... 29 3.8 – Human Resources Management ........................................................................................................... 32 3.9 – Caseload ............................................................................................................................................... 34 3.10 – Absenteeism ......................................................................................................................................... 35 3.11 – Health Worker Knowledge ................................................................................................................... 36 3.12 – Patient experience ................................................................................................................................. 44 3.13 – Health Facility Management and Finance ............................................................................................ 50 6 List of Tables Table 1: SDI Health survey instrument description ................................................................................................ 13 Table 2: Description of Core SDI indicators ........................................................................................................... 15 Table 3: Survey characteristics ................................................................................................................................ 16 Table 4: Basic Facility Characteristics .................................................................................................................... 18 Table 5: Facility Inputs ............................................................................................................................................ 21 Table 6: Infection Control and Waste Management ................................................................................................ 22 Table 7 Reception of funds and goods .................................................................................................................... 24 Table 8: Spending of funds, by Facility type........................................................................................................... 25 Table 9: Health Worker Characteristics .................................................................................................................. 27 Table 10: Consultation Frequency: more than once per week ................................................................................ 29 Table 11: Health Worker Training in past 12 months ............................................................................................. 31 Table 12: Staff Perceptions of Management ........................................................................................................... 33 Table 13: Caseload .................................................................................................................................................. 34 Table 14: Absenteeism ............................................................................................................................................ 35 Table 15: Health Worker Knowledge ...................................................................................................................... 42 Table 16: Consultation Observations ...................................................................................................................... 45 Table 17: Consultation observation; basic illness screening ................................................................................... 46 Table 18: Use of director's time, by facility type .................................................................................................... 52 Table 19: Limiting factor, percent of facilities who mention as the limiting factor, by rural / urban ..................... 53 Table 20: Percentage of facilities in which a staff attendance register exists, average number of absent days during the past 30 days and percentage of absent days in which a replacement was found, by rural/urban ...................... 55 Table 21: Supervision visits by MoH in 2017, by Facility type .............................................................................. 56 Table 22: Reception of feedback and themes/issues mentioned in supervision book (percent of facilities), by region ....................................................................................................................................................................... 57 Table 23 Existence and quality of work plan, by region ......................................................................................... 58 Table 24 Financial Management, by facility type ................................................................................................... 59 Table 25: Facility Fees ............................................................................................................................................ 61 Table 26: Facility Fee exemptions........................................................................................................................... 63 Table 27 Stock ruptures, by strategy for receiving medicines................................................................................. 65 Table 28: Health Management committee by rural/urban ....................................................................................... 67 Table 29 Decisions in committee meetings, by rural/urban .................................................................................... 68 Table 30 Community contributions by rural/urban ................................................................................................. 69 Table 31 Complaints by rural/urban ........................................................................................................................ 70 Table 32 Response to feedback, by urban/rural ...................................................................................................... 71 List of Figures Figure 1: Distribution of Health Facilities in Guinea Bissau .................................................................................. 16 Figure 2: Bed Occupancy ............................................................................................ Error! Bookmark not defined. Figure 3: Availability of Basic Inputs as compared with other SDI countries ........................................................ 23 Figure 4: Type of health worker training received in past 12 months ..................................................................... 30 Figure 5: Facilities that undertake individual performance review (percent of facilities, by region) ..................... 32 Figure 6: Factors influencing performance review (percent of facilities, national) ................................................ 33 Figure 7: Adherence to Clinical Guidelines by Case and Training ......................................................................... 38 7 Figure 8: Diagnostic Accuracy by Case and Training ............................................................................................. 39 Figure 9: Treatment Accuracy by Case and Training.............................................................................................. 40 Figure 10: Management of Maternal and Neonatal Complications by Case and Training ..................................... 41 Figure 11: Health Worker knowledge by recent training topic ............................................................................... 44 Figure 12: Consultation score by health worker function ....................................................................................... 46 Figure 13: Patient Satisfaction by type of consultation and region ......................................................................... 48 Figure 14: Patient Satisfaction by type of consultation and facility type ................................................................ 48 Figure 15: Percentage of patients paying fees for child consultations by facility type ........................................... 49 Figure 16: Percentage of patients paying fees for adult consultations by facility type ........................................... 50 Figure 17: Characteristics of Health Facility Directors across Facility Types ........................................................ 51 Figure 18: Characteristics of Health Facility Directors Across Regions ................................................................. 51 Figure 19: Relationship between number of supervision visits and duration, excluding visits > 10 hours and centers with > 24 visits per year .............................................................................................................................. 56 Figure 20: Number of stock outs by facility type .................................................................................................... 64 Figure 21: Committee Member Appointment Mechanism ...................................................................................... 67 Figure 22: Percentage of health centers in which the complaint was listed among the three most frequent claims 71 Figure 23: Information sharing with community .................................................................................................... 72 8 Acknowledgments This report has been prepared by the World Bank Health, Nutrition and Population Global Practice (HNP GP). The report was led by Edson C. Araujo (TTL and Senior Economist, HNP GP). The core team included, from the World Bank: Manuela Uribe Villar (Consultant, World Bank), Julius Koll (Economist, Ministry of Public Health/ODI fellow), Charlotte Albin (Consultant, World Bank) and Alejandra Mia Garcia-Meza (Consultant, World Bank); and from the Government of Guinea-Bissau: Van Hanegem Menezes Moreira (Ministerio da Saude Publica, MINSAP) e Andreia Silva (Instituto Nacional de Estatistica, INE). The team received insightful comments and inputs from Christoph Herbst (Senior Health Specialist, GHN05) and Jaime Bayona (senior Health Specialist, GHNGE). The report was prepared under the supervision of Sophie Naudeau (Program Leader, AFCF1) and cleared by Gaston Sorgho (Practice Manager, GHN13) and Amadou Ba (Resident Representative, AFCF1). 9 1. Introduction 7. Guinea-Bissau meets many if not all the criteria that characterize health systems in fragile states. The country’s health system faces persistent challenges related to low public spending, poor infrastructure, inadequate supply of health workers, inadequate clinical and managerial training systems, malfunctioning referral system, non-operational health-information systems, weak governance and inadequate management capacity and systems (such as budgeting, public financial management and human resources management). Public spending accounts for about 20 percent of total health spending and is mostly used to pay staff salaries, while donors finance nearly 90 percent of the recurrent costs of the sector, including medicines and other critical health inputs. 8. The country faces persistent challenges in the health sector with a high burden of infectious diseases and high rates of child mortality. The country's life expectancy is 55 years, which is lower than the average for Guinea- Bissau’s regional (59) and income peers (60). Malaria is the single biggest cause of deaths, followed by HIV/AIDS, neonatal disorders, lower respiratory infections, diarrheal diseases and nutritional deficiencies. The burden of HIV in Guinea-Bissau is the highest in West Africa and it disproportionately affects more women than men (female adults with HIV represent 58.6 percent of the population above 15 years old with HIV).1 Progress has been made to reduce infant mortality, but both the infant mortality rate (IMR) and under-five mortality rate (U5MR) remain among the highest in the world, 60 and 88.8 per 1,000 live births, respectively. 9. Guinea-Bissau has one of the highest maternal mortality rates in the world. According to the last Multi Indicators Cluster Survey (MICS) the maternal mortality rate (MMR) is estimated at 900 maternal deaths per 100,000 live births, which is higher than the average among West Africa countries (579), among other low-income countries (542) and in Sub-Saharan Africa (494). The country did not achieve the Millennium Development Goal (MDG) for maternal health, set to lower MMR to 229 per 100,000 live births and is unlikely to achieve the Sustainable Development Goals (SDGs) target for 2030 along the current trend.2 10. Neonatal mortality rate (NMR), 35.8 per 1,000 live births, is higher than the average for West Africa and is strongly associated with birth spacing and birth order, indicating a lack of access to reproductive health services. The rate of NMR is comparable for any of the first six children born to a woman (approximately 36 per 1000 live births), but is 2.5 times higher for children born seventh or later in the birth order. This pattern is also true for birth spacing; children born less than two years after their previous sibling are almost twice as likely to die than if they were born at least three years after their previous sibling. These same patterns hold true for U5MR, currently at 89 per 1000 live births.5 Given constraints in the access pointed out above, birth spacing and maternal knowledge seem to be more important factors influencing child health outcomes. Unsurprisingly, only 16 percent of women ages 15-49 who are married or in a stable union report using any contraceptive method,3 and the adolescent pregnancy rate is estimated at 28 percent. 11. The utilization of obstetric services by expecting mothers in Guinea-Bissau is significantly low. Only 45 percent of the deliveries take place within health facilities.5 A recent assessment by a European Union (EU) funded health project showed only 38 percent of women met the standard four antenatal consultations and that out of 1 World Bank, 2016. Guinea-Bissau Health Sector Diagnostic. World Bank, Washington, DC. 2 World Development Indicators, 2016. 3 Guinea-Bissau Multiple Indicator Cluster Survey (MICS5), 2015. 10 every 100 women having at least one antenatal care visit, only 37 percent delivered their babies in a health facility.4 In addition, there is large variation in the burden of maternal and child health deaths distribution within Guinea-Bissau. U5MR, for example, varies from 41.8 per 1,000 in the region of Biombo to 158.9 per 1,000 in Gabú and 125.6 per 1,000 in Bafatá (Table 1). When looking across regions, Gabú and Bafatá consistently underperform in nearly all potential factors affecting child health outcomes. 12. This report aims to provide a comprehensive diagnostic of the health service delivery system in Guinea-Bissau. The diagnostic complements previous World Bank report which used secondary data to analyze the main health system’s building blocks in Guinea-Bissau. The previous report identified areas to be further explored, among them quality of care, workload/productivity and absenteeism. To do so the team implemented a service delivery indicators (SDI) survey which covers these and other key aspects of the health system for which no data were available in Guinea- Bissau (such as public expenditure tracking survey). The report provides an extensive analysis of the SDI data and links it to other analytical and operational work under implementation in the country. 4 PIMI Report. European Union, 2016. 11 Box 1: The Service Delivery Indicators (SDI) Program A significant share of public spending on education is transformed to produce good schooling outcomes at schools. Likewise, in health, public spending should contribute to good health outcomes. Understanding what takes place at these frontline service provision centers is the starting point in establishing where the relationship between public expenditure and outcomes is weak within the service delivery chain. Knowing whether spending is translating into inputs that teachers or health providers have to work with (e.g. basic equipment in health facilities, textbooks in schools), or how much work effort is exerted by health providers or teachers (e.g. how likely are they to come to work), and their competency would reveal the weak links in the service delivery chain. Reliable and complete information on these measures is lacking, in general. To date, there is no robust, standardized set of indicators to measure the quality of services as experienced by the citizen in Africa. Existing indicators tend to be fragmented and focus either on final outcomes or inputs, rather than on the underlying systems that help generate the outcomes or make use of the inputs. In fact, no set of indicators is available for measuring constraints associated with service delivery and the behavior of frontline providers, both of which have a direct impact on the quality of services that citizens are able to access. Without consistent and accurate information on the quality of services, it is difficult for citizens or politicians (the principal) to assess how service providers (the agent) are performing and to take corrective action. The SDI provides a set of metrics to benchmark the performance of health clinics and schools in Africa. The Indicators can be used to track progress within and across countries over time and aim to enhance active monitoring of service delivery to increase public accountability and good governance. Ultimately, the goal of this effort is to help policymakers, citizens, service providers, donors, and other stakeholders enhance the quality of services and improve development outcomes. The perspective adopted by the Indicators is that of citizens accessing a service. The indicators can thus be viewed as a service delivery report card on education and health care. H owever, instead of using citizens’ perceptions to assess performance, the Indicators assemble objective and quantitative information from a survey of frontline service delivery units, using modules from the Public Expenditure Tracking Survey (PETS), Quantitative Service Delivery Survey (QSDS), and Staff Absence Survey (SAS). The literature points to the importance of the functioning of health facilities and more generally, the quality of service delivery. The service delivery literature however is clear that, conditional on providers being appropriately skilled and exerting the necessary effort, increased resource flows for health can indeed have beneficial education outcomes. The SDI initiative is a partnership of the World Bank, the African Economic Research Consortium (AERC), and the African Development Bank to develop and institutionalize the collection of a set of indicators that would gauge the quality of service delivery within and across countries and over time. The ultimate goal is to sharply increase accountability for service delivery across Africa, by offering important advocacy tools for citizens, governments, and donors alike; to work toward the end goal of achieving rapid improvements in the responsiveness and effectiveness of service delivery. More information on the SDI survey instruments and data, and more generally on the SDI initiative can be found at: www.SDIndicators.org and www.worldbank.org/sdi, or by contacting sdi@worldbank.org. 12 2. Methodology 1.1 – Survey Instrument design The SDI questionnaire in Guinea-Bissau includes additional sections and modules to those traditionally collected by the SDI survey. The SDI survey traditionally collects information (see description Table) in 5 modules that include (i) facility characteristics and inputs which include infrastructure, equipment, medicine and vaccine availability, (ii) health worker characteristics and absenteeism, (iii) health worker knowledge and (iv) facility management practices and financing. The survey was designed in 2009 and piloted in Senegal and Tanzania, underwent a redesign in 2012 that was implemented in Kenya and has been generally only adjusted to each country context since, keeping intact all variables necessary for the calculation of the core SDI indicators. Guinea-Bissau is a relatively small country with a smaller number of health facilities as compared to other countries where SDI has been implemented, allowing for the possibility of undertaking a more in-depth analysis. As noted above, the Ministry of Health, World Bank and Development partners have previously identified human resources for health, in Guinea-Bissau, as a main constraint for service delivery. The opportunity to reach all health facilities allowed SDI in Guinea-Bissau to include additional questions on health worker characteristics, additional sections for the observation of patient consultations as well as an additional module to assess the patient experience, using patient exit interviews. The module used to assess facility management and finances in Guinea-Bissau is based on that developed for Mozambique in 2014 with some additional questions that assess specific transfer of inputs and finances from programs and donors of interest. Table 1: SDI Health survey instrument description Module of Module Main respondent Description Instrument Title Module 1 Facility Head of facility Information about the facility’s: functioning, infrastructure, information equipment, materials, supplies, and tracer drugs. Module 2A Health 2A: Head of facility 2A: Administered to head of facility to obtain a list of all and 2B Worker health workers. Roster 2B: Administered to randomly selected health workers to 2B: selected medical measure absence rates and to collect information about staff worker characteristics. Module 3 Clinical Selected Medical Administered to up to 10 randomly selected medical knowledge staff personnel who regularly treat patients to evaluate their assessment competency in the diagnosis and treatment of routine pathologies. Done using vignettes and consultation observations. Module 4 Facility Head of facility and Collection of information about revenues, expenditures, finances accountant (where management, governance, and drug provision for the and relevant) facility. governance Module 5 Patient Exit Adult patient Information about the patient experience in the facility and Interview consultation process, satisfaction with care, care Adult accompanying expenditures and socio-economic characteristics a child patient 13 To undertake the in-depth analysis of health worker knowledge and action, the sections for the observation of patient consultations and patient exit interview that were included in the questionnaire were designed using Kenneth Leonard’s [give reference] instruments for health worker assessments. These sections assess care for children under the age of 5 and adults separately. Their objective is to allow for triangulation of what health workers know, using vignettes, what health workers do, using observations and what patients perceive, using patient exit interviews. 