91325 A STUDY ON THE IMPLEMENTATION OF JAMPERSAL POLICY IN INDONESIA DISCUSSION PAPER SEPTEMBER 2014 Endang L. Achadi Anhari Achadi Eko Pambudi Puti Marzoeki A STUDY ON THE IMPLEMENTATION OF JAMPERSAL POLICY IN INDONESIA Endang L. Achadi, Anhari Achadi, Eko Pambudi, Puti Marzoeki September 2014 Health, Nutrition, and Population (HNP) Discussion Paper This series is produced by the Health, Nutrition, and Population (HNP) Global Practice of the World Bank Group. The papers in this series aim to provide a vehicle for publishing preliminary results on HNP topics to encourage discussion and debate. The findings, interpretations, and conclusions expressed in this paper are entirely those of the author(s) and should not be attributed in any manner to the World Bank Group, to its affiliated organizations, or to members of its Board of Executive Directors or the countries they represent. Citation and the use of material presented in this series should take into account this provisional character. For information regarding the HNP Discussion Paper Series, please contact the Editor, Martin Lutalo at mlutalo@worldbank.org or Erika Yanick at eyanick@worldbank.org. © 2014 The International Bank for Reconstruction and Development / The World Bank 1818 H Street, NW Washington, DC 20433 All rights reserved. ii Health, Nutrition, and Population (HNP) Discussion Paper A Study on the Implementation of Jampersal Policy in Indonesia a a b b Endang L Achadi, Anhari Achadi, Eko Pambudi, Puti Marzoeki a Center for Family Welfare, University of Indonesia b Health, Nutrition, and Population Global Practice, the World Bank Group, Jakarta Office, Indonesia Paper prepared by the World Bank Group and supported by funding from the government of Japan through the Japan–World Bank Partnership Program for Universal Health Coverage (P125669). Abstract: Indonesia launched Jampersal in 2011, a nationwide program to accelerate the reduction of maternal and newborn deaths. The program was financed by central government revenues and provided free and comprehensive maternal and neonatal care with an emphasis on promoting institutional deliveries. Jampersal providers were public and enlisted private facilities at the primary and secondary levels. In 2013, the World Bank and the Center for Family Welfare, University of Indonesia conducted a qualitative and quantitative study to assess the implementation and impact of the program in Garut District and Depok Municipality in West Java Province. The study found that Jampersal utilization was highest among women who were least educated, poor, and resided in rural areas. Utilization was also high among women with delivery complications. The study showed Jampersal only had an impact where institutional delivery coverage was still low such as in Garut District. In this district, women were 2.4 times more likely to have institutional deliveries after Jampersal. The finding suggests implementation of Jampersal policy may have to be adjusted according to the utilization pattern for efficiency and effectiveness. The government discontinued Jampersal with the launching of the National Health Insurance Program (JKN) on January 1, 2014. The study’s findings indicate the merit in reevaluating the policy to terminate the program, given that Jampersal helped increase institutional deliveries while voluntary participation in JKN remains low. . Keywords: Jampersal, comprehensive maternal and neonatal care, national health insurance Disclaimer: The findings, interpretations, and conclusions expressed in the paper are entirely those of the authors, and do not represent the views of the World Bank, its Executive Directors, or the countries they represent. Correspondence Details: Puti Marzoeki at pmarzoeki@worldbank.org. iii Table of Contents Table of Contents .......................................................................................................................................... iv ACKNOWLEDGMENTS ............................................................................................................................. vi PREFACE ..................................................................................................................................................... vii ACRONYMS AND GLOSSARY ................................................................................................................ viii LIST OF FIGURES ..........................................................................................................................................x LIST OF TABLES ...........................................................................................................................................xi EXECUTIVE SUMMARY ............................................................................................................................... 1 PART 1: BACKGROUND .............................................................................................................................. 4 PART 2: OBJECTIVES .................................................................................................................................... 7 PART 3: JAMPERSAL PROGRAM DESCRIPTION ..................................................................................... 8 3.1. Program Target......................................................................................................................... 8 3.2. Benefit Package......................................................................................................................... 8 3.3. Financial Scheme .................................................................................................................... 10 3.4. Socialization ............................................................................................................................ 11 PART 4: STUDY FRAMEWORK ................................................................................................................. 12 PART 5: METHODOLOGY ......................................................................................................................... 14 5.1. Study Design .......................................................................................................................... 14 5.2. Population and Sample Size .................................................................................................. 14 5.3. Study Limitations ................................................................................................................... 14 PART 6: RESULTS ........................................................................................................................................ 15 6.1. DESCRIPTION OF THE STUDY AREAS ............................................................................. 15 6.2. Implementation of Jampersal Policy (Findings from the Qualitative Study) ..................... 16 6.2.1. Organizational Arrangement .............................................................................. 16 6.2.2. Financing .............................................................................................................. 17 6.2.3. Verification, Claim, and Reimbursement System.............................................. 18 6. 2. 3. 1. Primary Care Level ......................................................................... 18 6. 2. 3. 2. Hospital Level ................................................................................. 20 6.2.4. Provider Payment Scheme .................................................................................. 21 6. 2. 4. 1. Primary Care ................................................................................... 