1.2 – Implementation The SDI survey was implemented in Guinea-Bissau between March and April of 2018 over a period of six weeks. The survey collected information from 1522 health care providers across 132 health facilities, reaching all public health facilities, of all levels, in the country except for the Military Hospital for which access was not granted. Preparation for implementation began several months prior to data collection. An extensive process of consultation with the Ministry of Health and Development Partners, in country, was undertaken for the purpose of adjustment and validation of the survey instrument and data collection methodology to the context of Guinea-Bissau. In this process, the SDI questionnaire was carefully reviewed with different stakeholders and specifics related to facility locations, facility types, medicines, health worker cadres, clinical guidelines and facility management and financing practices were incorporated as revisions. Having reached consensus on the survey instrument content and data collection methodologies, ethical clearance for the implementation of the survey and analysis was granted by the National Institute of Health (Instituto Nacional de Saude- INASA) of Guinea-Bissau. The survey questionnaire was programmed using the Survey Solutions platform designed and managed by the World Bank. All survey modules, including the clinical vignettes, observations and patient interviews were programmed for tablet-based data collection. The training and data collection process was undertaken as a partnership between the National Institute of Statistics (Instituto Nacional de Estadistica-INE) and the National Institute of Health (Instituto Nacional de Saude- INASA) of Guinea-Bissau. The Ministry of Health provided the most updated list of public health facilities in the country, which served as the basis for the census of facilities visited by the survey and participated in the supervision of enumerator training and data collection. INE and INASA recruited a total of 16 enumerators, with a nursing background, who were trained over a period of two weeks on the details of data collection for each survey module. The training required much practice of each survey question as well as practice of the management of the tablet on which survey answers were recorded. Enumerators that received training were tested and only those that displayed in-depth knowledge of the survey instrument and data collection methodologies were recruited to participate in the field data collection process. After completion of the training of enumerators, a pilot test was undertaken in private facilities near the city of Bissau where the survey instrument, methodologies and enumerators were assessed. Small glitches in the survey data entry program, grammatical errors in the survey questions and specific areas of further training were identified and adjusted over a couple of days following the completion of the pilot test. The data collection teams were organized as a supervisor for two enumerator pairs, who were assigned to a specific region of the country. The enumerators were each assigned a role for the data collection process that was maintained throughout the survey; one enumerator collected information about the facility and its management, while the second enumerator collected information about health care workers and the patients they treated. The supervisor oversaw the completion and accuracy of the survey for each health facility as well as the online transfer of data, on a daily basis to supervisors in INE and the World Bank who checked data consistency and completeness on a daily basis as well. Any issues with the quality of the data were communicated immediately to the supervisors and corrected on the ground. As per SDI protocol, health facilities were visited twice to ensure the adequate assessment of health care worker absenteeism. The first visit was previously announced and confirmed with local authorities to allow for the participation of health facility directors or managers for the completion of data collection related to health facility 14 characteristics and management as well as health worker characteristics and knowledge and patient experiences. The second visit was unannounced and used to assess the presence of randomly selected health care workers, in each facility. 1.3 – Analysis methodology Data resulting from a tablet-based data collection system requires a shorter data cleaning process than paper-based surveys as the entry program was designed to control for outlier values, skip patterns and inconsistencies. The data management process for this survey, nonetheless, included a review of indicator values and a number of consistency checks. The data collected with this survey represent a census of health facilities and hence don not require the use of survey weights. The data cleaning process for this survey did not undertake any imputations. Data were analyzed around three survey respondent pillars: facilities and their directors, health workers and patients. It is for this reason that the denominator values change across measures. The SDI indicator values were calculated as described in Table 2. The survey collects a wealth of other information for which analyses are generally presented to distinguish across regions, locations, facility types, and health worker types. Table 2: Description of Core SDI indicators Indicator Definition Infrastructure availability Share of facilities with electricity, clean water and improved sanitation. Equipment availability Share of facilities with thermometer, stethoscope and weighing scale, refrigerator and sterilization equipment. Drug availability Share of basic drugs which at the time of the survey were available at the health facilities. Caseload per health provider Number of outpatient visits per clinician per day. Absence rate Share of a maximum of 10 randomly selected providers absent from the facility during an unannounced visit. Adherence to clinical Unweighted average of the share of relevant history taking questions, the share guidelines of relevant examinations performed. Management of maternal and Share of relevant treatment actions proposed by the clinician. neonatal complications Diagnostic accuracy Average share of correct diagnoses provided in the four clinical cases. 3. Results 3.1 – Respondent Characteristics The SDI survey in Guinea-Bissau visited a total of 132 of the 133 public health facilities reporting a total staff of 2338 of which 1522 are health workers. The number of health facilities per region varies between 8 facilities in Biombo and 21 in the region of Cacheu. Over two thirds (62%) of health facilities in the country are located in rural areas and a majority are categorized as Type C health centers which are in turn distributed across the different regions. Guinea-Bissau has a total of 7 hospitals of which 5 are Regional Hospitals located in the Region’s main 15 city while the two reference hospitals are located in the city of Bissau and in Biombo. See Figure 1 for exact location of health facilities in Guinea Bissau in relation to existing road system. Figure 1: Distribution of Health Facilities in Guinea Bissau The distribution of health workers across regions is not commensurate with the number of facilities. The region of SAB (Bissau) has 43.6% of all the country’s health workers in only 11 facilities while Quinara has only 5.1 % of all health workers in nearly the same number of facilities, the region with the fewest health workers. The large majority of health workers (86.5%) work in urban areas. Slightly more health workers are assigned to work in 2 reference hospitals (31.8%) than are assigned to work in the 103 Type C health centers around the country (27.6%). Table 3: Survey characteristics Facilities Health Workers n % n % Region Bafata 15 11.36 142 9.3 Biombo 8 6.06 128 8.4 Bolama Bijagos 16 12.12 82 5.4 Cacheu 21 15.91 133 8.7 Gabu 20 15.15 110 7.2 16 Oio 15 11.36 99 6.5 Quinara 10 7.58 77 5.1 SAB 11 8.33 663 43.6 Tombali 16 12.12 88 5.8 Location Rural 82 62.12 206 13.5 Urban 50 37.88 1316 86.5 Facility Type Health Center Type C 103 78.03 420 27.6 Health Center Type B 14 10.61 261 17.1 Health Center Type A 4 3.03 78 5.1 Maternal and Child health center 4 3.03 49 3.2 Regional Hospital 5 3.79 230 15.1 Reference Hospital 2 1.52 484 31.8 Total 132 100 1522 100 3.2 – Facility Characteristics Guinea-Bissau has six types of health facilities that have different basic characteristics. Of the 132 health facilities visited by the SDI survey, two are Referral Hospitals, five Regional Hospitals, four Maternal and Child Health Centers and a majority of health facilities are categorized as Health Center Type A, B or C (121, 92%). 2.2.1 – Health Centers The Health Centers, which are envisioned to provide increasingly complex primary health care, with Health Center Type C providing more basic care than Health Center Type A and B, are distributed across the country and report different basic characteristics. Health Centers Type C are the most common type of health facility in the country; there are a total of 103 of these facilities. The different regions have between 6 and 17 of these facilities, Bissau and Quinara each have 6 while Cacheu and Gabu have 17. A total of 82% of Health Centers Type C are located in rural areas, these are the only facilities that provide care to rural populations. The majority of Health Centers Type C, report that they provide services 7 days per week (86%) and 71% report that they provide health services for 8 to 12 hours per day; 27% report that they provide services 24 hours per day 5. These facilities are staffed, on average by 4 health workers who conduct 160 consultations per month. Only a handful (4%) of Health Centers Type C report that they provide hospitalization services with one in every two of these facilities having available one bed to be used for hospitalization. Occasionally these facilities keep patients over night; on average there were 8.5 hospitalizations per month. Nearly all Type C facilities perform deliveries; on average they deliver 9 babies per month. Three of these facilities reported a maternal death in the three months prior to the survey. Health Centers Type A and B are far fewer in number than Type C facilities; there are a total of 14 Health Centers Type B and 4 Health Centers Type C across the country. There are between one and three Health Centers Type B in each region except in Bolama Bijagos while Health Centers Type A are located in Bolama Bijagos, Cacheu and Quinara. All of the Type A and B facilities are located in Urban areas, the large majority report being open for consultations 7 days per week and between 8 and 12 hours per day. Health Centers Type A and B report having similar number of health workers in their staff, on average 19.5 and 18.6 accordingly. Type B facilities report that they provide, on average, nearly twice as many outpatient consultations per month (336) than Type A 5 Anecdotal evidence, reported by enumerators, suggests that Type C facilities that provide services 24 hours per day are those in which the health worker(s) of the facility are on call for emergency care. 17 facilities (186). The majority of these facilities provide hospitalization services and Type A facilities have, on average, slightly more inpatient beds (24.3) than Type B facilities (18.9). Type A and B facilities hospitalize on average, approximately one patient per day (35 and 27 patients per month, accordingly) but there is a large variation between facilities. Both types of facilities perform deliveries; Type B facilities report delivering nearly twice as many babies as Type A facilities, 26 and 16 accordingly. Three Type B facilities reported having had a maternal death in the three months prior to the survey. There are four Maternal and Child Health Centers in Guinea-Bissau, located in urban areas of Bafata, Cacheu, Gabu and Bissau. Three of these centers report being open less than 7 days per week and between 8 to 12 hours per day. On average they have 12 health workers that provide an average total of nearly 625 patients per month. None of the Maternal and Child Health Centers provide hospitalization or delivery services. 2.2.2 – Hospitals There are eight hospitals in Guinea-Bissau, seven of which were included in the SDI survey. Two Referral Hospitals are located in Biombo and Bissau, while the five Regional Hospitals are located in Tombali, Oio, Bafata, Gabu and Cacheu. All hospitals are located in urban areas of the country and provide outpatient consultations 7 days per week. One hospital, the national referral hospital, provides outpatient consultations 24 hours per day. The Regional Hospitals have on average 46 health workers on their staff while the referral hospitals have an average of 242 health workers. Referral Hospitals provide nearly four times the number of outpatient consultations that the Regional Hospitals provide (1779 and 489 accordingly). Both types of hospitals provide hospitalization services; Referral Hospitals have an average of 262 inpatient beds, admitting 355 patients per month, while Regional Hospitals have an average of 78 beds and admit 168 patients per month. All hospitals perform deliveries with Regional Hospitals having reported an average of 135 deliveries per month and Referral Hospitals an average of 169 per month. Regional Hospitals reported an average of 1.1 maternal deaths per month while Referral Hospitals reported 10.3 maternal deaths per month. Table 4: Basic Facility Characteristics Maternal Health Health Health and Child Center type Center Center Health Regional Reference Total C type B type A Center Hospital Hospital n (%) Region Bafata n (%) 12 (12) 1 (7) 0 (0) 1 (25) 1 (20) 0 (0) 15 (11) Biombo n (%) 6 (6) 1 (7) 0 (0) 0 (0) 0 (0) 1 (50) 8 (6) Bolama Bijagos n (%) 14 (14) 0 (0) 2 (50) 0 (0) 0 (0) 0 (0) 16 (12) Cacheu n (%) 17 (17) 1 (7) 1 (25) 1 (25) 1 (20) 0 (0) 21 (16) Gabu n (%) 17 (17) 1 (7) 0 (0) 1 (25) 1 (20) 0 (0) 20 (15) Oio n (%) 12 (12) 2 (14) 0 (0) 0 (0) 1 (20) 0 (0) 15 (11) Quinara n (%) 6 (6) 3 (21) 1 (25) 0 (0) 0 (0) 0 (0) 10 (8) SAB n (%) 6 (6) 3 (21) 0 (0) 1 (25) 0 (0) 1 (50) 11 (8) Tombali n (%) 13 (13) 2 (14) 0 (0) 0 (0) 1 (20) 0 (0) 16 (12) Location Rural n (%) 82 (80) 0 (0) 0 (0) 0 (0) 0 (0) 0 (0) 82 (62) Urban n (%) 21 (20) 14 (100) 4 (100) 4 (100) 5 (100) 2 (100) 50 (38) Basic Characteristics # days open for consultations <7 days n (%) 14 (13.6) 1 (7.1) 0 (0) 3 (75) 0 (0) 0 (0) 18 (13.6) 7 days n (%) 89 (86.4) 13 (92.9) 4 (100) 1 (25) 5 (100) 2 (100) 114 (86.4) 18 # hours providing outpatient consultations per day 1 to 7 hours n (%) 2 (1.9) 0 (0) 0 (0) 1 (25) 1 (20) 1 (50) 5 (3.8) 8 to 12 hours n (%) 73 (70.9) 10 (71.4) 4 (100) 3 (75) 4 (80) 0 (0) 94 (71.2) 24 hours n (%) 28 (27.2) 4 (28.6) 0 (0) 0 (0) 0 (0) 1 (50) 33 (25) 18.6 19.5 # of health workers Mean (SD) 4.1 (5) (10.1) (9.7) 12.3 (5.9) 46 (14.8) 242 (237.6) 132 (11.5) # outpatient consultations per 159.7 335.8 186 488.7 1769.7 month Mean (SD) (159.8) (271) (74.7) 624.4 (483) (275.8) (777.3) 131 (230.6) Provides Hospitalization services % (SD) 4 (19.4) 93 (26.7) 100 (0) 0 (0) 100 (0) 100 (0) 132 (21.2) 18.9 78.2 # of inpatient beds Mean (SD) 0.5 (3.3) (8.6) 24.3 (9) 0 (0) (31.8) 262 (270.1) 132 (10.1) # of hospitalizations per 27.2 35.1 168.3 354.8 month Mean (SD) 8.5 (15.9) (27.8) (18.8) NA (141) (303.8) 28 (74.3) Performs deliveries % (SD) 98 (13.9) 100 (0) 100 (0) 0 (0) 100 (0) 100 (0) 132 (95.5) 15.8 135.1 169.2 # of deliveries per month Mean (SD) 9.2 (9) 26 (16.4) (13) NA (83.6) (171.8) 126 (18.8) # maternal deaths per month Mean (SD) 0 (0.1) 0.1 (0.1) 0 (0) NA 1.1 (1) 10.3 (1.9) 126 (22.5) Total n (%) 103 (78) 14 (11) 4 (3) 4 (3) 5 (4) 2 (2) 132 (100) Although there are 28 health facilities across Guinea-Bissau that provide hospitalization services, the overall mean reported inpatient bed occupancy rate is 13.5%. Only two facilities have occupancy rates above 40%: those are the Health Center Type B in Bandim and the Hospital Simao Mendes in Bissau. 3.3 – Basic Infrastructure 19 Methodological Note The infrastructure indicator captures the availability of three inputs: water, sanitation and electricity. The indicator is an unweighted average of these three components. Eligible sources are: Electricity sources-electric power grid, a fuel operated generator, a battery-operated generator or a solar powered system as their main source of electricity. Water sources-piped into the facility, piped onto facility grounds or comes from a public tap/standpipe, tube well/borehole, a protected dug well, a protected spring, bottled water or a tanker truck. Sanitation sources-functioning flush toilets or Ventilated and Improved (VIP) latrines, or covered pit latrine (with slab). Across all health facilities in Guinea-Bissau, 45% have available a sustainable source of electricity (grid or solar panel), a reliable water source and a modern toilet. The availability of basic infrastructure in facilities varies greatly across regions; in Bissau 73% of facilities have basic infrastructure while in Bafata, only 20% meet these basic criteria. Reference Hospitals and Health Centers Type C stand slightly above other types of facilities in their availability of basic infrastructure with 50% and 49% availability accordingly. Only 25% of Health Centers Type A and Maternal and Child Health Centers have the three components of basic infrastructure available. Surprisingly, facilities in rural areas have higher availabilities of basic infrastructure than those in urban areas. 3.3.1– Basic Equipment Methodological Note The equipment indicator focuses on the availability (observed and functioning by the enumerator) of minimum equipment expected at a facility. The pieces of equipment expected in all facilities are: a weighing scale (adult, child or infant), a stethoscope, a sphygmomanometer and a thermometer and a refrigerator, and additionally sterilization equipment at health center and hospital levels. Health facilities in Guinea-Bissau have a high availability of basic equipment defined as a functional weighing scale, thermometer and stethoscope (92%), but when this measure is adjusted for the availability of a working refrigerator and sterilization equipment, the measure drops dramatically to 55%. Although high throughout, the availability of basic equipment varies from one region to the next; all facilities in Tombali and Oio have the three pieces of equipment available while 81% of facilities in Bolama-Bijagos, do. When we consider the availability of a refrigerator and sterilization equipment, the region with better equipped facilities is Bissau (82%) followed by Biombo (75%) while in Tombali only 25% of facilities have these two pieces of equipment additionally available. 3.3.2 – Essential Medicines Methodological Note This indicator is defined as the number of drugs of which a facility has one or more available, as a proportion of all the drugs on the list. The drugs have to be unexpired and have to be observed by the enumerator. The drug list contains tracer medicines for children and mothers identified by the World Health Organization (WHO) following a global consultation on facility-based surveys. 20 On average, health facilities in Guinea-Bissau had 69% of a set of essential medicines available and not expired, when visited by the survey team. There is over 20 percentage point variation across regions; facilities in Bolama- Bijagos were found to have only 59% of the essential medicines available while facilities in Biombo and Oio had 81% of these tracer medicines. The differences also span across health facility types with the four Health Centers Type A displaying the highest availability of medicines (91%) and the Maternal and Child Health Center, the lowest (39%). Facilities in urban areas have a higher percentage of available essential medicines (72%) than facilities in rural areas (67%). 3.3.3 – Essential Vaccines Although, on average health facilities in Guinea-Bissau had 69% of a defined set of essential vaccines available and not expired, the variation across regions and facility types is greater than that of essential medicines. The region of Bolama-Bijagos has the lowest availability of vaccines among the different regions (41% on average) while Biombo has the highest availability with 90%, more than twice as high. The availability of vaccines in Regional Hospitals is very low (17%) followed by Reference Hospitals (42%), the four Health Centers Type A have an average availability of vaccines of 92% while the most common type of facility (Health Center Type C) has 69% of vaccines available on average. Table 5: Facility Inputs Minimum Minimum Basic Infrastructure equipment Medicines Vaccines Equipment adjusted* n mean SD mean SD mean SD mean SD mean SD Region Tombali 16 56% 51% 100% 0% 25% 45% 61% 14% 52% 42% Bafata 15 20% 41% 93% 26% 53% 52% 60% 21% 74% 32% Biombo 8 38% 52% 88% 35% 75% 46% 81% 13% 90% 9% Bolama Bijagos 16 44% 51% 81% 40% 38% 50% 59% 20% 41% 34% Cacheu 21 24% 44% 95% 22% 57% 51% 79% 16% 73% 34% Gabu 20 65% 49% 89% 32% 60% 50% 63% 20% 81% 30% Oio 15 47% 52% 100% 0% 67% 49% 81% 15% 81% 34% Quinara 10 50% 53% 90% 32% 50% 53% 73% 15% 50% 44% SAB 11 73% 47% 91% 30% 82% 40% 72% 11% 82% 28% Health Facility Type Health Center Type 103 49% 50% 93% 25% 52% 50% 69% 19% 69% 35% C Health Center Type 14 36% 50% 79% 43% 64% 50% 77% 8% 83% 26% B Health Center Type A 4 25% 50% 100% 0% 100% 0% 91% 3% 92% 10% Maternal and Child 4 25% 50% 100% 0% 75% 50% 38% 21% 67% 45% Health Center Regional Hospital 5 40% 55% 100% 0% 20% 45% 65% 17% 17% 37% Reference Hospital 2 50% 71% 100% 0% 50% 71% 65% 16% 42% 59% Location Rural 82 50% 50% 94% 24% 49% 50% 67% 19% 66% 38% Urban 50 38% 49% 90% 30% 64% 48% 72% 17% 74% 34% Total 132 45% 50% 92% 27% 55% 50% 69% 19% 69% 36% *Adjusted to include the availability of a working refrigerator and sterilization equipment. 21 3.4 – Infection Control and Waste Management Methodological Note The infection control and waste management indicators are not core SDI indicators. However, these two indicators are key to understanding basic patient safety practices in a health facility. The indicators focus on the average availability of a minimum set of infection control items (running water, disposable gloves, disposable syringes and sharps boxes) or waste management practices (burning or incineration of sharps and biomedical waste that is disposed of in a non-visible location with the availability of waste management guidelines). On average, across all facilities in Guinea-Bissau, 87% of basic infection control items can be found, however, on average only 17% of the three basic waste management practices are available. The availability of infection control items varies across the different regions from a low of 78% in Bafata to 95% in SAB. The two reference hospitals in the country have available all the basic infection control items but only 87% of these items are available in Health Centers Type C, the most common type of facility in the country. There is little difference in the availability of infection control items when comparing facilities in urban and rural areas, rural areas have 87% of the basic items while urban areas have 88%. The availability of waste management practices in health facilities across the country is very low, ranging from 6% in Bolama Bijagos and SAB to 51% in Bafata. The highest percentage of these three practices is available in Type C facilities although this is only 19% on average. Rural areas have slightly less available waste management practices than urban areas, 16% and 19% accordingly. For waste management, the low performance in this index is largely driven by the complete lack of biohazard and sharps disposal practices where only a handful of Type C facilities practice adequate biomedical waste disposal (10% of Type C facilities) and sharps disposal (13% of Type C facilities), while no other facilities were observed to undertake these practices at a basic level of adequacy. Table 6: Infection Control and Waste Management Waste Infection Control Management N mean SD mean SD Region Tombali 16 84% 24% 17% 21% Quinara 10 93% 12% 17% 32% Oio 15 87% 13% 9% 15% Biombo 8 94% 12% 25% 39% Bolama_Bijagos 16 84% 20% 6% 18% Bafata 15 78% 16% 51% 33% Gabu 20 93% 12% 15% 17% Cacheu 21 85% 15% 13% 17% SAB 11 95% 10% 6% 13% Facility Type Health Center type C 103 87% 16% 19% 27% Health Center type B 14 93% 12% 12% 17% Health Center type A 4 75% 20% 17% 19% 22 Maternal and Child health center 4 81% 13% 8% 17% Regional Hospital 5 90% 22% 7% 15% Reference Hospital 2 100% 0% 17% 24% Location Rural 82 87% 16% 16% 25% Urban 50 88% 17% 19% 26% Total 132 87% 16% 17% 26% 3.5 – Availability of Basic Inputs as compared with other SDI countries Although low, the availability of medicines and infrastructure in Guinea Bissau, was found to be higher than the average for other countries that have implemented SDI surveys. As can be seen in Figure 2, the average percentage of essential medicines that are available and not expired across facilities in Guinea Bissau is greater than that of Tanzania (60%), Sierra Leone (56%) and Kenya (50%) among others. The availability of equipment in Guinea Bissau (56%) was found to be lower than countries like Tanzania (84%), Kenya (76%), Madagascar (62%), Togo (93%) and Mozambique (80%). Similarly, Guinea Bissau is approximately in the average of infrastructure availability when compared to other countries that have implemented SDI. Figure 2: Availability of Basic Inputs as compared with other SDI countries 100 93 90 84 78 80 80 76 69 70 64 60 62 60 56 56 54 53 50 50 49 49 48 47 48 47 50 45 43 39 39 40 36 34 32 28 30 22 24 22 20 13 10 0 Drug availability Equipment availability Infrastructure Availability 23 3.6 – Financial Resources User fees and sale of medicines are by far the most important source of income for health facilities. These amount to an average of 41 percent and 38 percent of total funds received by facilities, respectively, meaning that together these make up close to 80 percent of facility funds, on average. Not only are the average amounts received from user fees large compared to amounts received from other sources, but also the coverage among facilities, in the sense that almost all facilities received funds from either or both sources. In contrast, financing received by the Ministry of Health only contributed 0.6 percent of total funds (this does not include salary payments of health workers, which the state government also pays). This is in line with a recent PFM report, which notes that the government does not bestow to facilities funds for operating expenses or investment. Table 7 shows the percentage of facilities that received funds from each of the specified sources, the average value that facilities received if they received any, the share of total funds contributed by the source, and the share of the received funds that is dedicated to general expenses. Only 5 facilities received financial assistance from the MoH; three out of the five reference hospitals, two health centers type C and one health center type B. Only one of the two reference hospitals (HNSM) reported having received funds from the ministry of health and these amounted to 5 percent of their total funding. Funding received from MoH increases with the facility level, since higher-ranked facilities will serve more patients. In consequence, a higher percentage of urban centers (10 percent) received funding from MoH, compared to only 2 percent of rural centers. 18 percent of facility funds come from the PIMI program, financed by the European Union and distributed through either of the two NGOs EMI and IMVF. These are funds received for providing maternal-child services and for satisfying certain management and hygiene criteria. The national reference hospital received about five times as many funds as the next largest institution (Gabu Regional Hospital) and about 10 times as many funds as the third largest institution. It received 45 percent of its funding from the sale of medicine, 38 percent from user fees, 9 percent from the Ministry of Health and 8 percent from PIMI/EMI. Being a reference hospital increases total funds on average by 52.000.000 XOF, around 90.000 USD per year, even controlling for number of inpatients and outpatients, whether the facilities offer deliveries, other types of facilities and urban/rural areas. The second largest significant predictor is whether a facility admits inpatients, increasing annual funding by around 13.000.000 XOF (22.000 USD). Each bed-day gives a facility an additional funding of 28.000 XOF (48 USD). Whether facilities offer deliveries has no significant effect on funding, despite facilities being compensated by PIMI per delivery. Table 7 Reception of funds and goods Received funds Average value received in Share of total Share of funds to general (percent of facilities) 2017 (‘000 XOF) funds expenses (percent) MoH 5 10928 0.6 91 PIMI/EMI 44 2765 15 25 PIMI/IMVF 13 1492 3 25 International donors / 7 317 1 15 NGOs Private donors / 5 370 0.8 22 individuals Community 1 500 0.2 0 User fees 89 6099 41 63 Sale of medicine 87 5205 38 57 24 Almost all facilities (92 percent) spent funds on medicines and on cleaning materials (Table 8: Spending of funds, by Facility typeTable 8). This is relatively constant across regions, urban and rural facilities and type of health centers. Among facilities that do not spend funds on medicines, the main provider of medicines is the government (7 facilities), a NGO (2 facilities) and the patients (1 facility). The second most important item in terms of spending is the salary of non-health staff (janitors, maids, guards etc.); 87 percent of facilities spend funds on this. While 82 percent of urban centers spend money on construction and maintenance, only 57 percent of rural centers spend money on this. Urban facilities tend to be larger and have more inpatients, and in consequence they have about 14 times as much funds as rural facilities. This might allow them to more frequently spend money on items such as construction and maintenance which are not day-to-day expenses. More than double the share of urban centers (52 percent) spend money on public utilities (electricity, water, gas, etc.) compared to rural centers (24 percent). These utilities are often not available from a charging provider in rural regions. This is in line with the fact that in the least urbanized regions (Tombali, Quinara, Bolama/Bijagos and Gabu) no or almost no facility spends funds on public utilities. In addition, the higher funding of urban facilities might explain part of this pattern. Approximately 20 percent of health facilities reported spending their funds on health worker salaries. This is surprising, since salaries are generally paid directly by the government. However, during the fiscal year 2017, the period which the survey question referred to, there were several periods in which government salaries were delayed or not paid. It might be that facilities used their internal funds during this time to substitute the government’s salary. Alternatively, it might be that funds are used to top up salaries or to pay non-salaried interns. No Health Center type A and no Maternal-Child clinic spent funds on health worker salaries. Two out of five regional hospitals did and the Reference Hospital of Cumura also did so. The latter is not owned by the government and religious organizations pay the generality of its salaries. The regions in which the highest share of facilities spent funds on health-worker salaries are Cacheu (48 percent), Biombo (38 percent) and SAB (36 percent). In contrast, 0 percent, 6 percent and 7 percent of facilities spent funds on health-worker salaries in Gabu, Tombali and Bafata, respectively. Table 8: Spending of funds, by Facility type Health Health Health Maternal and Regional Reference Total Center Type Center Center Child Health Hospital Hospital A Type B Type Center C N 4 14 103 4 5 2 132 Percent of facilities that spent funds on: Medicines 100 93 93 100 60 100 92 Cleaning material 100 93 92 100 80 100 92 Other staff salary 100 86 87 75 80 100 87 Construction / 75 77 63 75 80 100 66 Maintenance Medical material 50 69 62 50 80 100 63 Medical equipment 50 71 35 75 40 100 42 Public utilities 25 36 34 50 40 50 35 Health worker salary 0 31 17 0 40 50 19 Other 75 17 29 25 20 50 29 3.7 – Health Workforce Characteristics 25 Detailed information was collected from 940 health workers across all health facilities. The average age of health workers in Guinea Bissau is nearly 41 years. There are some, but not significant, differences in health worker age across regions. The average age of health workers in SAB is 45.5 years while health workers in Quinara and Bolama-Bijagos, are on average, 10 years younger (36 and 37.2 years respectively). The average age of health workers across each of the different types of facilities is not significantly different from the national average, suggesting very similar distributions by age within health facilities. The differences in health worker age are slightly larger (but not significant) when we compare those posted to urban and rural locations; health workers in urban locations are on average 4 years older than those in rural areas. Midwives are on average, nearly 7 years older than the average health worker in Guinea Bissau while nurses are slightly younger than average. Less than half of health workers in Guinea-Bissau are men (41%) but there is variation in their distribution across regions, location and function. Tombali is the region with the greatest proportion of men among its health workers (65%) while SAB (Bissau) is the region with the smallest proportion of men (19%) followed by Biombo where over two-thirds of the health workforce is female. There are more women health workers than men across all types of health facilities except for Health Centers Type A. Nearly two-thirds of health workers in Health Centers Type C are women, 37% are men. The gender distribution across geographic locations suggests that male health workers are more likely to be posted in rural locations than female health workers; 48% of health workers in rural health facilities are men while 33% of health workers in urban facilities are men. The gender distribution across cadres also shows some differences; the majority of facility directors (90%) and doctors (66%) are men while most midwives (98%) and nurses (62%) are women. Place of birth is often seen as an indication of the likelihood that a health worker will serve a facility or community for an extended period; if the health worker was born in the locality of the facility where they are posted, they are more likely to serve there longer. On average, one fifth (21%) of health workers in Guinea Bissau report having been born in the same locality as the facility they work in. There are some significant differences in these proportions between regions; only 1% of health workers in Gabu, 7% in Oio and 10% in Bafata were born in the same locality where they work in contrast with SAB where 38% report being born in the locality of their facility. When comparing across types of health facilities we find that only 10% of health workers posted to Regional Hospitals were born in the locality of this facility, in contrast with those posted to Reference Hospitals (37%), 19% of health workers posted to Health Centers Type C were born in the same locality as their facility. Health workers posted to facilities in rural areas are less likely (16%) than those posted to urban areas (25%) to have been born in the locality of the facility. Doctors and midwives are more likely than the average health worker to have been born in the locality of their health facility, given that the majority of doctors are posted to urban areas and midwives to rural areas, this is likely further indication of the importance of recruitment of health workers from the localities where they were born, to ensure equity in their distribution. Health workers who are members of a union are seen to be more likely to negotiate their salaries and working conditions. On average, 44% of health workers in Guinea-Bissau report being members of a worker’s union. There is a large variation in the percentage of health workers that are part of a union, when comparing across regions. In Bolama-Bijagos 82% of health workers report being unionized while 16% in Quinara and 17% in Tombali report being unionized. The differences in union membership is also apparent when looking at different types of facilities where health workers are posted; all health workers of Maternal and Child Health Centers for which we have data available are unionized, 84% of health workers posted to Health Centers Type B while 39% of those posed to Health Centers Type C belong to a worker’s union. Among health workers posted to rural locations 50% report union membership compared to 38% of those posted to urban areas. Doctors and Directors are more likely than average to be members of a worker’s union, 60% and 65% respectively. Health worker experience is highly correlated with age. On average, health workers in Guinea Bissau have 11 years of experience in the health sector. The most experienced health workers are posted to facilities in SAB with an average experience of 17.2 years, this is in contrast to health workers in Bolama-Bijagos and Quinara who have, on average 6.8 and 7.1 years of experience, respectively. Health workers posted to Health Centers Type A have, on average less experience than those working in other types of facilities (9.7 years), while those working in the 26 Reference Hospital have slightly more experience (12.3). The difference in the years of experience of health workers posted to urban and rural areas is large; health workers in rural areas have an average of 8.7 years of experience while those in urban areas have 15 years of experience. Midwives and technicians are the health workers with the greatest average number of years of experience, 16 and 17.7 years respectively, while nurses are less experienced (9.7 years). Delays in payment were not reported as a major issue by health workers across Guinea-Bissau. This is potentially due to the payment of health workers coming