21 6. 2. 4. 2. Hospital Level ................................................................................. 21 6.2.5. Socialization ......................................................................................................... 22 6.2.6. Monitoring and Evaluation................................................................................. 23 6.2.7. Supply-Side Capacity .......................................................................................... 23 6. 2. 7. 1. Human Resources ........................................................................... 23 6. 2. 7. 2. Public Health Facilities ................................................................... 23 6. 2. 7. 3. Private Providers............................................................................. 24 6.2.8. Commitment to Jampersal .................................................................................. 25 6.2.9. Garut’s Dilemma in Implementing Jampersal Policy ....................................... 26 6.3. Results of the Household Survey .......................................................................................... 27 6. 3. 1 Respondent Characteristics................................................................................. 27 6. 3. 2 Community Understanding about Jampersal Program .................................... 28 6. 3. 3 Jampersal Utilization ........................................................................................... 30 6. 3. 4 Utilization of Maternal Health Services ............................................................. 31 1.3.4.1. Antenatal Care................................................................................. 31 1.3.4.2. Delivery ........................................................................................... 32 1.3.4.3. Cesarean Section ............................................................................. 34 1.3.4.4. Postnatal Care.................................................................................. 34 1.3.4.5. Family Planning .............................................................................. 35 1.3.4.6. Source of Payment for Last Delivery ............................................. 36 6. 3. 5 Impact of Jampersal............................................................................................. 37 6.3.5.1. Impact of Jampersal on Institutional Deliveries ............................ 37 6.3.5.2. Impact of Jampersal on Cesarian Section ...................................... 39 iv PART 7: DISCUSSION ................................................................................................................................. 42 7.1. Breadth of the Coverage: Who Is Insured? .......................................................................... 43 7.2. Depth of the Coverage: the Benefit Package......................................................................... 44 7.3. The Height of the Coverage: Financial Protection ............................................................... 45 PART 8: CONCLUSION .............................................................................................................................. 46 8.1. Policy Formulation and Implementation.............................................................................. 46 8.2. Health Service Provision and Performance .......................................................................... 46 8.3. Community Acceptance to Jampersal Policy ....................................................................... 47 8.4. Impact of Jampersal on MCH Service Coverage .................................................................. 47 PART 9: RECOMMENDATIONS................................................................................................................ 48 REFERENCES ............................................................................................................................................... 50 ANNEX ......................................................................................................................................................... 53 v ACKNOWLEDGMENTS This report is published by the World Bank Office, Jakarta, as part of the Japan–World Bank Partnership Program for Universal Health Coverage. Data management and analysis were conducted by Anhari Achadi, Endang L. Achadi, Trisari Anggodowati, Kamaluddin Latief, Fitri Nandiaty, Poppy Elvira, and Nurfitri Rachmadaniawati from the Center for Family Welfare, University of Indonesia (CFW-UI), and by Eko Pambudi from the World Bank. Report writing was led by Puti Marzoeki from the World Bank. The team coordinated closely with, and benefited from, valuable inputs during consultations with the following representatives of the Indonesian government: Usman Sumantri (P2JK/ Center for Health Financing and Health Insurance, MoH) and Riskiyana Sukandi Putra (Dit Kesehatan Ibu/ Directorate of Maternal Health, MoH). Valuable comments on the draft report were received from Ajay Tandon, Senior Economist, World Bank. The report was peer reviewed by Marjorie Koblinsky, International Maternal Health Expert, and Karima Saleh, Senior Economist, World Bank The paper was written under the overall guidance of Toomas Palu, Practice Manager, GHNDR. The authors are grateful to the World Bank for publishing this report as an HNP Discussion Paper. vi PREFACE In 2011, Japan celebrated the 50th anniversary of achieving universal health coverage (UHC). To mark the occasion, the government of Japan and the World Bank conceived the idea of undertaking a multicountry study to respond to this growing demand by sharing rich and varied country experiences from countries at different stages of adopting and implementing strategies for UHC, including Japan itself. This led to the formation of a joint Japan–World Bank research team under the Japan–World Bank Partnership Program for Universal Health Coverage. The program was set up as a two-year multicountry study to help fill the gap in knowledge about the policy decisions and implementation processes that countries undertake when they adopt the UHC goals. The program was funded through the generous support of the government of Japan. This country report on Indonesia is one of the 11 country studies on UHC that was commissioned under the program. The other participating countries are Bangladesh, Brazil, Ethiopia, France, Ghana, Japan, Peru, Thailand, Turkey, and Vietnam. A synthesis of these country reports is in the publication “Universal Health Coverage for Inclusive and Sustainable Development: A Synthesis of 11 Country Case Studies,” available at http://www.worldbank.org/en/topic/health/brief/uhc-japan. These reports are intended to provide an overview of the country experiences and some key lessons that may be shared with other countries aspiring to adopt, achieve, and sustain UHC. The goals of UHC are to ensure that all people can access quality health services; to safeguard all people from public health risks; and to protect all people from impoverishment due to illness, whether from out-of- pocket payments or loss of income when a household member falls sick. Although the path to UHC is specific to each country, it is hoped that countries can benefit from the experiences of others in learning about different approaches and avoiding potential risks. vii ACRONYMS AND GLOSSARY ANC Antenatal Care APBD Anggaran Pendapatan dan Belanja Daerah/ Regional Government Budget APBN Anggaran Pendapatan dan Belanja Negara/ State Budget APN Asuhan Persalinan Normal/Normal delivery care ARSADA Asosiasi Rumah Sakit Daerah Seluruh Indonesia/Indonesian District Hospital Association Askeb Asuhan