directly from the user fees charged in the facilities where they are posted and not from a central source of funds. Only 7% of health workers reported a delay in their last payment. Although only very few health workers across most regions and health facilities experienced a delay in payment, those in a Health Center Type A of the region of Quinara reported a much higher rate of this delay. Table 9: Health Worker Characteristics Gender Born in Age (yrs) Unionized Experience (yrs) Payment delay (Male) locality Mea Mea Mea Mea Mea Mea N n SD n SD n SD n SD N n SD N n SD Region Bafata 100 37.3 9.1 44% 50% 10% 30% 49% 50% 81 8.9 10.4 36 6% 23% Biombo 79 41.4 9.2 27% 44% 25% 44% 70% 46% 58 12.3 11.1 39 5% 22% Bolama_Bijagos 74 37.2 9.3 53% 50% 24% 43% 82% 38% 59 6.8 7.8 25 8% 28% Cacheu 120 40.2 9.5 44% 50% 16% 37% 65% 48% 94 11.4 11.2 54 4% 19% Gabu 75 42.0 10.1 56% 50% 1% 11% 21% 41% 62 11.4 12.1 42 2% 15% Oio 95 40.8 10.9 44% 50% 7% 26% 57% 50% 68 11.6 14.1 39 5% 22% Quinara 77 36.0 11.4 61% 49% 21% 41% 16% 37% 55 7.1 10.9 25 28% 46% SAB 246 45.5 10.9 19% 39% 38% 49% 31% 46% 126 17.2 13.5 62 8% 27% Tombali 72 39.3 9.8 65% 48% 17% 38% 17% 38% 53 8.8 10.5 23 0% 0% Facility Type Health Center type C 469 40.6 10.6 37% 48% 19% 40% 39% 49% 385 11.1 12.3 220 6% 24% Health Center type B 219 41.0 10.9 42% 50% 24% 43% 34% 48% 134 11.5 12.5 63 8% 27% Health Center type A 70 39.5 10.7 51% 50% 20% 40% 84% 37% 46 9.7 9.5 11 36% 50% Maternal and Child 100 health center 8 45.3 12.1 25% 46% 13% 35% % 0% 8 20.0 12.4 6 0% 0% Regional Hospital 97 41.2 10.4 48% 50% 10% 30% 44% 50% 48 11.9 10.9 21 0% 0% Reference Hospital 75 43.2 10.3 41% 50% 37% 49% 57% 50% 35 12.3 12.3 24 0% 0% Location Rural 474 38.1 9.6 48% 50% 16% 37% 50% 50% 381 8.7 10.2 216 8% 27% Urban 464 43.8 10.9 33% 47% 25% 44% 38% 49% 275 15.0 13.5 129 5% 21% Function Director 20 43.4 10.8 90% 31% 15% 37% 65% 49% 13 12.9 9.9 9 0% 0% Doctor 87 40.2 10.1 66% 48% 28% 45% 60% 49% 57 11.4 11.9 28 7% 26% Nurse 586 39.3 10.3 38% 49% 19% 40% 39% 49% 429 9.7 11.2 227 7% 26% Midwife 65 47.2 8.9 2% 12% 26% 44% 50% 50% 41 16.0 10.4 13 15% 38% Technician 138 46.0 11.3 43% 50% 17% 38% 49% 50% 88 17.7 14.9 57 2% 13% Other 39 35.8 6.3 51% 51% 36% 49% 38% 49% 26 8.0 10.2 11 9% 30% Total 940 40.9 10.7 41% 49% 21% 41% 44% 49% 656 11.3 12.1 345 7% 25% 27 3.7.1 – Consultation Frequency Health workers randomly selected for a knowledge assessment and observations were asked about the frequency with which they deliver consultations. Details on consultation frequency are available from 238 health workers. The frequency with which health workers deliver specific types of consultations gives us an indication of the rate of de facto specialization of health workers in a given field. Among the health workers interviewed, 87% conduct pediatrics and adult consultations at least once per week while 70% undertake obstetrics and gynecology consultations and 66% family planning and antenatal care consultations at least once per week. There is some variation in pediatrics consultation exposure for health workers across different regions, in Quinara and Tombali, all health workers conduct pediatrics consultations at least once per week while the proportion is lower, 70%, for health workers in SAB. There are some differences across facilities in health worker exposure to pediatrics consultations, while all health workers in Health Centers Type B report conducting pediatrics consultations more than once per week, 88% in Health Centers type C and 38% in Reference Hospitals do so as well. The two health workers in this sample that are posted to Maternal and Child Health Centers do not report undertaking pediatrics consultations. Health workers in urban areas appear to be more specialized, as on average a fewer (77%) percentage of health workers posed to urban areas undertake pediatrics consultations on a weekly basis as compared to those posted to rural areas (92%). When comparing across health worker functions, 92% of nurses reported conducting pediatrics consultations more than once per week, while 76% of doctors and 45% of midwives did as well. Interestingly, most facility Directors in this sample conduct pediatrics consultations at least once per week. There is some variation in adult consultation exposure for health workers across different regions, as with pediatrics, in Quinara and Tombali, all health workers conduct adult consultations at least once per week while the proportion is lower, 60%, for health workers in SAB. There are some differences across facilities in health worker exposure to adult consultations, while 97% of health workers in Health Centers Type B report conducting adult consultations more than once per week, 87% in Health Centers type C and 50% in Reference Hospitals do so as well. Health workers from Maternal and Child Health centers, as with pediatrics, do not undertake general adult consultations in their practice. Health workers in urban areas appear to be more specialized, as on average a fewer (74%) percentage of health workers posed to urban areas undertake adult consultations on a weekly basis as compared to those posted to rural areas (93%). When comparing across health worker functions, 91% of nurses reported conducting adult consultations more than once per week, while 74% of doctors and 45% of midwives did as well. All facility Directors in this sample conduct adult consultations at least once per week. A smaller percentage of health workers conduct OBGYN consultations on a regular basis (70%) than other types of consultations. Across different regions there is some variation in OBGYN consultation exposure for health workers, in Tombali, 90% of health workers report conducting these types of consultations on a weekly basis, while the proportion is significantly lower, 13%, for health workers in SAB. There are some differences across facilities in health worker exposure to OBGYN consultations, while 100% of health workers in Maternal and Child Health Centers and 79% in Health Centers Type C report conducting OBGYN consultations on a weekly basis, 29% in Regional Hospitals and 50% in Reference Hospitals do so as well. Health workers in urban areas appear to be more specialized, as on average a fewer (47%) percentage of health workers posed to urban areas undertake OBGYN consultations on a weekly basis as compared to those posted to rural areas (81%). When comparing across health worker functions, 91% of midwives reported conducting OBGYN consultations more than once per week, while 37% of doctors and 76% of nurses did as well. A smaller percentage of health workers conduct family planning and antenatal care (FP and ANC) consultations on a regular basis (66%) than other types of consultations. Across different regions there is some variation in FP and ANC consultation exposure for health workers, in Bolama-Bijagos, 92% of health workers report conducting these types of consultations on a weekly basis, while the proportion is significantly lower, 13%, for health workers in SAB. There are some differences across facilities in health worker exposure to FP and ANC consultations, while 28 100% of health workers in Maternal and Child Health Centers and 76% in Health Centers Type C report conducting FP and ANC consultations on a weekly basis, 14% in Regional Hospitals and 13% in Reference Hospitals do so as well. AS with the other types of consultations, health workers in urban areas appear to be more specialized, as on average a fewer (44%) percentage of health workers posed to urban areas undertake FP and ANC consultations on a weekly basis as compared to those posted to rural areas (78%). When comparing across health worker functions, 91% of midwives reported conducting FP and ANC consultations more than once per week, while 18% of doctors and 75% of nurses did as well. Table 10: Consultation Frequency: more than once per week Family Planning Pediatrics Adults OBGYN and ANC N Mean SD Mean SD Mean SD Mean SD Region Bafata 22 82% 39% 86% 35% 77% 43% 64% 49% Biombo 20 85% 37% 95% 22% 65% 49% 50% 51% Bolama_Bijagos 36 94% 23% 94% 23% 89% 32% 92% 28% Cacheu 37 81% 40% 81% 40% 73% 45% 68% 47% Gabu 33 94% 24% 91% 29% 79% 42% 76% 44% Oio 22 82% 39% 82% 39% 68% 48% 73% 46% Quinara 18 100% 0% 100% 0% 78% 43% 72% 46% SAB 30 70% 47% 60% 50% 13% 35% 13% 35% Tombali 20 100% 0% 100% 0% 90% 31% 90% 31% Facility Type Health Center type C 170 88% 33% 87% 34% 79% 41% 76% 43% Health Center type B 31 100% 0% 97% 18% 42% 50% 45% 51% Health Center type A 13 92% 28% 92% 28% 69% 48% 69% 48% Maternal and Child health center 2 0% 0% 0% 0% 100% 0% 100% 0% Regional Hospital 14 86% 36% 86% 36% 29% 47% 14% 36% Reference Hospital 8 38% 52% 50% 53% 50% 53% 13% 35% Location Rural 160 92% 27% 93% 26% 81% 40% 78% 42% Urban 78 77% 42% 74% 44% 47% 50% 44% 50% Function Director 6 83% 41% 100% 0% 33% 52% 33% 52% Doctor 38 76% 43% 74% 45% 37% 49% 18% 39% Nurse 160 92% 27% 91% 28% 76% 43% 75% 43% Midwife 11 45% 52% 45% 52% 91% 30% 91% 30% Technician 18 89% 32% 89% 32% 83% 38% 83% 38% Other 5 100% 0% 100% 0% 60% 55% 80% 45% Total 238 87% 34% 87% 34% 70% 46% 66% 47% 3.7.2 – Recent Health Worker Training 29 Among the 231 health workers for which we have data available, from across all health facilities in Guinea- Bissau, 81% reported having received training on one or more service provision topics in the 12 months prior to the survey. The most common topic of training received was Neonatal Resuscitation (35%) followed by HIV counseling and testing (34%) while the least common was Adolescent Health (7%). Figure 3: Type of health worker training received in past 12 months 100% 81% 80% 60% 47% 40% 35% 34% 27% 23% 20% 15% 7% 0% Although health workers across all regions report having received training in the past year, there are regional variations in their responses: among health workers in Oio, Tombali and Bolama-Bijagos, 95%, 95% and 94% respectively, reported receipt of training while in SAB only 59% did so as well. All health workers posted to Health Centers Type A and Maternal and Child Health centers reported having received some form of training while 64% of those posed in Regional Hospitals did as well. A higher proportion of health workers posted to rural areas (85%) received training in the past year than those posted to urban areas (72%). Among the different health worker functions, the largest proportion having received any training in the previous year, are technicians (89%), while 83% of nurses and 68% of doctors received some form of training in the past year. 30 Table 11: Health Worker Training in past 12 months Adolescent Neonatal HIV counseling Any Training Safe Deliveries health Infection control IMCI Resuscitation and testing Other training N Mean SD Mean SD Mean SD Mean SD Mean SD Mean SD Mean SD Mean SD Region Bafata 21 81% 40% 43% 51% 0% 0% 25% 44% 10% 30% 33% 48% 25% 44% 35% 49% Biombo 20 80% 41% 15% 37% 10% 31% 10% 31% 32% 48% 25% 44% 25% 44% 53% 51% Bolama_Bijagos 35 94% 24% 43% 50% 26% 44% 38% 49% 3% 17% 49% 51% 75% 44% 34% 48% Cacheu 37 78% 42% 11% 31% 0% 0% 6% 24% 29% 46% 28% 45% 24% 43% 61% 49% Gabu 29 76% 44% 28% 45% 10% 31% 18% 39% 7% 26% 34% 48% 32% 48% 53% 51% Oio 22 95% 21% 27% 46% 0% 0% 0% 0% 14% 35% 23% 43% 27% 46% 70% 47% Quinara 18 72% 46% 28% 46% 6% 24% 17% 38% 11% 32% 39% 50% 11% 32% 24% 44% SAB 29 59% 50% 3% 19% 7% 26% 17% 38% 18% 39% 24% 44% 15% 37% 38% 50% Tombali 20 95% 22% 60% 50% 0% 0% 80% 41% 15% 37% 60% 50% 50% 51% 50% 51% Facility Type Health Center type C 163 83% 37% 29% 46% 9% 28% 25% 43% 17% 38% 40% 49% 36% 48% 46% 50% Health Center type B 31 65% 49% 13% 34% 0% 0% 13% 34% 3% 18% 23% 43% 31% 47% 52% 51% Health Center type A 13 100% 0% 23% 44% 15% 38% 23% 44% 8% 28% 23% 44% 62% 51% 42% 51% Maternal and Child health center 2 100% 0% 0% 0% 0% 0% 0% 0% 0% 0% 0% 0% 0% 0% 100% 0% Regional Hospital 14 64% 50% 29% 47% 0% 0% 29% 47% 21% 43% 29% 47% 8% 28% 50% 52% Reference Hospital 8 88% 35% 50% 53% 13% 35% 13% 35% 25% 46% 25% 46% 14% 38% 50% 55% Location Rural 157 85% 35% 32% 47% 8% 27% 24% 43% 14% 35% 38% 49% 39% 49% 49% 50% Urban 74 72% 45% 16% 37% 7% 25% 19% 39% 16% 37% 27% 45% 23% 43% 43% 50% Function Director 6 83% 41% 0% 0% 0% 0% 33% 52% 33% 52% 33% 52% 50% 55% 50% 55% Doctor 37 68% 47% 22% 42% 5% 23% 14% 35% 16% 37% 22% 42% 15% 36% 41% 50% Nurse 154 83% 38% 27% 45% 5% 23% 21% 41% 15% 36% 32% 47% 35% 48% 51% 50% Midwife 11 82% 40% 55% 52% 0% 0% 18% 40% 9% 30% 73% 47% 18% 40% 45% 52% Technician 18 89% 32% 28% 46% 28% 46% 39% 50% 6% 24% 61% 50% 50% 51% 24% 44% Other 5 80% 45% 40% 55% 40% 55% 60% 55% 20% 45% 40% 55% 80% 45% 50% 58% Total 231 81% 39% 27% 45% 7% 26% 23% 42% 15% 36% 35% 48% 34% 47% 47% 50% 31 3.8 – Human Resources Management Management practices in facilities were investigated from the perspectives of the facility managers and the health workers. Only one third of facility directors conduct individual performance reviews of their staff (Figure 4). This number is consistent across urban and rural areas but not across different regions. Only 55 percent of facilities in SAB conduct performance reviews, compared to 19 percent in Tombali. Figure 4: Facilities that undertake individual performance review (percent of facilities, by region) SAB Oio Bolama-Bijagos Biombo National Cacheu Bafata Gabu Quinara Tombali 0 10 20 30 40 50 60 The reported factors used to influence health worker performance reviews vary greatly across regions. Patient satisfaction is the most frequently mentioned factor influencing performance review (mentioned in 44 percent of facilities). However, patient satisfaction does not play any role during performance review in Tombali and in Biombo but is assessed as part of the performance review in 86 percent of facilities in Oio. Interestingly, absenteeism does not play a role in performance review in any health center. In urban areas, direct supervision of patient care is the most frequent factor influencing performance review (mentioned in 60 percent of centers). Directors of rural facilities are less prone to conduct direct supervision of patient care for performance review; this could be because directors conduct consultations themselves and have less time to supervise the consultation of other staff. In Quinara, all facilities that conduct performance reviews report doing so based on direct knowledge assessment, and 50 percent of facilities do so exclusively. This stands in contrast to most other regions, where direct knowledge assessment frequently does not influence performance review (mentioned in 29 percent of facilities nationally). Figure 5 shows the share of facilities in each region that conduct individual performance review while figure … shows the factors influencing performance review on the national level. 32 Figure 5: Factors influencing performance review (percent of facilities, national) Patient satisfaction Knowledge assessment Direct supervision of patient care Patient results Cases per day Other Absenteeism 0 10 20 30 40 50 *Note: Percentages do not add to 100 because multiple options could be named. Overall, health workers interviewed were generally positive about specific aspects of how their facility is being managed; the difference between those that agree with statements of good management practices and those that strongly agree, tell a more nuanced story. Health workers agree or strongly agree (95%) that data is used to improve the facility, staff have the autonomy to decide on their work programs (95%), conflicts within the facility are adequately resolved (92%) and effort is equally distributed across health workers and staff (91%). Although health workers are generally positive in their views, a large difference is seen between the percentage that agree and those that strongly agree with statements of good management practices in their facilities. Only about one third of health workers that that have positive views about use of data for improvement, staff autonomy, conflict resolution and distribution of efforts would go so far as to say that this is definitely the case in their facilities. A more nuanced assessment of the findings points to large perceived gaps related to health worker performance management and compensation practices. Only 26% of health workers strongly believe that their manager communicates a vision, 21% strongly believe that staff performance measures are objective, 20% strongly believe that better staff are promoted and a dire 3% strongly believe that staff are compensated for their efforts and 1% strongly believe that staff that underperform are fired. Table 12: Staff Perceptions of Management Proportion Agree or of strongly Strongly Strongly agree Agree agree Agree Use of data for improvement 95% 63% 32% 34% Staff autonomy 95% 65% 30% 31% Conflicts resolved adequately 92% 59% 33% 36% Effort equally distributed 91% 61% 30% 33% Senior staff available to mentor 89% 60% 29% 33% Staff encouraged to contribute 87% 55% 32% 37% Good communication between staff 86% 50% 37% 43% Staff receive training 86% 56% 30% 35% Manager communicates vision 86% 60% 26% 30% Staff participate in decisions 84% 52% 31% 38% 33 Regular staff meetings 78% 48% 30% 39% Staff performance measures are objective 78% 57% 21% 27% Staff like to work in facility 76% 46% 30% 39% Staff trust each other 70% 43% 26% 38% Better performing staff are promoted 58% 38% 20% 35% Better staff are compensated for efforts 23% 20% 3% 12% Low performing staff are fired 7% 5% 1% 22% Note: 238 health workers were interviewed on their views of facility management practices. 