Kebidanan/ Maternity Care Plan Askeskin Asuransi Kesehatan Miskin/Insurance for the poor Bappenas Badan Perencanaan dan Pembangunan Nasional/National Development Planning Agency BOR Bed occupancy rate BPK Badan Pemeriksa Keuangan/ National Finance Audit Board BPS Badan Pusat Statistik/ Central Bureau for Statistics Bumil Ibu Hamil/Pregnant woman Bupati Kepala Daerah/Head of district DTP Dengan Tempat Perawatan/ Primary health center with in-patient care Dinkes Kab Dinas Kesehatan Kabupaten/District health office Dinkes Prop Dinas Kesehatan Propinsi/Provincial health office EMAS Expanding maternal and newborn survival FGD Diskusi Kelompok Terarah/Focus group discusion GDON Gawat Darurat Obstetri Neonatal/Emergency Obstetric and Neonatal HDI Indeks Pembangunan Manusia/Human Development Index HRH Sumber Daya Kesehatan/Human resourches for health IBI Ikatan Bidan Indonesia/Indonesia Midwives Association ICU Intensive care unit IGD Instalasi Gawat Darurat/Emergency room INA-CBGs Indonesia case-based groups IUD Intrauterine device Jamkesda Jaminan Kesehatan Daerah/Local health insurance scheme Jamkesmas Jaminan Kesehatan Masyarakat/National health insurance scheme Jampersal Jaminan Persalinan/Universal delivery care Jamsostek Jaminan Sosial Tenaga Kerja/Social insurance for private sector workers JKN Jaminan Kesehatan Nasional/National Health Insurance Program Juknis Petunjuk Teknis/Technical guidelines KB Keluarga Berencana/Family planning Kemenkes Kementerian Kesehatan/Ministry of Health Kemendagri Kementerian Dalam Negri/Ministry of Home Affairs Kemenkeu Kementerian Keuangan/Ministry of Finance KK Kartu Keluarga/Family card KPPN Kantor Pusat Perbendaharaan Negara/State Treasury Office KTP Kartu Tanda Penduduk/Identity card K1 Kunjungan ANC Pertama/ANC in first trimester K4 Kunjungan ANC keempat/2nd ANC in third trimester Linakes Persalinan oleh tenaga kesehatan/Delivery assisted by health professional Linfaskes Persalinan di fasilitas kesehatan/Institutional delivery MoU/SKK Surat Keterangan Kerjasama/Memorandum of understanding MDGs Millennium Development Goals Menkokesra Menteri Koordiantor Kesejahteraan Rakyat/Ministry of People’s Welfare NICU Neonatal intensive care unit Ob-gyn Spesialis Kebidanan/Obstetrics and gynecology Perbup Peraturan Bupati/Head of district regulation Perda Peraturan daerah/Local regulation Permenkes Peraturan Menteri Kesehatan/ MoH regulation PERSI Perhimpunan Rumah Sakit Seluruh Indonesia/Indonesian Hospital Association Perwali Peraturan Walikota/Mayor regulation PKH Program Keluarga Harapan/Conditional cash transfer program PNC Postnatal care viii PNBP Penerimaan Negara Bukan Pajak/Nontax payment PNPM-GSC Program Nasional Pemberdayaan Masyarakat — Generasi Sehat dan Cerdas/The National Program for Community Empowerment (PNPM). Smart and Health Generation PNS Pegawai Negeri Sipil/Civil servant PPH 21 Pajak Penghasilan/Personal income tax PPM Bidan Praktek Swasta/Private practice midwife Polindes Pos Bersalin Desa/Village delivery post PONED/BEONC Pelayanan Obstetrik dan Neonatal Esensial Dasar/Basic essential obstetric and neonatal care (BEONC) PONEK/CEONC Pelayanan Obstetrik dan Neonatal Esensial Komprehensif/Comprehensive essential obstetric and neonatal care (CEONC) Posyandu Pos Pelayanan Terpadu/Integrated service post PP Peraturan Pemerintah/Government regulation PPS Probability Proportional to Size PPK1 Pemberi Pelayanan Kesehatan — Tingkat 1/Primary health service PTT Pegawai Tidak Tetap/Contracted worker (temporary civil service) P2JK Pusat Pembiayaan dan Jaminan Kesehatan/Center for Health Financing and Health insurance Puskesmas Pusat Kesehatan Masyarakat/Primary health center Pustu Puskesmas Pembantu/Satellite puskesmas RB Rumah Bersalin/Maternity clinic Riskesdas Riset Kesehatan Dasar/National Basic Health Research RKA Rencana Kerja Anggaran/Budget work plan SBA Persalinan oleh tenaga terlatih/ Skilled birth attendance SC Sectio Cesarea (Operasi Sesar)/Cesarian section SIP Surat Izin Praktek/License to practice SK Surat Keputusan/Decree SP2D Surat Perintah Pencairan Dana/Disbursement letter TBA Dukun Bersalin/Traditional birth attendants UNFPA United Nations Population Fund Walikota Kepala Daerah/Mayor WHO Organisasi Kesehatan Dunia/World Health Organization ix LIST OF FIGURES Figure 1.1 Maternal Mortality Ratio in Indonesia................................................................................ 4 Figure 1.2 Jampersal Fund Channeling............................................................................................ 10 Figure 1.3 Conceptual Framework of Policy Implementation ............................................................ 12 Figure 1.4 Study Framework ........................................................................................................... 13 Figure 1.5 Claim and Reimbursment Mechanism at Primary Care Level .......................................... 19 Figure 1.6 Claim and Reimbursment Mechanism at the Hospital Level............................................. 21 Figure 1.7 Proportion of Women Who Have Heard of Jampersal ...................................................... 28 Figure 1.8 Source of Jampersal Information in Depok and Garut ..................................................... 28 Figure 1.9 Perception about Jampersal ........................................................................................... 29 Figure 1.10 ANC 1-1-2 Visit Pattern ................................................................................................ 31 Figure 1.11 Place of Delivery .......................................................................................................... 32 Figure 1.12 Skilled Birth Attendance before and after Jampersal ..................................................... 33 Figure 1.13 Percentage of Postnatal Care....................................................................................... 35 Figure 1.14 Percentage of Family Planning, before and after Jampersal, by Method ....................... 35 Figure 1.15 Three Dimentsions of Universal Health Coverage......................................................... 42 x LIST OF TABLES Table 1.1 Comparison between Jampersal and Other Social Insurance Programs in Indonesia prior to January 2014................................................................................................................................. 5 Table 1.2 Comparison of 2011 and 2012 Jampersal Guidelines ......................................................... 9 Table 1.3 Facilities and Human Resources for Health in Garut and Depok, 2011 ............................ 15 Table 1.4 The Number of Health Facilities in Garut and Depok, 2010–2012 ..................................... 16 Table 1.5 Population and Recipients of Social Health Insurance in 2012 .......................................... 16 Table 1.6 Roles and Responsibilities of Actors under Jampersal ...................................................... 17 Table 1.7 Budget Allocation and Utilization of Jampersal in Garut, 2011–13 ..................................... 17 Table 1.8 Budget Allocation of Jampersal in Depok, 2011–13 .......................................................... 18 Table 1.9 Respondent Characteristics ............................................................................................. 27 Table 1.10 Utilization of Jampersal for the Last Delivery, by Respondents’ Characteristics, Depok and Garut, Post-Jampersal (after 2011) ................................................................................................. 30 Table 1.11 Number of ANC Visits .................................................................................................... 