3.9 – Caseload Methodological Note The caseload indicator is defined as the number of outpatient visits (recorded in outpatient records) in the three months prior to the survey, divided by the number of days the facility was open during the 3-month period and the number of health workers who conduct patient consultations. In hospitals, the caseload indicator was measured using outpatient consultation records; only providers doing outpatient consultations were included in the denominator. The term caseload rather than workload is used to acknowledge the fact that the full workload of a health provider includes work that is not captured in the numerator, notably administrative work and other non-clinical activities. From the perspective of a patient or a parent coming to a health facility, caseload—while not the only measure of workload—is arguably a critically important measure. The average caseload for facilities in Guinea-Bissau is 2.25 patients per health worker per day. There is a large variation in caseload across the different regions in the country with Oio and Bafata reporting the greatest number of patients per health worker per day (3.59 and 3.55 respectively) while Quinara and Bolama-Bijagos have the lowest caseload among regions (0.91 and 0.94 respectively). There are some differences in caseload across facility types whereas the Maternal and Child Health centers have nearly three times the number of patients per health worker per day than the national average (6.89), Health Centers Type A and reference hospitals have lower caseloads than average (1.26 and 1.81 respectively). Health facilities in urban areas see more patients per health worker per day, 2.52, than do health facilities in rural areas,2.08. Table 13: Caseload Caseload N mean sd Bafata 15 3.55 3.1 Biombo 8 2.39 1.5 Bolama_Bijagos 16 0.94 0.7 Cacheu 21 2.40 1.3 Gabu 19 2.05 1.3 Oio 15 3.59 3.6 Quinara 10 0.91 0.3 SAB 11 2.12 1.8 Tombali 16 1.97 2.2 Facility Type 34 Health Center type C 102 2.14 2.0 Health Center type B 14 2.29 2.3 Health Center type A 4 1.26 1.1 Maternal and Child health center 4 6.89 3.6 Regional Hospital 5 1.54 1.2 Reference Hospital 2 1.81 2.1 Location Rural 81 2.08 2.1 Urban 50 2.52 2.2 Total 131 2.25 2.2 3.10 – Absenteeism Methodological Note The average rate of absence at a facility is measured by assessing the presence of at most ten randomly selected clinical health staff at a facility during an unannounced visit. Only workers who are supposed to be on duty are considered in the denominator. Thus, workers on call and off duty were excluded from the analysis. The approach of using unannounced visits is regarded best practice in the service delivery literature. Across all facilities, 605 health workers were randomly selected for an assessment of absenteeism. The average rate of absenteeism across all health facilities in Guinea-Bissau is 34%; meaning that over a third of health workers are absent from a health facility when they ought to be present. The rate of absenteeism varies across regions, where facilities in Tombali have an average absenteeism rate of 55%, followed by Quinara where on average 50% of health workers are absent from their facility at a given moment. Absenteeism is lowest, albeit high, in Gabu (19%) followed by Biombo (25%). Health workers in Health Facilities Type A have the highest rate of absenteeism with 52% being absent from their post during the unannounced visit, while health workers in Reference Hospitals have the lowest rate of absenteeism, 13%. Absenteeism is slightly higher in rural facilities than in urban facilities, 35% versus 34%. Table 14: Absenteeism Absence n mean SD Region Bafata 77 43% 50% Biombo 56 25% 44% Bolama_Bijagos 44 45% 50% Cacheu 90 29% 46% Gabu 57 19% 40% Oio 64 28% 45% Quinara 52 50% 50% SAB 112 27% 44% Tombali 53 55% 50% 35 Facility Type Health Center type C 381 38% 48% Health Center type B 118 31% 47% Health Center type A 29 52% 51% Maternal and Child health center 8 25% 46% Regional Hospital 39 15% 37% Reference Hospital 30 13% 35% Location Rural 359 35% 48% Urban 246 34% 47% Total 605 34% 47% 3.11 – Health Worker Knowledge Methodological Note The choice of tracer conditions was guided by the burden of disease among children and adults, and whether the condition is amenable to use with a simulation tool, i.e., the condition has a presentation of symptoms that makes it suitable for assessing provider ability to reach correct diagnosis with the simulation tool. Three of the conditions were childhood conditions (malaria with anemia; diarrhea with severe dehydration, and pneumonia), and three conditions were adult conditions (pregnancy, pulmonary tuberculosis and diabetes). Two other conditions were included: post-partum hemorrhage and neonatal asphyxia. The former is the most common cause of maternal death during birth, and neonatal asphyxia is the most common cause of neonatal death during birth. The successful diagnosis and management of these seven conditions can avert a large share of child an adult morbidity and mortality. These indicators were measured using the patient case simulation methodology, also called clinical cases. Clinical cases are a widely used teaching method used primarily to measure clinicians (or trainee clinicians) knowledge and clinical reasoning. A vignette can be designed to measure knowledge about a specific diagnosis or clinical situation at the same time gaining insight as to the skills in performing the tasks necessary to diagnose and care for a patient. According to this methodology, one of the fieldworkers acts as a case study patient and he/she presents to the clinician specific symptoms from a carefully constructed script while another acts as an enumerator. The clinician, who is informed of the case simulation, is asked to proceed as if the fieldworker is a real patient. For each facility, the case simulations are presented to up to ten randomly selected health workers who conduct outpatient consultations. If there are fewer than ten health workers who provide clinical care, all the providers are interviewed. There are two other commonly used methods to measure provider knowledge and ability, and each has pros and cons. The most important drawback in the patient case simulations is that the situation is a not a real one and that this may bias the results. The direction of this potential bias makes this issue less of a concern —the literature suggests that the direction of the bias is likely to be upward, suggesting that our estimates can be regarded as upper bound estimates of true clinical ability. The patient case simulation approach offers key advantages given the scope and scale of the Service Delivery Indicators methodology: (i) a relatively simple ethical approval process is required given that no patients are observed; (ii) there is standardization of the case mix and the severity of the conditions 36 presented to the clinician; and (iii) the choice of tracer conditions is not constrained by the fact that a dummy patient cannot mimic some symptoms. Among the 255 health workers, from across all health facilities, who were interviewed using clinical vignettes, on average, health workers asked about or mentioned conducting 29% of actions outlined in basic clinical guidelines, gave a correct diagnosis to 31% of cases and mentioned the correct treatment for 52% of cases, when presented with the two cases of maternal and neonatal complications, health workers on average mentioned 39% of key steps for their management. There is nearly a 20-percentage point difference in adherence to clinical guidelines when we compare across regions ( 37 Table 15); while health workers in Oio display a knowledge of 38% of the clinical guidelines to assess the 6 basic cases presented in the vignettes, in Gabu health worker knowledge of guidelines is 20%. The knowledge for adherence to clinical guidelines also varies across health facility types where health workers in Health Centers Type A display the lowest knowledge (24%), the two health workers posted to the Maternal and Child Health Centers, the highest knowledge (61%). Health workers in Health Centers Type C know to adhere to less than a third of basic recommended clinical guidelines (29%) while health workers in Regional and reference hospitals display a slightly higher knowledge of 35% of clinical guidelines. Health workers posted to rural areas display slightly higher knowledge of clinical guidelines (31%) than do those posted to urban area facilities (24%). Health worker with higher level training, generally, have slightly higher knowledge of clinical guidelines than do those with less training but the differences are not statistically significant. Adherence to clinical guidelines also varies by the case presented in the vignettes (Figure 6). Overall, adherence to clinical guidelines is greater for Tuberculosis (53%) and significantly lower for the case of malaria with anemia (14%). The difference in adherence between Diabetes Mellitus (33%), Diarrhea with Dehydration (30%) and Pneumonia (31%) cases is not significant while adherence to ANC guidelines is slightly lower (23%). Across most cases, there are no significant differences between health workers with different training in their knowledge of clinical guidelines for each case. General Physicians display a higher knowledge of clinical guidelines than health workers with other training in the cases of Diarrhea, Pneumonia, Tuberculosis and Diabetes. In the case of Diabetes Mellitus both the general and the specialized physicians have significantly greater knowledge of clinical guidelines than health workers with other training. Figure 6: Adherence to Clinical Guidelines by Case and Training Adherence to Clinical Guidelines by Case and Training 70% 60% 50% 40% 30% 20% 10% 0% Diarrhea Pneumonia Malaria + anemia Tuberculosis Diabetes Mellitus ANC Specialist Physician General Physician Superior Level Nurse Medium Level Nurse Auxiliary Nurse Superior Level Technician Medium Level Technician Other There is a 15-percentage point difference in diagnostic accuracy across regions ( 38 Table 15); while health workers in Oio display a diagnostic accuracy of 37% of the 6 basic cases presented in the vignettes, in Bolama-Bijagos health worker diagnostic accuracy is 22%. Diagnostic accuracy also varies across health facility types where health workers in Health Centers Type A display the lowest knowledge (31%) and health workers in Regional Hospitals have a diagnostic accuracy of 39%. Health workers posted to rural areas display slightly higher, but not significant, diagnostic accuracy (31%) than do those posted to urban area facilities (30%). Health worker with higher level training, generally, have slightly higher diagnostic accuracy than do those with less training but the differences are not statistically significant. There is much variation in diagnostic accuracy across cases; cases with a combined diagnosis were very poorly identified by health workers of all levels of training (Figure 7). On average, diarrhea with dehydration was only diagnosed correctly by 14% of health workers, malaria with anemia was diagnosed correctly only by 2% of health workers and anemia in a normal pregnancy ANC consultation was diagnosed correctly only by 7% of health workers. As can be expected based on the results for knowledge of clinical guidelines, 85% of health workers are able to correctly diagnose a case of tuberculosis, 32% are able to identify a case of diabetes mellitus and 51% are able to diagnose a case of childhood pneumonia. General physicians have a higher diagnostic accuracy of Tuberculosis and Diabetes mellitus than do other health workers with different training. For Diarrhea with dehydration, pneumonia and normal pregnancy with anemia, there is no significant difference in diagnostic accuracy between physicians and nurses. Figure 7: Diagnostic Accuracy by Case and Training Diagnostic Accuracy by Case and Training 120% 100% 80% 60% 40% 20% 0% Diarrhea Pneumonia Malaria +anemia Tuberculosis Diabetes Mellitus ANC Specialist Physician General Physician Superior Level Nurse Medium Level Nurse Auxiliary Nurse Superior Level Technician Medium Level Technician Other The percentage of correct treatment knowledge of health workers across Guinea Bissau is low but slightly higher than the diagnostic accuracy; health workers might not know that an illness is or is called but have a slightly higher knowledge of how to treat it ( 39 Table 15). There nearly a 30-percentage point difference in treatment accuracy across regions; while health workers in Bolama-Bijagos display a treatment accuracy of 65% of the 6 basic cases presented in the vignettes, in Biombo health worker treatment accuracy is 36%. Treatment accuracy also varies across health facility types where health workers in Health Centers Type A display the lowest knowledge (43%) and health workers in Regional Hospitals have a treatment accuracy of 58%. Health workers posted to rural areas display slightly higher, but not significant, treatment accuracy (58%) than do those posted to urban area facilities (42%). Health workers with higher levels training, generally, do not have a higher treatment accuracy than do those with less training as the differences are not statistically significant. There is some variation in treatment accuracy between the different cases presented in the vignettes (Figure 8); 76% of health workers can correctly treat a case of malaria while 36% are able to correctly recommend treatment to a pregnant woman during an antenatal consultation. On average, 50% of health workers know to correctly provide full treatment to children with diarrhea, 62% can correctly treat a case of childhood pneumonia, 60% can treat a case of tuberculosis or correctly refer them to another facility and 49% can treat a case of diabetes or refer them to a higher-level facility. There is a significant difference in treatment knowledge between physicians and nurses in the treatment of malaria where nurses have significantly higher treatment accuracy than physicians, while physicians have a significantly higher treatment accuracy of diabetes mellitus cases than nurses. Across other cases, the difference in treatment accuracy between health workers of different training is not significant. Figure 8: Treatment Accuracy by Case and Training Treatment Accuracy by Case and Training 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Diarrhea Pneumonia Malaria + Anemia Tuberculosis Diabetes Mellitus ANC Specialist Physician General Physician Superior Level Nurse Medium Level Nurse Auxiliary Nurse Superior Level Technician Medium Level Technician Other The knowledge for the management of maternal and neonatal complications in Guinea Bissau is low, especially as most health facilities report the provision of delivery services where these complications would be faced ( 40 Table 15). There are nearly a 20-percentage point difference knowledge for the management of maternal and neonatal conditions across regions; while health workers in Bolama-Bijagos and Tombali display a management accuracy of 49% of the 2 cases presented in the vignettes, in SAB health worker management accuracy is 27%. Maternal and neonatal management accuracy also varies across health facility types where health workers in Regional Hospitals display the lowest knowledge (27%) and health workers in Maternal and Child Health, Type B and Type C health centers have a management accuracy of 42%. As with previous knowledge measures, health workers posted to rural areas display slightly higher, but not significant, management accuracy (44%) than do those posted to urban area facilities (30%). Health worker with higher levels training, generally, do not have a higher management accuracy than do those with less training as the differences are not statistically significant. Albeit low, there is significantly greater knowledge of the management of neonatal asphyxia (55%) than of post- partum hemorrhage (28%) among health workers in Guinea-Bissau (Figure 9). There are no significant differences across health workers with different levels of training for the management of either of the two cases presented in the vignettes. Figure 9: Management of Maternal and Neonatal Complications by Case and Training Management of Maternal and Neonatal Complications by Case and Training 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% PPH Asphyxia Specialist Physician General Physician Superior Level Nurse Medium Level Nurse Auxiliary Nurse Superior Level Technician Medium Level Technician Other 41 Table 15: Health Worker Knowledge Management of mat/neonatal Adherence to guidelines Diagnostic Accuracy Treatment Accuracy complications n mean sd mean sd mean sd mean sd Region Bafata 24 31% 13% 31% 16% 52% 22% 46% 23% Biombo 21 24% 16% 28% 14% 36% 23% 31% 23% Bolama_Bijagos 36 27% 9% 22% 17% 65% 23% 49% 19% Cacheu 38 36% 16% 34% 15% 62% 25% 44% 23% Gabu 34 20% 10% 35% 12% 61% 20% 26% 19% Oio 24 38% 13% 37% 16% 61% 27% 45% 23% Quinara 18 28% 13% 29% 11% 44% 24% 38% 27% SAB 40 22% 18% 31% 18% 37% 29% 27% 24% Tombali 20 34% 15% 30% 13% 44% 22% 49% 27% Facility Type Health Center type C 177 29% 14% 30% 15% 54% 26% 42% 23% Health Center type B 13 26% 13% 33% 18% 48% 24% 42% 28% Health Center type A 38 24% 16% 31% 15% 43% 28% 29% 25% Maternal and Child health center 2 61% 5% 33% 24% 82% 2% 42% 12% Regional Hospital 15 35% 15% 39% 12% 58% 27% 27% 20% Reference Hospital 10 35% 24% 37% 19% 44% 25% 38% 28% Location Rural 164 31% 13% 31% 15% 58% 22% 44% 22% Urban 91 24% 17% 30% 16% 42% 30% 30% 25% Health Worker Training Specialist Physician 11 35% 19% 35% 19% 50% 29% 36% 30% General Physician 32 39% 10% 44% 11% 58% 20% 46% 20% Superior Level Nurse 25 27% 14% 32% 18% 51% 24% 31% 25% Medium Level Nurse 145 29% 12% 29% 15% 57% 23% 43% 22% Auxiliary Nurse 4 31% 16% 33% 14% 65% 25% 48% 35% Superior Level Technician 4 36% 23% 13% 8% 42% 32% 44% 21% Medium Level Technician 15 23% 15% 31% 14% 49% 33% 31% 21% Auxiliary Technician 1 57% 50% 81% 50% Other 3 23% 3% 28% 10% 67% 15% 47% 10% 42 Total 255 29% 15% 31% 15% 52% 26% 39% 24% 43 Some recent training of health workers has had an effect on health worker performance in the vignette assessment. Health workers who received training on safe deliveries, adolescent health and neonatal resuscitation display a higher knowledge of maternal and neonatal complication management than those who received no training. Health workers who received training on HIV counseling and testing show a higher treatment accuracy than those who received no training. None of the different types of training seem to have increased the knowledge of adherence to guidelines or diagnostic accuracy for health workers when compared to those that did not receive recent training. Figure 10: Health Worker knowledge by recent training topic Knowledge and Recent Training 80% 70% 60% 50% 40% 30% 20% 10% 0% No training Safe Deliveries Adolescent Infection IMCI Noenatal HIV counseling other Health Control resuscitation and testing Adherence to guidelines Diagnostic Accuracy Treatment Accuracy Management of mat neonat compl 3.12 – Patient experience A total of 1,112 consultations of adults and children, undertaken by 229 health workers were observed across facilities in Guinea-Bissau: the assessment included 192 observations of adult and 47 observations of child (under the age of 5) consultations. From a set of defined actions based on clinical guidelines for general adult and child consultations, health workers on average conducted 29% of actions stipulated for an adult consultation and 28% of actions for a child consultation. When comparing the comprehensiveness of adult consultations observed, across different regions we find that health workers in Bolama-Bijagos conduct the smallest percentage of actions (22%) while health workers in Biombo are the most comprehensive (43%). Health workers in Maternal and Child Health Centers (45%) and Reference Hospitals (34%) undertake more comprehensive consultations than those in other health facilities. Health workers in Health Centers type A conduct the least comprehensive consultations (23%). There is no significant difference between the comprehensiveness of adult consultations in rural and urban 44 areas. When comparing across health worker functions, midwives conduct significantly more comprehensive consultations than health workers in other functions. Observed child consultations follow a similar pattern to that observed with adult consultations. Although the sample size is small, child consultations in Bolama-Bijagos were observed to be less comprehensive (14%) than those undertaken in Biombo (53%), Bafata (35%) or Quinara (35%). There is no significant difference between the comprehensiveness of child consultations observed in Health Centers Type C and B. Child consultations in rural areas (32%) were observed to be more comprehensive than those of urban areas (26%). Doctors and nurses conduct consultations that are similarly comprehensive (28 and 29% respectively). Table 16: Consultation Observations Adult Consultation Child Consultation Average Consultation score Total # of #HWs #HWs #HWs Mean SD Mean SD Mean SD Observ observed observed observed ations Region Bafata 88 16 26% 12% 5 35% 10% 19 29% 12% Biombo 40 12 43% 12% 1 53% 13 44% 11% Bolama_Bijagos 35 12 22% 9% 2 14% 1% 13 21% 9% Cacheu 106 29 32% 13% 5 31% 11% 36 32% 13% Gabu 132 25 33% 9% 4 23% 3% 26 32% 9% Oio 95 24 29% 11% 4 30% 11% 26 29% 11% Quinara 94 18 23% 9% 6 35% 10% 22 27% 13% SAB 445 39 28% 9% 18 25% 8% 54 27% 9% Tombali 77 17 28% 14% 2 30% 10% 20 29% 14% Facility Type Health Center type C 463 114 30% 12% 19 27% 9% 132 30% 12% Health Center type B 375 41 27% 12% 19 29% 11% 54 29% 13% Health Center type A 54 12 23% 9% 4 26% 14% 13 23% 10% Maternal and Child 19 3 45% 16% 0 3 45% 16% Health Center Regional Hospital 142 13 31% 7% 4 29% 9% 17 32% 8% Reference Hospital 59 9 34% 8% 1 38% 10 35% 8% Location Rural 403 116 29% 12% 17 32% 10% 128 29% 12% Urban 709 76 30% 11% 30 26% 10% 101 30% 11% Function Director 71 9 25% 10% 3 36% 17% 10 27% 12% Doctor 322 37 29% 9% 17 28% 7% 48 30% 8% Nurse 508 112 29% 13% 22 29% 12% 135 30% 13% Midwife 147 21 34% 10% 0 21 34% 10% Technician 39 10 30% 11% 3 19% 5% 11 28% 10% Other 25 3 24% 13% 2 28% 8% 4 24% 11% 45 Total 1112 192 29% 12% 47 28% 10% 229 30% 12% From an average of all consultations, results indicate that the more comprehensive consultations, by better skilled health workers are given in Biombo, Reference Hospitals with no apparent difference between nurses and doctors (Figure 11). Figure 11: Consultation score by health worker function Although the sample size is reduced, for child cases that presented with fever, pneumonia or diarrhea and adults presenting with a cough, we were able to analyze the adherence to guidelines by health workers as in the vignette cases, for the screening of malaria, pneumonia, diarrhea and tuberculosis accordingly . Results show that for children presenting with a fever, 39% of actions outlined in clinical guidelines to identify a case of malaria were followed by health workers. For pneumonia the percentage of these actions is higher when patients that presented with a cough were observed (42%) and slightly lower for patients presenting with diarrhea (25%). For adults presenting with a cough, 45% of recommended actions are taken for screening of tuberculosis. Table 17: Consultation observation; basic illness screening Pneumonia Tuberculosis Malaria Screening Screening Diarrhea Screening Screening # of #H Me S #H Mea #H Mea #H Mea Observations Ws an D Ws n SD Ws n SD Ws n SD Region Bafata 86 5 48 18 2 50 47 1 29 5 20 12 Biombo 40 1 100 0 0 3 56 35 Bolama_Bijagos 35 2 25 0 0 0 4 25 22 Cacheu 106 3 45 7 1 42 1 43 11 59 34 Gabu 132 4 27 8 0 1 14 3 67 29 Oio 95 4 36 9 2 54 6 1 39 3 75 22 Quinara 94 6 50 18 1 100 2 7 10 5 46 23 SAB 445 17 33 14 7 28 22 4 29 39 14 30 29 46 Tombali 77 1 46 0 0 2 75 12 Facility Type Health Center type C 463 17 37 14 5 45 32 6 30 30 22 42 33 Health Center type B 375 18 38 21 5 29 17 3 10 8 11 53 21 Health Center type A 54 4 43 22 1 100 0 7 30 29 Maternal and Child Health Center 19 0 0 0 1 50 Regional Hospital 140 3 51 5 1 17 1 43 6 57 36 Reference Hospital 59 1 38 1 58 0 3 39 54 Location Rural 401 15 42 17 4 56 34 3 1 22 26 51 31 Urban 709 28 38 18 9 35 27 7 28 28 24 38 30 Function Director 71 3 55 40 1 50 1 14 4 11 8 Doctor 320 14 40 11 4 38 17 1 43 16 49 33 Nurse 508 21 38 17 6 47 41 8 25 28 27 47 31 Midwife 147 0 0 0 2 62 31 Technician 39 3 26 21 2 31 20 0 1 17 Other 25 2 46 12 0 0 0 Total 1110 43 39 18 13 42 29 10 25 26 50 45 31 Patient satisfaction with outpatient services received, is generally high in Guinea Bissau but it varies across types of consultations and region. Results point to higher satisfaction for persons that had received a consultation for a child patient (89%) than those that received an adult consultation for themselves (85%), see Figure 12. The study finds large variation in satisfaction across regions. Satisfaction with child consultations were higher in Oio (95%) and in Cacheu (94%) than in all other regions. Satisfaction with child consultations was lowest in SAB (79%). Satisfaction with adult consultations was found to be highest in Bafata (100%) and lowest in Biombo (41%). Satisfaction for child consultations is highest among those attending Health Centers type C (86%), while satisfaction is higher for adults attending reference hospitals (91%), see sFigure 13. 