31 Table 1.12 Place of Delivery by Time and by Women’s Characteristics, ........................................... 32 Table 1.13 Breakdown of Preferred Birth Delivery Facilities ............................................................ 33 Table 1.14 Skilled Birth Attendance before and after Jampersal, by Urban and Rural Areas, ............ 34 Table 1.15 Results of Bivariate Analysis on Association between Jampersal and Cesarian Sections 34 Table 1.16 Source of Payment for the Last Delivery ......................................................................... 36 Table 1.17 Type of Other Payments during Last Delivery, among Jampersal Users ......................... 37 Table 1.18 Association between Jampersal and Institutional Deliveries ............................................ 38 Table 1.19 Change in Institutional Deliveries by Mother’s Education, Socioeconomic Status, Insurance Ownership, and Residence before and after Jampersal in Garut and Depok (%) .............. 38 Table 1.20 Change in Cesarian Sections by Women’s Characteristics before and after Jampersal in Garut and Depok (%)....................................................................................................................... 39 Table 1.21 Association between Jampersal and Cesarian Sections ................................................. 40 Table 1.22 Results of Multivariate Analysis on Association between Wealth Quintiles and Cesarian Section ............................................................................................................................................ 41 xi EXECUTIVE SUMMARY Indonesia has made progress in reducing maternal mortality in recent years. Despite this progress, Indonesia’s maternal mortality ratio (MMR) level remains high relative to its income, to regional peers, and for a country that has high utilization of maternal health services such as antenatal care (ANC) and skilled birth attendance rates. In 2011, the government launched the Jampersal program to accelerate the reduction of maternal and newborn deaths. The program was financed by central government revenues, and provided free and comprehensive maternal and neonatal care with an emphasis on promoting institutional deliveries. The beneficiaries were those without coverage from existing social health insurance schemes such as Askes, Jamsostek, and Jamkesmas/Jamkesda. Jampersal providers were public facilities and enlisted private facilities at the primary and secondary levels. The government reimbursed providers for Jampersal services through fee claims. The tariff for primary level services was set by the Ministry of Health (MoH), while the tariff for hospital services followed the INA-CBG (Indonesia case-based groups). The World Bank in collaboration with the Center for Family Welfare, University of Indonesia (CFW-UI) conducted a study in 2013 to assess the implementation of Jampersal and the impact of the program on the coverage of maternal and neonatal health services. At the subnational level, the study involved Garut District and Depok Municipality in West Java Province. The study comprised qualitative and quantitative components and was funded through the Japan–World Bank Partnership Program for Universal Health Coverage. Jampersal was implemented by the MoH and local governments, although other stakeholders such as the Ministry of Finance, Bappenas (National Development Planning Agency,) Office of the Vice President, Ministry of Home Affairs, and the coordinating Ministry of People’s Welfare were involved in the policy formulation process. Awareness and socialization of Jampersal were lower than expected: even in the third year of implementation: 30 percent of respondents (women of child-bearing age) in both Garut and Depok were still unaware about Jampersal, in part due to limited involvement of other sectors during implementation. Even among those respondents who knew about Jampersal, most perceived Jampersal was a program that was limited to provision of free deliveries at puskesmas and public hospitals. Respondents doubted that the services really were for free, and some perceived that “free” meant lower service quality. The percent of deliveries (last birth for each woman) financed by Jampersal in Garut and Depok were 28.1 percent and 10.3 percent, respectively. Jampersal utilization was highest among women who were least educated, poor, and residents of rural areas. Jampersal use was also high among women with delivery complications. Contrary to design, Jampersal use was higher among those who already had health insurance coverage, including Jamkesmas/Jamkesda beneficiaries. Respondents who knew about the program considered the requirements for using Jampersal services simple (identity card and MCH [maternal and child health] book) and were aware that, unlike Jamkesmas, the choice of providers at the primary level was not limited to public providers. Almost 60 percent of Jampersal users in both Garut and Depok reported paying additional out-of-pocket (OOP) costs for delivery services at health facilities, and about half of them could not explain the reason for the extra OOP payments. Among those that provided explanations for OOP costs, 16 percent reported paying for drugs and injections, even though these should have been covered by the program. In addition, respondents noted that private midwives sometimes requested extra payments for long-term contraceptives, as Jampersal fee reimbursement for family planning services was lower than the cost of long-term contraceptives. Households also reported paying additional OOP costs for referral transport. Puskesmas was the main Jampersal provider at the primary level in Garut. In Depok, the private sector was the dominant maternal and neonatal care provider. However, although 67 percent of private practice midwives and all private hospitals have enlisted as Jampersal providers, their actual participation in the program was minimal. The reluctance for involvement in Jampersal was mostly due to dissatisfaction about service fees. Although the government increased the fees in the second year, the amount was still considered much lower than the regular private sector fees. Moreover, 1 Jampersal reimbursement processes were considered cumbersome because verification requirements and claim payments were often delayed. Private providers preferred non-Jampersal clients who paid OOP directly. The qualitative study revealed that Jampersal resulted in a reported workload increase among public sector providers and that this affected the income of midwives, particularly those engaging in dual practice. Jampersal negatively influenced those dual practice midwives who had a high number of patients in their private practices before Jampersal. Conversely, private practice midwives with fewer patients before Jampersal reported benefitting positively from the program. In Garut District, the institutional delivery increase after Jampersal was 54.4 percent, and the increase was statistically significant; women were 2.4 times more likely to have institutional deliveries after Jampersal. However, around 30 percent of deliveries in Garut were still assisted by traditional birth attendants (TBAs). Coverage of institutional deliveries in Depok was already high (92.3 percent) prior to Jampersal, and after almost three years of program implementation, the coverage remained the same. An interesting phenomenon was the slight shift of institutional deliveries in Depok from the private to the public sector, and from private midwives to private clinics/obstetricians after Jampersal. There was almost no change in antenatal care (ANC) and postnatal care (PNC) visits after Jampersal, although the use of long-term contraceptives has increased. Delivery by Cesarean section (C-section) increased by 26.4 percent, but the increase was not statistically significant, and the study could not convincingly confirm that Cesarean sections were strictly for delivery complications. The odds of having a C-section were higher in Depok than in Garut, suggesting higher access to C-section services in Depok. Recommendations A program like Jampersal requires strong support from stakeholders beyond the health sector. Jampersal encompassed horizontal (across sectors within the same level) and vertical (between the central and local government) collaboration. Jampersal has brought sectors together during policy formulation, but sector collaboration is important for building the momentum during implementation as well. For example: a. Given the decentralized system in Indonesia, the involvement of the Ministry of Home Affairs (MoHA) could help build local government commitment to the program. A strong buy-in from local governments may increase subdistrict and village government support in increasing community awareness and potentially in removing other barriers to care. b. Joint collaboration between MoH, MoHA, and Ministry of Finance (MoF) in reviewing existing regulations would be helpful in finding ways to reduce the complexity of Jampersal (or other insurance) fund management and the reimbursement process. c. Multisector involvement is required in planning and implementing long-term investment for improving road infrastructure, transportation, and health facilities in geographically difficult areas, to improve access to institutional deliveries. Jampersal could help to increase institutional deliveries while National Health Insurance Program (Jaminan Kesehatan Nasional, JKN) voluntary participation is still low. The study has shown the potential of Jampersal to increase institutional deliveries where coverage was low, as in Garut. The government may want to consider this finding and reevaluate the policy to terminate Jampersal implementation with the launching of the JKN on January 1, 2014. Nevertheless, before continuing the implementation, it is important to review the cost-effectiveness of the program. Addressing “nonservice” cost is important. The study reported families still pay additional OOP cost for referral transport, which might be a barrier to accessing care. Building linkages with other 2 programs such as PNPM GSC1 might be an option for addressing this issue. Moreover, reducing OOP would also require better understanding about the reason for the OOP; for example, whether it is an issue of supply chain and drug shortages, an issue of exclusion of certain drugs from the basic benefit package, or an issue of provider prescription behavior. For the latter, it will be useful to monitor possible cream skimming.2 “Free” is not enough; service readiness and quality also matter. Women who experienced low quality of care tended to stop using the care, or even dissuaded other women from using care (Wairimu 2013). In Kenya, women’s refusal to use free maternity care was due to poor quality of facilities and the rude attitude of health providers, among other factors. Continuous quality improvements should follow any effort to improve access to health care. Complementary work shows that maternal health supply-side readiness problems remain, especially in some parts of the country; these issues could deter patients from utilizing care, despite the removal of financial barriers via programs such as Jampersal. Getting private providers on board requires a carefully designed provider payment system. In areas where private providers are dominant and demand for private provision is high, buying services from the private sector to improve access might be more efficient than expanding public sector investment. Obviously, this would require reasonable fees and quality assurance of service provision. The study also showed that the support from professional associations like the Indonesia Midwives Association (Ikatan Bidan Indonesia, IBI) could facilitate private midwife participation in Jampersal. There is a need to explore more options to improve human resources for health (HRH) availability and distribution in remote areas. Skilled birth attendance in rural Garut increased after Jampersal despite the low presence of midwives, suggesting adding more midwives could increase skilled birth attendance even more if paying for service is not an issue. Experience shows that monetary incentives are often not enough for deploying HRH to “difficult” areas. Other attractive features, such as housing, children’s education, continuing education, and a clearly defined time period of service need to be considered. In developing HRH policies, Indonesia would benefit from a labor market analysis to understand labor market dynamics influencing HRH supply and demand. In areas with difficult access to hospitals, the presence of basic emergency obstetric and neonatal care (BEONC) is essential. Investment in BEONC should be followed by close monitoring and consistent support to ensure continuity of care as well as by introducing policies that deter high turnover of trained staff and increase utilization. Implementation of Jampersal (or UHC) policy may have to be adjusted according to the utilization pattern for efficiency and effectiveness. The household survey conducted under this study showed Jampersal only had an impact in an area where institutional delivery coverage was still low, such as Garut. Further assessment is needed to translate this finding into future policy changes. 1. PNPM GNC is the government’s community-driven development program, providing block grants to poor communities to attain selected health and education targets. 2. Midwives working in dual practice may decide to ask patients to come to their private practice as they can use the whole claim reimbursement for their own benefit, while in puskesmas they have to share with other staff. 3 PART 1: BACKGROUND Indonesia has made progress in reducing maternal mortality. Although there is uncertainty about Indonesia’s MMR level, all existing estimates show a decline of MMR during the last two decades with a recent plateau or even increase (figure 1.1). Despite the decline, Indonesia MMR level is still high for a country with good access to maternal health services. Figure 1.1 Maternal Mortality Ratio in Indonesia Maternal mortality ratio in Indonesia, 1990-2015 Maternal deaths per 100,000 live births 700 900 Joint WB-UN 2010 Census 500 DHS 300 IHME MDG target 100 1990 1995 2000 2005 2010 2015 Year Sources: Joint WB-UN estimates; Indonesia census DHS; Institute of Health Metrics and Evaluation (IHME) Note: Shaded area represents joint WB-UN estimation uncertainty By 2012, delivery by a skilled provider was 83.1 percent, although 36 percent of deliveries were at home, and 13.5 percent were assisted by traditional birth attendants (BPS et al. 2013). Many studies reported that possible reasons for maternal deaths included unskilled attendant during delivery, lack of knowledge of danger signs, and delayed referral. In addition, horizontal referrals have contributed to the delay in proper management of birth delivery complications. To achieve MDGs 4 (reduce child mortality) and 5 (improve maternal health), the government has promoted institutional deliveries and improving the referral system, including access to referral services. In 2011, the government launched the Jampersal program to provide free delivery assistance to those who do not have insurance coverage for delivery services. Under this definition, Jampersal is an expansion of Jamkesmas coverage for delivery services and aims at achieving universal coverage for maternal and neonatal health services. Table 1.1 show the various social insurance schemes implemented at the same time as Jampersal, as well as prior to the merger of those programs under the National Health Insurance Program (JKN) on January 1, 2014. 