47 Figure 12: Patient Satisfaction by type of consultation and region 100% 100% 94% 95% 90% 91% 90% 90%91% 87% 89% 90% 87% 86% 85% 85% 79% 79% 80% 68% 70% 60% 50% 41% 40% 30% 20% 10% 0% Bafata Biombo Bolama Cacheu Gabu Oio Quinara SAB Tombali Total Bijagos Child Adult Figure 13: Patient Satisfaction by type of consultation and facility type 100% 91% 91% 91% 87% 89% 90% 86% 84% 85% 82% 80% 67% 70% 60% 50% 40% 30% 20% 10% 0% Health Center Health Center Health Center Regional Hospital Reference Total type C type B type A Hospital Child Adult 2.14.3 Patient payment of user fees 48 Although Guinea Bissau has abolished user fees for children and mother care, patients continue to pay for their use of health services across the country. As can be seen in Figure 14, Only very few Health Centers type C and B currently charge for consultations of children while 25% of health centers type A and of Regional hospitals currently charge for these consultations. Although relatively few patients report payment for consultation fees, fees for tests and medicines continued to be charged all across different types of health facilities, on average 37% of patients currently pay for tests and 54% currently pay for medicines for a child consultation. Figure 14: Percentage of patients paying fees for child consultations by facility type 100% 100% 90% 80% 70% 67% 60% 57% 57% 54% 50% 50% 47% 43% 40% 37% 33% 30% 25% 25% 25% 20% 17% 10% 3% 4% 1% 0% 0% Health Center Health Center Health Center Regional Hospital Reference Total type C type B type A Hospital Fees for consultation Fees for tests Fees for medicines In contrast, adult patients are more likely to pay for consultations but fewer report payments for laboratory tests. Among adults interviewed, 18% reported payment for their consultation, 13% reported payments for laboratory tests and nearly half (47%) reported paying fees for medicines after their consultation. While Health Centers type C are less likely to charge for consultations and tests than other health centers, these facilities are more likely to charge for medicines. All patients interviewed at the regional hospitals reported having paid for consultations, tests and medicines. 49 Figure 15: Percentage of patients paying fees for adult consultations by facility type 100% 100%100%100% 100% 90% 80% 70% 60% 50% 43% 45% 40% 40% 33% 33% 30% 16% 17% 18% 20% 13% 10% 10% 0% 0% 0% Health Center Health Center Health Center Maternal and Regional Hospital Total type C type B type A Child health center Fees for consultation Fees for tests Fees for medicines 3.13 – Health Facility Management and Finance This section explores health facility management and finance. It first explores characteristics of directors and administrators in facilities – their education and experience, how they spend their time, what they identify as the bottleneck to the well-functioning of the health facility and what authority they have over various dimensions. The section then explores supervision and management practices, as well as existence of management tools such as work plans and staff registers. Then financial management, reporting and accounting practices are explored. Most importantly, we analyze sources of funding and fees in facilities. Following that, we consider in-kind receipts for the functioning of the health facilities. The main respondents of the health management and financing module were directors or administrator. Heads of facility have on average eight years of experience in the health sector. This ranges from an average of 13 years in SAB to an average of less than 5 years in Bolama. 83 percent of directors are responsible for a health facility for the first time. One third of directors have some experience outside of the health sector, on average 5.3 years. In Bolama and Tombali all directors are responsible for the first time, while in SAB 45 percent of directors were responsible for another facility prior to their current appointment. In both regions, Bolama and Tombali, centers are remote and can be difficult to access. In Bolama, 87 percent of centers are rural, in Tombali 75 percent. Thus, it appears that first-time directors are allocated in the least desirable regions, while more urbanized areas such as the capital receive more directors with prior experience. Figure 16 and Figure 17 shows directors’ experience by facility type and by region. Years of experience outside the health sector is calculated only for those who have experience. Directors of the reference hospitals have less experience in the health sector than the national average and the least experience as directors of any health facility. This might be because the top-level position at the National Hospital is more politicized than in lower-tier facilities. In consequence, it changes more frequently. Also, 50 lower-tier facilities are frequently managed by the most senior nurse or doctor, while reference hospitals can be managed by personnel without a medical background. Figure 16: Characteristics of Health Facility Directors across Facility Types 14 12 10 8 6 4 2 0 HC A HC B HC C MCC Reg Hosp Ref Hosp National Figure 17: Characteristics of Health Facility Directors Across Regions 14 12 10 Years in Health Sector 8 Years as Director of any health facility 6 Years as Director of this facility 4 Number of years outside of health 2 sector 0 The health center directors often live on the premises of the health facilities and are available seven days a week. In consequence, most (86 percent) responded to work more than five days a week and the average time worked is 13.7 hours per working day, with 33 percent responding to work 24 hours per day. Interestingly, by far the smallest average number of hours worked by directors is seen in Reference Hospitals (7.5 hours on average, compared to 13.7 hours national average). Most likely, these centers have more formalized work schedules, due to their size and number of staff, while in lower-tier facilities the responsible staff is permanently on duty for lack of staff. Logistic regressions show that whether a facility offers deliveries is a significant predictor for whether directors work seven days a week (Odds-ratio (OR) 120). However, this result needs to be interpreted with caution, because of the small sample size (n=132) and only six facilities do not perform deliveries. The clear majority of facility managers, 91 percent, give consultations in addition to their management duties. This number becomes smaller as the facility size increases: while directors do not give consultations in reference hospitals, in type-C health centers almost all of them (96 percent) give consultations and this is the activity 51 that takes most of their time in a typical day (26 percent of working time). In comparison, treatment of patients only takes 10 percent of directors’ time in a health center type A, 15 percent in regional hospitals and 17 percent in reference hospitals. In small, lower-tier facilities, all duties of the health center are concentrated on the few individuals working there and the director will assume various duties (administration, consultation, management, etc.). In hospitals, it is more likely that formal administration systems with staff exist which are separate from consultations. However, only a small difference is detected in the time dedicated to treatment of patients between rural (23 percent) and urban (25 percent) areas and the time dedicated to management (65 percent in rural, 64 percent in urban areas). Table 18: Use of director's time, by facility type Health Health Health Maternal Regional Reference National center type center type center type and child Hospital Hospital (%) A (%) B (%) C (%) care center (%) (%) (%) N 4 14 102 4 5 2 131 Consultation by director 50 93 96 75 60 0 91 (percent of directors) Share of time dedicated to: Consultations Treatment of patients 10 21 26 19 15 17 24 Management Patient flux 13 13 11 8 12 4 11 Supervision 7 9 6 11 12 16 7 Task distribution 3 5 4 3 8 4 4 Administration 40 17 24 28 23 10 24 Verifying equipment & 3 11 11 10 10 20 11 medicine Relations management 14 11 9 10 12 10 10 Other activities 12 12 10 11 9 20 10 *Note: Numbers do not add to 100% due to rounding. After treatment of patients, the highest share of time is dedicated to report writing and administrative activities. In regions with more rural facilities (Bolama/Bijagos, Tombali, Gabu), a higher share of time is dedicated to these tasks. Directors of type-A centers dedicate 40 percent of their time to administrative duties, while directors of reference hospitals dedicate only 10 percent of their time to this task, spending instead 20 percent of time on verifying equipment and medicine (compared to 11 percent nationally) and 20 percent on other activities (compared to 10 percent nationally). Directors identified lack of adequate infrastructure and lack of human resources as the main constraints to the well-functioning of their facility. 33 percent of directors responded that the lack of adequate infrastructure is 52 considered the main limiting factor, both in rural and in urban areas. This number reaches 50% in Bolama/Bijagos and in Biombo and is lowest, with 20 percent, in Quinara and Gabu. Availability of personnel was identified by 23 percent as the main limiting factor and 32 percent mentioned either availability of personnel or lack of trained personnel as the limiting factor. Lack of autonomy was mentioned by only 2 percent as the main limiting factor. Table 19 shows the difference between what directors of rural and of urban areas identify as the main limiting factor. Table 19: Limiting factor, percent of facilities who mention as the limiting factor, by rural / urban Rural Urban National N 82 50 132 Medicine Shortage 10 16 12 HR Shortage 23 22 23 Lack of trained 6 14 9 personnel Lack of adequate 34 32 33 Infrastructure Equipment shortage 13 6 11 Lack of autonomy 0 4 2 Other 13 6 11 There is some discrepancy when comparing these assessments by directors to the main complaints received by patients. Even though the answer options about patient complaint were not the same as the limiting factors, patient complaints nevertheless point to the main problems that patients perceive. 42 percent of patients mention either long waiting times or unavailability of health workers among their three main complaints. Long waiting times can, among other factors, be caused by a lack of staff or absent staff. 40 percent mention a long distance to the health center. This share is highest in Bolama/Bijagos (75 percent), where patients might have to take boats to reach the centers. Bolama/Bijagos is also the region with the highest share of rural centers (88 percent). Long distance was mentioned the least as a complaint in the small region of Biombo (0 percent), where 63 percent of centers are urban. Overall, facility managers are the main decision makers on most issues. Facility managers are the main decision makers on spending of fees and repairs, the two topics most relevant for ensuring adequate infrastructure. That lack of adequate infrastructure was most frequently identified as the bottleneck for ensuring the well-functioning of the facility is thus likely related either to scarcity of funds to make the relevant acquisitions and repairs or to mismanagement. For worker recruitment, national and regional governments were identified as the main decision makers. HR shortage and lack of trained personnel was the second-most cited problem. To the extent that this problem is caused by two few hired staff, it is thus up to the government, especially the national government, to solve this, rather than to facility directors. However, a recent Public Financial Management report pointed out the weak HR management in the health sector and that the inability of MoH to manage employment and placement of health-care staff was a major impediment to service delivery. The Ministry of Public Administration ( Função Pública) is responsible for hiring decisions, while the Ministry of Health determines the placement of personnel. 53 To the extent that this problem is caused by absenteeism, the solution lies with facility managers, since in 80 percent of facilities they are the main responsible for absenteeism approval. Decisions about fees lie with governments rather than with the facilities. Patient fees and sale of medicine are among the most important sources of income for health facilities. A great majority of facilities mentioned facility directors as the main decision maker over medicine requests (73 percent). Nevertheless, 16 percent of facilities (20 percent in rural, 10 percent in urban areas) mention unavailability of medicine as one of three main complaints by patients. A great majority of facilities mentioned facility directors as the main decision maker when it comes to absenteeism approval (80 percent) and disciplinary actions (58 percent). Staff attendance registers are present in 85 percent of facilities. However, this number is lower in rural facilities (79 percent) as compared to urban facilities (96 percent). There is great variation between regions: While staff attendance registers exist in all facilities in Biombo, Bolama/Bijagos, Oio and SAB, only 38 percent of facilities in Tombali have an attendance register. Interestingly, number of reported absent days per health worker does not appear to be negatively correlated with the existence of a register (r=0.11). Instead, the number of reported absent days per health workers in facilities with a register was 1.7, while the number in facilities without a register is 0.7 over a 30-day period. However, when observing attendance of health workers among a random sample during unannounced visits, 34 percent of health workers were absent (excluding those doing field work, those who were reported to have finished for the day during time of visit and those off duty and on holidays). This translates to 10 absent days per health worker over a 30-day period, suggesting that the number of absent days are underreported by a factor of roughly 6. In self-reported data, there is a higher number of absentees in rural facilities with 1.27 worker absences per day, as opposed to 0.47 in urban facilities over a 30-day period. Rural facilities managed to replace more absent staff than urban facilities. With a replacement rate of 67 percent in rural facilities and an absence rate of 1.27 worker-days per health worker (self-reported) this means that effectively there is a shortage of almost half a day per health worker in rural areas. For urban facilities, this number is halved. However, this is only among the staff the facility reported as absent and we cannot tell which percentage of staff not reported absent is replaced. There is one health facility with 300 absent health worker days in the last 30 days (14.3 per health worker, health center type A in Cacheu) and one facility with 18.5 absent days per health worker in the last 30 days (health center type C in Bolama/Bijagos). These are extreme events. Both centers have attendance registers, both centers report HR shortage to be the biggest issue. In both facilities, patients rated either long waiting times or health workers unavailable as their biggest complaints. The center in Cacheu also reports that the director of the health facility is the main decision maker on absenteeism approval. It is strange then why no action is taken. 