4 Table 1.1 Comparison between Jampersal and Other Social Insurance Programs in Indonesia prior to January 2014 Jampersal Jamkesmas3 (established Askes (established Jamsostek (established (established 2011)* 2005)** 1960)*** 1992)*** Groups All women who were Poor and the near-poor Civil servants, Private employers with mandated not covered by any retired civil >10 employees or pay type of health servants, retired salary >Rp 1 million a insurance military personnel, month and veterans Number n.a. 76.4 million 16.6 million 5.0 million enrolled Premium n.a. Rp 6,500 ($0.67) per capita 2% of basic + 1% 3% of salary for per month government; no bachelors; ceiling 6% of salary for married employees; ceiling Rp 1 million per month (not changed since 1993) Contributor Government 100% Government 100% Employees 66%; Employers 100% employer 34% Carrier Ministry of Health Ministry of Health PT Askes (for profit) PT Jamsostek (for profit) Benefits Maternal and Comprehensive; drugs are Comprehensive, no Comprehensive; cancer newborn health care covered if prescribed within specific exclusion; treatment, cardiac formulary; no cost-sharing drugs are covered if surgery, hemodialysis, prescribed within and congenital diseases formulary; are excluded; 4 drugs Cost-sharing are covered if available when prescribed within services fall outside formulary; basic benefit no cost-sharing package Dependents n.a. All family members Spouse + 2 children Spouse + 3 children under 21 years who under 21 years who are are not working and not working and not not married married Providers All puskemas and All puskemas and public Mostly contracted Mixed: public and public hospitals and hospitals and selected public health centers private providers selected empanelled empanelled private and public hospitals private clinics and hospitals private hospitals Provider Fee-for-service at Fee-for-service at Special fee schedules Fees are negotiated; payment puskesmas; diagnosis- puskesmas; DRG for for civil servants; extra billing mechanisms related group hospitals extra billing depending on (DRG)/case-based depending on negotiated fees group (CBG) for negotiated fees hospitals Source: *Indonesia, Ministry of Health 2011; **Indonesia, Ministry of Health 2012; *** Harimurti et al. 2013. 3. Health insurance for the poor was introduced as Askeskin in 2005, which was expanded and renamed Jamkesmas in 2007. 4. Starting in 2012, Jamsostek expanded the benefits package to cover catastrophic cases as well. 5 Jampersal benefit package includes antenatal, delivery, and postnatal services at the primary care facilities, and referral services for maternal and neonatal complications at secondary and tertiary hospitals. Table 1.2 in the next section explains the details of the benefit package. The primary care facilities are public health centers (puskesmas) and their network, including the polindes (village delivery post). Enlisted private providers include private midwives, private midwifery clinics, and private hospitals. Referral/in-patient care is provided at class-3 hospital beds at public and enlisted private hospitals. 6 PART 2: OBJECTIVES This study examined the financing, payment, and organization policies of the Jampersal program. Specific objectives of the study include the following: (i) a review of how Jampersal policy was formulated at the central level and implemented at the local level; (ii) a review of service provider and community response to the policy; and (iii) an assessment of the impact of Jampersal on the coverage of maternal services, particularly the impact on institutional deliveries at the primary level and Cesarean section (C-section) coverage at the secondary level (See annex 3 for study methodology). 7 PART 3: JAMPERSAL PROGRAM DESCRIPTION 3.1. PROGRAM TARGET Jampersal is a government-supported program for maternity care, specifically targeting pregnant women who are not covered by any other health insurance scheme regardless of their socioeconomic status. The government of Indonesia is committed to improving the country’s health system as stated in the Ministry of Health Strategic Plan 2010–2014 and the government’s Roadmap to Accelerate the Achievement of the MDGs in Indonesia. The commitments translate into the provision of social health insurance for the poor, among others. Until the end of 2013, the main social health insurance for the poor in Indonesia was Jamkesmas, previously known as Askeskin, financed by the central government and targeting the poor and near-poor. Jamkesmas, which became operational in 2005, was managed by the MoH, and provided beneficiaries with free health services in puskesmas and hospitals. Another type of insurance for the poor is Jamkesda, which is funded by the subnational government (province/district level), and finances health care for the poor who are not covered by Jamkesmas. The aim of most social health insurance programs is to enable the poor and near-poor to gain access to health services to reduce mortality, morbidity and inequality. Jampersal provides comprehensive maternal health service coverage to those not covered by Jamkesmas, Jamkesda, or any other health insurance scheme. The government terminated the implementation of Jampersal in December 2013 with the launch of the National Health Insurance Program (JKN) on January 1, 2014. 3.2. BENEFIT PACKAGE Jampersal covers pregnancy, delivery, and postpartum services including antenatal care, delivery care, postpartum care for mother and newborn, and family planning. The coverage includes standard drugs used in the national MCH program. Referral care is provided in hospital outpatient clinics and class 3 hospital beds, regardless of the income level of the user. The Ministry of Health released the Jampersal guidelines under Ministry of Health Regulation no.631/MENKES/PER/III/2011. This regulation was revised by Ministry of Health Regulation no. 2562/MENKES/PER/XII/2011 in 2012, to adjust the benefit package and the unit costs in response to findings from the evaluation of the first year implementation, indicating dissatisfaction among health providers with the service fees. Table 1.2 compares Jampersal 2011 with 2012 guidelines on services, frequency, and fees. 8 Table 1.2 Comparison of 2011 and 2012 Jampersal Guidelines Jampersal 2011 vs 2012 2011 2012 Type Benefit package (No.631/Menkes/PER/III/201 (No.2562/Menkes/PER/X11/2011) 1) Rate ANC in basic Antenatal care Rp 10,000/time Rp 20,000/time service Specifically for ANC with complications, the claim can be done according to the number of Frequency 4 times only ANC provided as long as it is consistent with the standards (ANC can be more than 4 times) Fee for normal Normal delivery in primary Rp 350,000/time Rp 500,000/time delivery level Limited to obstetric Obstetric and nonobstetric Maternal Maternal complication complications (related to complications (heart and other life- complication is covered pregnancy, delivery, and threatening pregnancy postpartum) complications) Rate for management of hemorrhage, vaginal delivery with basic Rp 500,000 Rp 650,000 emergency care in BEONC facilities In-patient services for complications during Covered, rate based on standard Not covered pregnancy, postpartum rate for inpatient services in BEONC facilities In-patient services for Covered, rate based on standard Not covered sick newborns rate for inpatient services Postpartum care, e.g., Separate at a rate of Rp 150,000 placenta removal Included in delivery package Fee for PNC in Postnatal care Rp 10,000/time Rp 20,000/time primary level 4 times to achieve the target pof KF1,KF2,KF3&KN1, KN 2, KN35 For PNC with complications, Frequency 3 times only according to the number of services given as long as consistent with standards (PNC can be more than 4 times) Separate: Contraception services6 Included in PNC component a. IUD/implant Rp 60,000 b. Injection Rp 10,000 Complication of contraception Not covered Rp 100,000/time postdelivery Transport Transport for referral Covered Covered management of sick Newborn Not covered Covered newborn Source: Indonesia, Ministry of Health 2011–12. 5. KF = Postpartum visit KN = Neonatal visit First visit for KF1 and KN1 (6 hours to 2nd day); second visit for KN2 (3rd to 7th days); third visit for KF2 and KN3 (8th to 28th days); fourth visit for KF3 (29th to 42nd days 6. Although Jampersal also reimburses the use of short-term contraceptives, the guidelines emphasized prioritizing long-term contraceptives. 9 3.3. FINANCIAL SCHEME Jampersal is entirely financed through central government revenues. Figure 1.2 describes channeling of the Jampersal fund from the central to the district level. The fund was channeled directly from the State Treasury Office (Kantor Pusat Perbendaharaan Negara, KPPN Jakarta V) to the following: The head of district/municipality health office (DHO) account. A Jamkesmas management team at the DHO manages the fund for puskesmas and participating private health facilities/providers at the primary level. The center makes the transfer periodically, three to four times a year. The amount is based on the number of pregnant women projection for the district, and adjusted according to fund utilization during the previous reporting period. The hospital account for hospitals, which has a memorandum of understanding (MoU) with the DHO for Jamkesmas/Jampersal program. The center transfers the amount in advance based on service utilization in the hospital budget report. As a measure to maintain continuity of care and equity among regions, P2JK (Pusat Pembiayaan dan Jaminan Kesehatan, the Center for Health Financing and Health Insurance, MoH) has the authority to reallocate funds among districts and municipalities according to district utilization and need, and depending on the availability of funds at the national level. Figure 1.2 Jampersal Fund Channeing Feedbac Ministry of Health (P2JK) Provincial KPPN V Health Office (State Treasury Financial Report SP2 Financia with INA-CBG’s l Report Transfer Fund Transfer Fund (Jamkesmas/Jamper (Jamkesmas/Jamper sal) sal) Hospital District Health CLAIM by Hospital V ifi t POA+CLAIM Claim Local Government Coders & ADM by Indenpendent Verificator PH PH Hom NOTE: C C : Transfer Fund Clini PH PP : Claim M C : Financial Report JAMPERS : Feedback Report Source: Modified from Ministry of Health Jampersal Technical Guidelines 2012. Payment for the puskesmas is done through a transfer from the district treasury, while primary level private health providers receive payment directly from the DHO. The amount of payment is based on claims. The DHO verifies claims from primary level providers, while P2JK assigns independent verificators to verify hospital claims. Jampersal adopts the portability principle, allowing beneficiaries to get services from any Jampersal participating facility; the facility can make claims to the DHO 10 where it is located. The DHO must return unused funds to KPPN V at the end of the year, and submit the fund utilization report to MoH. 3.4. SOCIALIZATION MoH invited DHOs and hospitals to the socialization workshops about Jampersal it held at each provincial health office (PHO); Jampersal technical guidelines were distributed to all attendees. 11 PART 4: STUDY FRAMEWORK Policy Implementation Theory lists five factors influencing the success of policy implementation. The factors are content, context, commitment, client capacity, and coalition or the 5-C Protocol (Najam 1995). According to Cheema and Rondinelli (1983), policy performance and impact in a decentralized setting is affected by four variables as illustrated in figure 1.3 below: Figure 1.3 Conceptual Framework of Policy Implementation Relation between organizations Environment Characteristics Performa condition & capability of nce & institutions impact Organizational resources Source: Cheema and Rondineli,1983. The study adopted the 5-C Protocol and the above conceptual framework to comprehensively analyze Jampersal implementation, focusing on the following: Jampersal policy design at the national level: (a) policy formulation, (b) soliciting political support to improve policy acceptance, (c) development of implementation guidelines, and (d) monitoring and evaluation. Jampersal implementation at the provincial and district levels: (a) policy implementation issues, including analysis of barrier and facilitating factors; and (b) impact of the policy on continuum of care and provider performance. 12 Figure 1.4 Study Framework Jampersal policy formulation at central level Involvement of related stakeholders in the design and formulation of the policy, in its legitimation, development of implementation guidelines, monitoring and evaluation Jampersal policy Implementation at Municipality/District level Contextual factors (Population, geographic, culture, revenue, Health Office’ capacity Commitment, awareness, advocacy t ) Local gov’t support & commitment (Counter budget, local regulation, etc) Health services provision & Community’s performance (hospital, health acceptance and Stakeholder support & center, doctors, midwife) use of services commitment (awareness, l t ) Continuum of care (by time and place) o % ANC o Delivery:  % facility-based delivery  % Cesarean section o % postnatal care o % postpartum family Source: Authors. Jampersal policy is developed centrally and implemented by local governments to provide maternal and newborn health services at public health service delivery points and private institutions enlisted by the local government. Because of decentralization, implementation varied among the districts, depending on local government commitment, health service, and community factors. This study examined the implementation of Jampersal policy in Garut District (mostly a rural area) and Depok Municipality (an urban area) in West Java Province, covering contextual factors (geography, population, revenue, and cultural aspects); local government commitment and support; stakeholder participation and support; and the capacity of the lead implementing unit (that is, district health office) in implementing the policy. This study also looked at service provision by the responsible public and private health service institutions at the primary and secondary levels; human resources, particularly doctors and midwives working in the institutions, regarding their satisfaction toward Jampersal terms and conditions and financial incentives; and Jampersal utilization by the community, including involvement of the community leaders, client satisfaction, and reasons for not using Jampersal. The study analyzed Jampersal program output at the community level, covering the level of antenatal care, institutional delivery, Cesarean section, and postpartum family planning services. In interpreting the impact of Jampersal, the study also identified other programs implemented at the national, provincial, or district level in health and other sectors that may influence program outcomes. 