54 Table 20: Percentage of facilities in which a staff attendance register exists, average number of absent days during the past 30 days and percentage of absent days in which a replacement was found, by rural/urban Rural Urban National N 81 50 131 Staff attendance register (percent of facilities) 79 96 85 Number of absent worker-days per health worker 1.27 0.47 0.97 N* 41 39 80 Absent staff replaced (percent) 67 54 61 *The question of how many absent staff were replaced was only asked for centers who registered absent staff in the last 30 days. Therefore, the number of respondents for this question is different. Besides supervision of health workers by facility directors, facilities are also supposed to be visited regularly by ministry staff. On average health centers received seven visits by MoH in 2017, but this number ranges from 0 to 68 visits (more than one per week), which shows that facilities received very different attention. While facilities in Gabu received visits on average almost once a month, facilities in Bolama/Bijagos received visits only twice annually in 2017. In all regions except for Oio and Cacheu there is at least one center that has not received any supervision visit. Smaller centers received more visits than bigger centers: 9 and 7 visits on average for type B and type C centers respectively, 5 visits on average for type A health centers, 4 visits in regional hospitals and no visits were made in reference hospitals. This is likely because supervision is usually performed by regional health directorates, but reference hospitals fall under the authority of the central level. The average duration of visits also differs between regions, from only 1.7 hours in Bafata to 8.7 hours in Bolama/Bijagos. In this region, some visits were recorded as having taken 24 hours. This is likely because the supervisors had to travel by boat to the health center and couldn’t travel back the same day. Even when all supervision visits over 10 hours are excluded, Bolama/Bijagos still has the longest average time of visits (4.6 hours, compared to the shortest of 1.7 in Bafata). There is a slightly negative relationship between number of visits and duration of visit (r=-0.09), meaning that facilities that receive fewer visits are visited for longer. Once visits longer than 10 hours are excluded and we exclude centers that received more than 24 visits (2 per month), the relationship disappears, as can be seen in the table and graph below, summarizes the number and duration of supervision visits by facility type. 55 Table 21: Supervision visits by MoH in 2017, by Facility type Facility type HEALT HEALT HEALT MCC Reg Ref National H H H Hosp Hosp CENTE CENTE CENTE R TYPE R TYPE R TYPE A B C N 4 14 102 4 5 2 131 Average number of visits 5 9 7 2 4 0 7 Minimum number of visits 2 0 0 0 0 0 0 Maximum number of visits 7 27 68 4 8 0 68 N* 4 13 95 3 3 0 118 Average duration (hours) 10.3 4.7 4.1 2.7 3.3 N/A 4.3 Minimum duration of last 3 1 0.5 2 1 N/A 0.5 supervision visit (hours) Maximum duration of last 24 12 24 3 8 N/A 24 supervision visit (hours) Figure 18: Relationship between number of supervision visits and duration, excluding visits > 10 hours and centers with > 24 visits per year During supervision visits, the focus appears to be on medical aspects and inputs and less on financial management. Only in 34 percent was the finance registry reviewed. While the finance registry was not reviewed in any facility in Tombali, it was reviewed in 70 percent of visits in SAB. Similarly, items related to finance and administration were mentioned in a review book in only 26 percent of facilities while 43 percent remarked on 56 materials and medicine. Thus, there appears to be less awareness by supervisors on the importance of financial management in health facilities or they are lacking the knowledge to review these topics. In principle, most reviews appear to be conducted according to standard procedures: 92 percent of reviewers use guides or control lists, 89 percent organize meetings with personnel and 72 percent observe consultations. Also, 92 percent of facilities received a written report after their last supervision visit. Table 21 shows which percentage of facilities received a feedback report and which topics were mentioned in the supervision book. Table 22: Reception of feedback and themes/issues mentioned in supervision book (percent of facilities), by region Region Tombali Quinara Oio Biombo Bolama- Bafata Gabu Cacheu SAB National Bijagos N 12 9 15 7 15 12 18 21 10 119 Reception of 100 78 100 86 100 83 83 95 90 92 Feedback Topics mentioned in book: Finance and 0 0 21 40 20 29 67 25 33 26 Administration Infrastructure 45 17 36 33 20 33 0 40 60 31 and Equipment Material and 9 33 47 50 50 71 44 40 50 43 Medicines HR 50 17 36 33 67 38 50 20 57 42 Quality of care 30 33 13 50 46 63 11 21 75 33 Epidemiology 50 71 20 33 7 38 50 47 40 37 64 percent of centers have prepared a work plan for 2018, but only 38 percent of centers elaborated a quarterly workplan. In Bolama/Bijagos no facility has a quarterly work plan, while 75 percent of facilities in Quinara have a quarterly work plan. Urban areas are more likely to have a quarterly plan (48 percent vs. 32 percent in rural areas) even though there is little difference between urban and rural areas in having an annual work plan. No reference hospital has a quarterly plan, while all regional hospitals have one. On average, 31 days were needed for the approval of work plans. This ranges from only 3 days in Gabu to 93 days in SAB. The overwhelming majority of centers who experienced delays in their work plan lay the blame on delays in the ministry of health (83 percent vs. 17 percent of facilities who say the plan was submitted late). This is maybe not surprising, since the answers were self-reported and there was no way to verify. Table 23 shows the existence and time for approval of work plans. 57 Table 23 Existence and quality of work plan, by region Region Bolama- Tombali Quinara Oio Biombo Bafata Gabu Cacheu SAB National Bijagos N 16 10 15 8 16 15 20 21 11 132 Workplan for 2018 56 40 67 100 75 73 40 76 55 64 Quarterly workplan 33 75 70 25 0 36 25 63 17 38 Days for approval of 44 26 22 40 N/A 16 3 33 93 31 Quarterly workplan Late approval (percent of 0 0 0 0 N/A 50 50 30 0 19 facilities) Reasons for delay in approval Late submission N/A N/A N/A N/A N/A 50 0 0 N/A 17 Delay in approval by MoH N/A N/A N/A N/A N/A 50 100 100 N/A 83 84 percent of health centers submitted a financial report for the last quarter, with no difference between urban and rural facilities and only small differences between regions. This shows a strong commitment to financial reporting. However, it is not clear if the financial report is read or of what quality it is. A recent Public Financial Management (PFM) report notes that “accounting and reporting functions are not satisfactor y. Data and reports are replete with errors and often inconsistent with each other�. It also notes that financial reports prepared at the central level do not include internal revenues and expenditures. Among centers that did not submit a financial report, 57 percent mentioned that it was not ready, but also around 20 percent stated that it was not requested. None of the reference hospitals had completed reports, while all maternal-child centers, four out of five regional hospitals and almost all health centers of type A, B and C had presented their reports. Support of financial management by the ministry and its partner projects is generally good. Visits by an accountant from the ministry or one of its partner projects (PIMI/H4+/EU Saúde) usually occurs monthly (79 percent of facilities), but 31 percent of facilities in Bolama/Bijagos have never been visited. Most facilities (77 percent) have obtained some assistance in financial management in the form of receipts or cash books, payment receipts or other tools to keep track of their finances. However, while in Gabu 95 percent of facilities received assistance, only 44 percent in Tombali received any management tool. Centers that did not receive financial management tools and centers that were visited less than monthly by a superior accountant were less likely to present financial reports (OR 4.0 and 1.7, respectively). Whether a facility has a staff member responsible for accounting has no influence on whether the facility submits a financial report. However, the probability of having a staff member responsible for financial accounting increases with the facility’s size and level of service: While 25 percent of maternal-child centers and 46 percent of health centers type C have staff responsible for accounting, 100 percent of reference and regional hospitals have such a position. In consequence, more facilities in urban areas than in rural areas have an accountant (70 percent vs 41 percent). Table 24 summarizes financial management variables by facility level. 58 Table 24 Financial Management, by facility type Health Health Health Materna Regional Reference National Center Center Center l and Hospital Hospital type A type B type C Child Health Center N 4 14 103 4 5 2 132 Presentation of Financial report 75 93 84 100 80 0 84 (percent of facilities) Reason for not submitting Financial Report (percent of facilities) Not ready 100 100 44 100 100 100 57 Bank conciliation not done 0 0 6 0 0 0 5 Not requested 0 0 25 0 0 0 19 Other 0 0 25 0 0 0 19 Reception of financial management tools (percent of facilities) Receipt Book 50 50 44 50 40 0 45 Payment receipt 50 14 30 0 0 0 26 Cash book 75 64 51 25 60 0 52 All of the above 50 14 21 0 0 0 20 Other 25 7 14 0 0 50 13 None 0 21 22 50 40 50 23 Visited by MoH or partner accountant (percent of facilities) Staff member responsible for 75 79 46 25 100 100 52 accounting (percent of facilities) Monthly 75 93 78 100 80 0 79 Quarterly 25 7 10 0 20 50 11 Bimestrial 0 0 2 0 0 0 2 Semiannually 0 0 3 0 0 0 2 Never 0 0 8 0 0 50 7 User fees for services are only displayed in over half of the centers nationally. These fees are not standardized nationally, varying between and within regions. 45 percent of facilities do not display any fees, which leaves patients in the unknown about the charges they will face when seeking treatment, makes them vulnerable to arbitrary charges and is contrary to national legislation (DL No. 4/1997). Thus, it is possible that fees are even different from patient to patient within the same facility. Urban facilities display fees more frequently (63 percent vs. 49 percent in rural facilities), but services in these centers are generally more expensive. In every center in SAB a list of fees was 59 found, while only 19 percent of centers displayed fees in Tombali. One hypothesis to explain this is that control and monitoring of these practices is easier and more frequently done in urban areas. Table 25 shows the fees by region. It is of note, however, that facilities in which fees are displayed report higher fees for 20 out of 27 services. Thus, it appears that either there is a systematic bias in the facilities that display fees or facilities that do not display fees underreported their charges in the questionnaire. SAB is by far the most expensive region, showing the highest average fees among all regions for registration, consultations of over-5 year olds, consultation and treatment of typical child diseases (pneumonia, diarrhea, malaria), services for under-5 year olds, chronicle disease consultation and treatment, small surgeries, lab tests and rehydration salts. All facilities display that malaria testing and treatment is free of charge, which is in line with national policy and financed by the Global Fund. Despite the efforts to introduce user fee exemptions for maternal and child health services (gratuidade), fees are still charged for these population groups in 27 percent of the facilities. Despacho No. 24/GMSP from December 17th, 2013 declares reproductive health services and services for children below 5 years of age to be free of charge. This policy is supported by the EU-financed program PIMI II, which pays facilities for each of these services provided. The average fee is 229 XOF, with the highest average fees charged in Oio (856 XOF). Oio is also the region with the highest share of facilities charging for these services (87 percent of facilities), while in Tombali and Gabu no facilities charges for these services. Average charges in Bafata, Quinara and Bolama/Bijagos are relatively low (6, 7 and 52 XOF, respectively), while fees in Biombo, Cacheu and Oio are quite high (between 365 and 856 XOF, on average). All except regional hospitals charge for reproductive health services, but only health centers of type C charge for services for under-5 year olds (average 7 XOF, but up to a maximum of 500 XOF). Urban areas charge more for reproductive health services (273 XOF vs. 203 XOF) and for under-5 year olds (11 XOF vs. 3 XOF) than rural areas. 60 Table 25: Facility Fees Region Tombali Quinara Oio Biombo Bolama- Bafata Gabu Cacheu SAB National Bijagos N 16 10 15 8 16 15 19 20 11 130 Displayed (percent of facilities) 19 70 40 88 44 47 58 60 100 55 Average (Maximum) fee, in XOF, for Registration / admission 0 0 345 (1000) 125 (500) 313 (1000) 542 (1000) 1000 263 (1000) 750 (2000) 328 (2000) (1000) Consultation >5 years 297 (500) 500 (2500) 257 (600) 338 (1500) 188 (500) 367 (500) 221 (500) 265 (500) 550 (1500) 309 (2500) Consultation <5 years 0 0 0 144 (1000) 0 17 (250) 0 38 (500) 100 (500) 24 (1000) Child services** 188 (500) 125 (250) 57 (600) 186 (1000) 31 (250) 228 (1000) 117 (500) 132 (500) 533 (1500) 160 (1500) Child services <5 years*** 0 0 0 0 0 18 (250) 0 0 50 (500) 6 (500) Chronicle disease consultation & treatment* 182 (500) 142 (333) 55 (500) 78 (375) 194 (500) 169 (1000) 21 (250) 91 (333) 300 (1000) 128 (1000) Reproductive health services* 0 7 (67) 856 (1667) 365 (1667) 52 (833) 6 (83) 0 533 (1667) 183 (1000) 229 (1667) Small Surgery 394 (1000) 370 (750) 360 (2000) 625 (1300) 263 (1000) 439 (1000) 439 (2000) 495 (1000) 690 (2000) 437 (2000) Lab tests 145 (708) 319 (900) 106 (1000) 139 (375) 73 (583) 114 (917) 73 (583) 325 (2000) 775 (1167) 222 (2000) Chest X-Ray N/A* N/A 1000 (5000) 0 0 2500 (5000) N/A 3125 1250 926 (12500) (12500) (2500) Rehydration salt 0 25 (250) 0 0 9 (150) 0 18 (250) 0 28 (250) 8 (250) Oral Hypoglycemia 0 0 0 0 0 0 1500 0 533 (1000) 89 (1500) (1500) Malaria treatment 0 0 0 0 0 0 0 0 0 0 Amoxicillin 178 (750) 293 (1000) 138 (1250) 766 (1000) 56 (900) 280 (1000) 188 (1000) 245 (1000) 673 (1500) 268 (1500) *The values shown are averages, with the maximum fee in the category in parentheses. **Average for child consultation for pneumonia, diarrhea and malaria ***Include immunization, treatment of diarrhea and of acute respiratory infections for children below 5 years of age. 61 Every center has some group that is exempt from paying fees for services. However, these exemptions are self- reported by the centers and were not reported by patients themselves or displayed publicly in the facilities. Elderly patients are the main group exempted from fees (in 81 percent of centers) . This number reaches 92 percent in urban facilities and is lower in rural facilities (74 percent), possibly because elderly make up a smaller share of patients in rural areas. In Bolama/Bijagos, old people are exempt in only 19 percent of centers, but apart from that the share of facilities ranges from 73 percent of facilities (Bafata) to 100 percent of facilities (Tombali). The second group exempt from fees are pregnant women (in 76 percent of centers). Interestingly, pregnant women are exempt in all centers in Oio. It is thus not clear whether the fees displayed in this center are charged to the patients, if those fees are the fee for service received from PIMI or if the facilities misreported who is exempt. Again, the exemption is higher in urban areas as opposed to rural areas, but in line with national policy of free treatment for pregnant women, 100 percent of facilities should have exemptions. However, this target is only reached in Oio, Gabu and Cacheu. SAB and Biombo fall slightly short of it (91 percent and 88 percent of facilities, respectively), while in Quinara, Tombali and Bafata just over half of facilities exempt pregnant women (73 percent, 56 percent and 50 percent, respectively). Facility staff benefit from fee exemption in 61 percent of centers, and their family in 28 percent of centers. In addition, local politicians benefit from fee exemption in 17 percent of centers. In Bolama/Bijagos this group is exempt in 75 percent of centers, while pregnant women and children under 5 are exempt in only 13 percent of centers in this region. It seems that in Bolama/Bijagos fee exemptions are decided based on social status rather than patient need. Besides exempting local politicians, 75 percent of centers in Bolama/Bijagos do not charge management committee members, 63 percent do not charge family of staff and 44 percent do not charge their staff. In contrast, in only 19 percent of facilities patients with chronicle diseases and elderly are exempt, and in only 25 percent of facilities are poor people exempt. Nationally, however, it appears that needy patients are exempt rather than politicians and people connected to facility staff (see Table 26). However, given that these were self-reported exemptions, it might also be that centers in Bolama/Bijagos were simply more honest about their exemptions than other centers. 62 Table 26: Facility Fee exemptions Region Tombali Quinara Oio Biombo Bolama- Bafata Gabu Cacheu SAB National Bijagos N 16 10 15 8 16 15 20 21 11 132 Exemptions for (percent of facilities) Patients with chronic diseases 29 40 87 63 19 27 90 79 73 58 Old patients 100 90 93 75 19 73 90 95 91 81 Poor people 81 70 93 88 25 53 95 80 64 73 Staff of facility 69 90 53 38 44 47 85 67 45 61 Pregnant women 56 50 100 88 13 73 100 100 91 76 Family of staff 13 40 13 25 63 27 15 33 27 28 Management committee members 0 10 7 13 75 0 25 19 27 20 Local politicians 13 30 0 0 75 27 5 0 9 17 Children under 5 56 50 80 88 13 73 100 71 73 67 Others 0 20 0 0 25 7 10 5 9 8 63 On average, health facilities declare to have had 10 stock ruptures in the last 12 months . A stock rupture is counted for any continuous period of at least one day that a drug that is supposed to be available has been missing. Health centers of type A and type C experience most stock outs (17 and 11 on average, respectively). Maternal-Child Centers, health centers of type B and regional hospital experience 7, 6 and 6 stock outs annually on average, while reference hospitals had 2 to 3 stock outs. Reference hospital have much higher requirements on the variety of medicines they stock than lower-tier facilities. That they have fewer stock outs than smaller health centers suggest that they have a better stock management system. Figure 19: Number of stock outs by facility typeshows histograms of stock outs by type of facility. One urban center of type C in SAB recorded 150 stock outs, by far the highest number. One center in Biombo and one center in Bolama/Bijagos recorded 90 stock outs. Bolama/Bijagos and SAB recorded the highest average number of stock outs among all regions (23 in the last 12 months). For Bolama/Bijagos this can be explained by the long-distance and time required for the medicines to reach the centers. In the case of SAB this might show that the high use of medicine – due to higher number of patients – outpaces the delivery of medicine. The lowest average number of stock outs occurred in Tombali, Quinara and Cacheu, with 4 stock outs on average per center. The number of stock outs is uncorrelated with whether the directors perceive medicine shortage to be the main limiting factor and with whether availability of medicine ranks among the three main complaints among patients (r=-0.01 and r=0.03, respectively). In addition to stock outs, 56 percent of centers declare to have had expired medicine in the last quarter. In Health centers type A this affects three-quarters of all facilities, while two out of five regional hospitals are affected. In Bolama/Bijagos, 88 percent of facilities had expired medicine, compared to only 35 percent in Gabu. Over the past quarter, half of facilities purchased medical articles, and among them 70 percent made their purchase at the local level. In rural areas, where medical articles are scarcer, a smaller number of facilities purchased medicine at the local level (65 percent) as opposed to urban centers (78 percent). Figure 19: Number of stock outs by facility type 64 The majority of facilities, approximately 70 percent, adopt a pull strategy to receive medicine. This means that facilities actively ask for medicines, rather than receiving them automatically in regular intervals. Table 27 shows the use of a pull and push strategy and compares stock outs by strategy. Facilities that follow a push strategy are more likely to have experienced four to five or more than five stock outs over the past twelve months than the national average and centers that follow a pull strategy. Similar shares of rural and urban centers use pull strategies (70 percent in rural areas vs. 68 percent in urban areas). All centers in Oio and Cacheu use a pull strategy, compared to only 11 percent of centers in Gabu. In all regions besides Gabu, at least 50 percent of centers use a pull strategy. Reference hospitals (100 percent) and health centers type C (74 percent) are most likely to use a pull strategy, compared to centers type B (64 percent), MCC (50 percent), regional hospital (40 percent) and centers of type A (25 percent). Table 27 Stock ruptures, by strategy for receiving medicines Pull Push National N 91 40 131 Stock ruptures, past 12 months 8.5 13.5 10 (average) 0 (percent of facilities) 16 3 13 1-3 (percent of facilities) 36 28 33 4-5 (percent of facilities) 11 15 12 >5 (percent of facilities) 36 55 42 Health centers made on average 10 medicine requests to their suppliers over a 12-month period, almost one per month. Urban facilities made on average one more request than rural facilities, possibly because they are larger and require more medicine. Rural facilities wait on average 17 days between their request for medicine and its receipt. Urban facilities wait 23 days6. The longest average waiting time is endured in Bafata, where it takes 41 days from request to receipt. In Cacheu it also takes just over one month. The shortest waiting time is in Tombali (8 days), followed by Oio (10 days). SAB, with 20 days, is close to the national average (19 days). These results are a bit surprising, since it seems that distant areas (rural areas, Tombali) receive medicines faster than areas close to the harbor, CECOME and the headquarters of most NGOs (urban areas, SAB). It thus appears that distance to a distribution center is not the bottleneck when it comes to medicine distribution. On the other hand, reference and regional hospitals have the shortest waiting times (4 and 12 days, respectively), while MCCs have the longest expected waiting time (35 days). Health centers type A, B and C wait between 19 and 21 days. The survey enquired whether a variety of medicines, supplies and equipment were received in the past three months and who supplied these. The medicines asked for included antimalarial medicine, various common antibiotics, vaccines, nutritional supplements, and other essential medicines. Supplies included basic equipment for newborn, child and maternal health (balances, resuscitation masks, measuring bands, Doppler apparatus, etc.) and basic equipment for general care (stethoscope, thermometer, blood pressure monitor, glucometer, hemoglobin meter). 95 percent of facilities had received at least some medicine and received on average 57 percent of the medicines enquired. However, while facilities in Oio received 84 percent of medicines, facilities in Bolama/Bijagos received only 13 percent. The only facilities who have not received 6 This excludes one outlier in Gabu, where the time from request to delivery was recorded as two years. 65 any medicine were seven rural health centers type C. As expected, rural centers received less medicines compared to urban centers (52 percent vs. 65 percent). Unexpectedly, health centers type B reported a higher share (74 percent) of medicines received than centers type A (58 percent), than regional hospitals (66 percent) and reference hospitals (33 percent). 86 percent of facilities received supplies, on average 5 of those enquired, but only 36 percent of facilities received equipment, and on average only 2 items of the equipment enquired. The PIMI program through its implementing NGOs EMI and IMVF is by far the largest supplier of the enquired medicines, supplies and equipment, supplying 74 percent of medicines, 60 percent of supplies and 53 percent of equipment. It is important to note, however, that this is by type of medicine and not by volume or person treated. In Bolama/Bijagos, only 20 percent of medicines were from PIMI/EMI/IMVF, compared to at least 64 percent of medicines in other regions. PIMI accounts for only 57 percent of medicine received by MCCs, even though it is a program focused on maternal and child health. Other suppliers of MCCs include CECOME (15 percent), GAVI (13 percent) and Global Fund through its implementation unit UNDP (4 percent). The second largest national supplier is CECOME, supplying only 8 percent of the enquired medicines, 17 percent of supplies and 14 percent of equipment. The official national distributor of medicines is therefore dwarfed by other programs and organizations. Other suppliers such as UNFPA, UNICEF and UNDP (Global Fund) have a smaller share in the supply of medicines, supplies and materials as these organizations have a very specific role (provision of family planning, maternal and neonatal medicines and material, or malaria treatment). During the last supervision visit conducted by MoH, the medicine stock was reviewed in a vast majority of centers (69 percent). This frequent supervision can explain that in almost all facilities (91 percent), the stock monitoring register was updated. The percent of facilities with an updated stock monitoring register is relatively constant across regions, reaching from 69 percent in Bolama/Bijagos to 100 percent in Bafata, Biombo and Oio. However, medicine delivery is not constantly checked by MoH. The last delivery was verified for less than half of facilities (42 percent). In Gabu, medicine deliveries were verified in only 5 percent of facilities, whereas it was verified for 76 percent of facilities in Cacheu. Deliveries were verified for only 37 percent of health centers type C, for 50 percent of type B, A and reference hospitals, for 75 percent of maternal-child centers and for 80 percent of regional hospitals. This type of supervision is more common in urban facilities (56 percent) than in rural facilities (33 percent). This result is aligned with the fewer stock ruptures and requests made in rural facilities as seen previously. In addition, facilities were asked about the human resources responsible for MEMM stock. On average, two people oversee managing MEMM in health facilities, among which one received a specific training for this purpose. More training might be needed in regional hospitals, where only 18 percent of managing MEMM have received any. Over a third of facilities (36 percent) were visited monthly by a pharmacist in 2017 for supervision. However, another third (30 percent) was never visited in 2017. This shows that supervision visits by pharmacists are either very frequent or nonexistent. This result is even more visible in rural areas where 42 percent of facilities were visited monthly, while 32 percent were never visited. This section explores the links between health facilities and the community. It aims to present results related to community participation in management committees, and the contributions to facilities made by community members. In addition, the accountability and transparency of health facilities towards the community is discussed. Management committees only exist in 13 health centers, or 10 percent of all facilities in the country. Only centers of type B (21 percent) and type C (10 percent) have management committees. On average, committees are composed of 3 members, including a small share of female members (10 percent). While the share of facilities with a management committee is the same in urban and rural areas, committees of urban centers seem to be more active, meeting on average at least once a month, while rural centers meet only seven times per year (national average: 9 meetings in the past twelve months). The community leader participated in 70 percent of centers where these management committees exist. Again, this participation is higher in urban centers (80 percent) than in rural centers (63 percent). Meetings minutes are available 66 in 54 percent of centers where community committees exist. However, contrarily to the number of meetings, meeting minutes are more available in rural centers. Table 28: Health Management committee by rural/urban Rural Urban National N 82 50 132 Existence of Committee (% of facilities) 10 10 10 Number of members 2.3 3.4 2.7 Number of meetings, past 12 months 7.3 12.6 9 Average number of female members 0.5 0 0.3 Share of female members (%) 18 0 11 Community leader participation in management committee (% of centers) 63 80 69 Meeting minutes available (% of centers) 63 40 54 Note: Centers with a management committee are: Canjambari (type C, Oio), Mansaba (type C, Oio), Olossato (type C, Oio), Nhacra (type C, Oio), Bijimita (type C, Biombo), Dorse (type C, Biombo), Pitche (type C, Gabu), Majanco (type C, Gabu), Sonaco (type B, Gabu), Barro (type C, Cacheu), Ingore (type B, Cacheu), Caliquisse (type C, Cacheu), Bairro Militar (type B, SAB) 41 percent of committee members were appointed by the community. This is the first nomination method, before appointment by a local authority (28 percent). Although the existence of management committees is limited, this shows a strong role of the community in the management of some centers. The vast majority (84 percent) of community-appointed members were elected, while the others were simply nominated. The share of members elected by the community is higher in centers of type C (45 percent) in comparison with centers of type B (16 percent). Lower-tier facilities tend to be more integrated into communities, which can explain the stronger role played by the community in such centers. The education level of members appointed by the community varies: 46 percent studied in secondary levels but did not complete it, and 23 percent completed secondary schools. In the context of Guinea-Bissau, this shows that the members chosen by the community are rather educated. These numbers are good indicators but need to be taken with caution as the population studied is very small. Figure 20: Committee Member Appointment Mechanism 9% Appointed by community 22% 41% Appointed by local authority Appointed by facility head Appointed by syndicate 28% Management committees are included in decision making over expenses of many items. Almost all committees (92 percent) decide on expenses of medicine and non-health staff salary. But it is worth reminding that this is the case in only in 10 percent of centers. They also participate in the decision on expenses related to medical material, cleaning material (83 percent), and medical equipment (67 percent). The expenses on which management committees decide the least concern the salary of health workers (33 percent). However, it should be noted that in urban centers, committees 67 have a stronger decision power on this topic (60 percent) than in rural centers (14 percent). However, it is not clear what part of the health-worker salary the committee decides on (top ups, salary for interns off the official payroll, substitution of unpaid salaries, or something else entirely), since the salary of health staff is decided on the central level by the Ministry of Finance and the Ministry of Public Administration. Table 29 Decisions in committee meetings, by rural/urban Rural Urban National N 7 5 12 Committee meetings to decide on expenses of (% of facilities), past year Medicine 86 100 92 Medical material 71 100 83 Medical equipment 57 80 67 Cleaning material 71 100 83 Health worker salary 14 60 33 Other staff salary 88 100 92 Public utility 43 60 50 Construction / Maintenance 43 80 58 Other 50 50 50 Community contribution was also investigated as it is key to understand the relationship with health centers. Half of health facilities received the contribution of the community through different means, with little difference between urban and rural areas. The most common contribution from the community is made through work and time given to help centers with construction, cleaning, gardening etc. This is the case in about half of the centers, as much in rural as in urban areas. However, not all regions see the same extent of contribution of such type. The percentage of centers where the community engages with work ranges from one quarter in Tombali to three quarter in the region of Biombo. Interestingly, Biombo is the region where the lack of adequate infrastructure was mentioned the most by the facility head as the main limiting factor. This can explain the strong help received from the community to improve the infrastructure. In Bissau, this type of contribution is lower (36 percent) than the national average. In-kind contributions (donation of equipment, medicines, material etc.) only occur in 5 percent of centers country-wide. Such in-kind community contribution only benefits centers of type B and C. This type of contribution is more frequent in urban centers (10 percent) than in rural centers (1 percent). This might be due to the scarcity of equipment, medicines and material to give to facilities in rural areas. Additionally, no regular financial contribution is done at all in any centers. It is worth noting that there is no significant correlation between community contributions and the level of funding or the number of patients treated in centers. 68 Table 30 Community contributions by rural/urban Rural Urban National N 82 50 132 Any type of regular contribution (% of centers) 48 52 50 Regular in-kind contributions by community (% of centers) 1 10 5 Regular work-time contribution by community (% of centers) 48 50 48 Regular financial contribution by community (% of centers) 0 0 0 Over a third of facilities have a formal mechanism to receive patients’ feedback. This practice is more common in urban centers (42 percent) than in rural areas (32 percent), and it varies across regions. It does not exist in Tombali, but reaches 55 percent of centers in Gabu and SAB. In addition, maternal and child centers do not have formal feedback mechanisms, while it varies across other facility types: higher-tier centers that receive more patients tend to implement this practice more. 60 percent of regional hospitals and 50 percent of type-B centers have a feedback mechanism, while this the case only in one third of centers of type C. However, centers of type-A are not aligned with this trend, as only a quarter of them implement a formal mechanism to receive patients’ feedback. Among centers with a feedback mechanism, nearly 80 percent of centers have a survey or claim box available. Health centers were asked to indicate the three main complaints received from patients in 2017. Nationally, the three main issues that stand out are long distance (39 percent), long waiting times (25 percent) and medicine unavailability (16 percent). There is a difference between urban and rural areas regarding these patient complaints. In rural areas, the complaint about long distance is more frequently mentioned (44 percent). Consistently, it is a complaint mentioned in 40 percent of type-C centers, which are more rural. As could have been expected, long distance is a complaint reported in the large majority (75 percent) of centers in the region of Bolama/Bijagos. Unavailable medicines come as the second complaint (20 percent) in rural facilities, before long waiting times (16 percent). This reflects the fact that rural centers are less easily accessible. However, in urban areas, long waiting time is a more common complaint (40 percent). Urban areas being more populated, the centers in these areas tend to be more crowded resulting in longer waiting time as opposed to rural centers. This is coherent with the fact that long waiting times was reported as a complaint in 80 percent of regional hospitals, which are all urban. Long distance is the second complaint in urban areas (cited in 32 percent of centers) but is less common than in rural areas. However, unavailability of medicines is a complaint only cited in 10 percent of urban facilities. Complaints about health staff and the received care are not common – but this information was given by the facility management staff and not directly by patients. 69 Table 31 Complaints by rural/urban Rural Urban National N 82 50 132 Formal mechanism for patients’ feedback (% of facilities) 32 42 36 Survey or claim box for feedback 73 86 79 3 Main complaints* Consultation fees 6 10 8 Lab exam fees 1 8 4 Medicine fees 10 16 12 Medicines unavailable 26 20 23 Health workers unavailable 26 12 20 Long waiting times 27 48 35 Service hours 12 20 8 Unrespectful minor staff 12 6 3 Unrespectful health staff 1 0 1 Inadequate or wrong care 24 0 2 Long distance 4 0 2 Informal fees 1 2 2 Lack of equipment 4 2 3 Other 16 12 14 *Percentage of health centers in which the complaint was listed among the three most frequent claims. Centers in which informal fees were named among the three most frequent complaints were: Mansaba (type C, Oio) and Reference Hospital Simão Mendes (SAB). Long distance, informal fees and lack of equipment were not originally mentioned as categories in the questionnaire and relied on enumerators to make a comment in the questionnaire if such a response was mentioned. 70 Figure 21: Percentage of health centers in which the complaint was listed among the three most frequent claims National SAB Cacheu Gabu Human Resources Bafata Quality of Service Bolama-Bijagos Equipment & Supplies Biombo Fees Oio Quinara Tombali 0 10 20 30 40 50 60 70 80 45 percent of centers declared having made alterations in response to patients’ complaints. The alterations types that are the most reported are: shorter waiting times (60 percent), longer opening hours (25 percent) and purchase of medicine (23 percent). The first and third main alterations match the second and third most common complaints. Urban centers reported to have made efforts to shorten waiting times, as this is an issue more frequent in urban areas. However, no change regarding long distance was mentioned. This is not surprising as this would have involved building a new facility closer to patients and is difficult to achieve. It is important to note that these alterations were directly reported by facilities. Nevertheless, this shows the awareness of the facilities in relation to improvement needs. Table 32 Response to feedback, by urban/rural Rural Urban National N 82 50 132 Any alterations in response to feedback (% of facilities) 40 54 45 Any alteration in response to feedback if the center has a formal 73 71 72 feedback mechanism Any alteration in response to feedback if the center has a survey 84 72 78 or claim box 3 Main alterations* Fees 7 14 10 Equipment & Supplies 11 14 12 Quality of Service 24 42 31 Human Resources 7 4 6 71 Transparency of health facilities was also investigated. Only ten percent of centers share financial information with the community. Rural centers (13 percent) tend to share financial information more than urban centers (4 percent) do. When financial information is shared, it is mainly through meetings (92 percent). The region where financial information is communicated the most to the community is Bolama/Bijagos. It is interesting to note that facilities of this region do not receive any financial support from the Ministry of Health or from the PIMI/IMVF program. In addition, reference and regional hospitals, as well as maternal-child centers do not share financial information. A higher number of facilities shares information related to medicine and equipment reception (35 percent). This time again, rural centers (41 percent) share information to a greater extent than urban centers (24 percent). Here again, the region of Bolama/Bijagos is the most transparent, as 56 percent of centers declare sharing information on medicine and equipment. Reference and maternal-child facilities also do not share information of this type. However, in this case, 40 percent of regional hospitals share information. Once again, the most common mean to share this type of information nationally is through meetings (56 percent), whereas public display is never used. Community health agents serve as communication channels in 30 percent of centers that inform of medicine reception nationally. It is only in rural areas that community health agents are a medium of communication which shows their close link to the community in more remote areas. Sharing information about given consultations to the community is uncommon. It only occurs in seven percent of centers and is more common in rural areas (10 percent) than in urban areas (2 percent). Figure 22: Information sharing with community 45% 41% 40% 35% 35% 30% 24% 25% Urban 20% 16% Rural 15% 13% 11% 10% National 10% 7% 5% 2% 0% Financial Information Consultations Number Medicine & Equipment Acquisition 4. Conclusions & Recommendations This report provides unique evidence on critical aspects of the Guinea-Bissau’s health system which represent a powerful tool for designing policies to tackle health systems challenges in Guinea- Bissau. Although in-depth discussion and consultations with the Government of Guinea Bissau, the donor community and civil society in the country, are necessary to validate and further explore the findings of this report, an initial set of recommendations can be listed. These are: - Improve the public resource tracking system. There is an urgent need to strengthen regulation to eliminate informal payment, and to rationalize the purchase and distribution of medicines; 72 - Strengthen health workforce compensation and education policies. This includes defining clear career pathway for different health workers, revision of the remuneration policy to implement performance- based pay and introduce non-monetary incentives, and to support the nursing and medical schools and expand in-service training for health workers; - Improve coordination and provision of services by integrated frontline primary health care (PHC) teams composed primarily of paid community health workers, auxiliary nurses and clinical officers, trained midwives, with the support of graduate nurses and physicians; - Roll out a nationwide PFM training program to implement the new regulatory framework, mentioned above, at all central and decentralized health facilities and strengthen compliance with it; - Strengthen the oversight and accountability processes for collection and use of Internally Generated Funds (IGFs) and externally received donations, such as user fees and donors' financial and in-kind support, for improved health services provision, through increased transparency in processes and procedures and strengthened internal controls framework.; - Strengthen PFM and procurement capacity at both the MEF and MINSAP as well as the coordination of PFM procedures between the MEF and the MINSAP. 73 - Strengthen information systems for disease surveillance and rapid response to disease outbreaks. That includes for both human and animal health. 74