13 PART 5: METHODOLOGY 5.1. STUDY DESIGN The team implemented the study during May to August 2013 in Garut District and Depok Municipality in West Java Province. The study applied qualitative and quantitative approaches. The study design was cross-sectional with pre/post assessment. The qualitative approach provided information about the financing, payment, and organizational policies of Jampersal and reviewed the impact of the policies on HRH performance. The quantitative approach examined whether Jampersal has an impact in improving the coverage of services, particularly institutional deliveries and Cesarean sections. The sample size provided statistically representative data for each area. 5.2. POPULATION AND SAMPLE SIZE The qualitative approach used in-depth interviews and focus group discussions at the central, provincial, district/municipality, and community levels. A household survey measured the impact of Jampersal on maternity care utilization. The respondents were women of childbearing age (15 to 49 years). Women who delivered two years (the two-year cut-off point was used to reduce recall bias) before Jampersal implementation (2011) were the baseline/pre-Jampersal samples, and women who delivered after Jampersal implementation were the endline/post-Jampersal samples. The required minimum sample size was 453 households for each group (pre- and post-) in each study area. The sample size was 906 samples in total for each area, or 1,812 samples in the two study areas. The study collected information from 921 respondents in Depok and 918 respondents in Garut, or a total of 1,839 respondents. The households were randomly selected using a two-stage random sampling with Probability Proportional to Size (PPS) method. Inclusion criteria were mothers with a child/children who have lived in the study area at least since 2009 (See annex 3 for the full study methodology). 5.3. STUDY LIMITATIONS The study may have been undertaken too early to evaluate the real impact of Jampersal. Further, since it covered only two districts, the study cannot represent the real situation of Jampersal implementation in the country. 14 PART 6: RESULTS 6.1. DESCRIPTION OF THE STUDY AREAS The two study areas were widely different in factors influencing access to health care. The size of Garut District was approximately 3,000 kilometers,² with a population of over 2.4 million residing in 42 subdistricts. The majority of the district is rural/remote areas consisting of 403 rural-villages and only 21 urban-villages. In contrast, Depok Municipality is an urban area, with no geographical constraints 2 and with a high availability of health care facilities. Depok encompasses an area of 200.2 kilometers and has a population of approximately 1.8 million people. Depok is a rapidly growing city, located next to Jakarta, Indonesia’s capital. According to the 2011 Human Development Index (HDI), 7 Depok had a high HDI score (79.36), higher than the HDI of West Java Province (72.73). The HDI of Garut for the same year was 71.70. In 2012, 12 of the 65 puskesmas in Garut had no doctor. More than half of the midwives in two of the sampled puskesmas in Garut were temporary midwives, who were paid a monthly honorarium by the puskesmas. Around 52 percent of village midwives in Garut had completed Normal Delivery Care (Asuhan Persalinan Normal, APN) training while only around 29 percent of them had attended Obstetric and Neonatal Emergency Care training (Gawat Darurat Obstetrik Neonatal, GDON). In Depok, almost 91 percent of the midwives had completed APN training, and around 63 percent had attended GDON training. The list of facilities and human resources for health at the beginning of the Jampersal program (2011) in the two locations is shown in table 1.3 below. Table 1.3 Facilities and Human Resources for Health in Garut and Depok, 2011 Depok Garut District Facilities Municipality Number (*) Number (**) Public hospitals 2 1 Public hospital beds 504 71 Military hospitals 1 1 Private hospitals 1 14 Puskesmas (with in-patient care) 15 1 Puskesmas (without in-patient care) 50 31 Pustu (satellite puskesmas) 136 5 Private maternity clinics 7 25 Private physicians 2 158 Private clinics 89 19 Posyandu (integrated services post) 3,558 974 Human resources Specialist physicians 30 551 Obstetricians 5 35*** Anesthesia specialists 12 24 Physicians (general practitioners) 93 269 Midwives 586 377 Village midwives 431 — Nurses 673 1827 Pharmacists 70 284 Source: *Garut District Health Profile 2011; **Depok Municipality Health Profile 2011; ***based on license. 7. Human Development Index is a composite of three indexes: life expectancy rate, education (based on illiteracy rate and average school years), and decent living index. 15 In 2012, there were 15 BEONC facilities in Garut, or double the number of 2010. Garut DHO had converted 14 Puskesmas in geographically difficult areas into BEONCs. There were far fewer BEONC facilities in Depok: only one BEONC in 2010 and four in 2012 (table 1.4). Table 1.4 The Number of Health Facilities in Garut and Depok, 2010–12 Data Garut District* Depok Municipality** 2010 2011 2012 2010 2011 2012 Puskesmas (PHC) 64 65 65 32 32 32 BEONC8 facilities 6 15 15 1 2 4 Midwives 930 1,074 1,074 108 109 109 CEONC9 (public) facilities 1 1 1 1 1 1 CEONC (private) facilities 0 0 0 15 15 15 Source: *Garut District Health Profile 2010–12; **Depok Municipality Health Profile, 2010-12. Less than half of Garut’s population and around a quarter of Depok’s population were covered either by Jamkesmas, Jamkesda, 10 or Askes. 11 In Garut, Jamkesmas, Jamkesda, and Askes covered around 33 percent, 6 percent, and 2 percent of the population, respectively. In Depok, the coverage was around 8 percent, 10 percent, and 6 percent, respectively (table 1.5). Table 1.5 Population and Recipients of Social Health Insurance in 2012 Garut Depok Total population* 2,485,130 1,813,612 Poor population** 969,924 321,012 Estimate number of pregnant women* 67,414 47,899 Estimate number of birth deliveries 64,701 45,722 Jamkesmas** (%) 33 8 Jamkesda** (%) 6 10 Askes** (%) 2 6 Source: *Health Profile of Garut District and Depok Municipality, District Health Office 2010–12; ** Provincial Health Office Estimates, 2012. 6.2. IMPLEMENTATION OF JAMPERSAL POLICY (FINDINGS FROM THE QUALITATIVE STUDY) 6.2.1. Organizational Arrangement There were six main actors in Jampersal implemention: MoF, MoHA, Bappenas, MoH, PHO/DHO, and local government. In practice only the MoH and local governments were actively involved in Jampersal implementation. Both Garut and Depok collaborated with the Indonesia Midwives Association (IBI) to increase participation of private practice midwives. 8 BEONC is Puskesmas with ability to provide 24-hour Basic Emergency Obstetric and Neonatal Care (BEONC/PONED), including management of preeclampsia/eclampsia, shoulder dystocia, vacuum extraction, post-partum hemorrhage, puerperal infections, low birth weight, and other early neonatal conditions. 9 CEONC is hospital with ability to provide comprehensive emergency obstetric and neonatal care 10. Local government health insurance. 11. Health insurance for civil servants. 16 Table 1.6 Roles and Responsibilities of Actors under Jampersal MoF MoHA Bappenas MoH PHO/D Local HO govern ment Oversight scheme √ Financing scheme √ √ √ Benefit package determination √ Accreditation/empaneling providers √ √ Financial management/planning √ √ √ √ Setting reimbursement rates √ Claims processing/payment √ √ √ Outreach/social marketing √ √ Service delivery √ √ Monitoring local utilization √ Monitoring national utilization √ Customer service √ √ Source: Modified from the Joint Learning Network 2012. 6.2.2. Financing Utilization of Jampersal and disbursement of the allocated money for Garut increased gradually (table 1.7). In 2011, the absorption of the Jampersal budget was 43.5 percent (Rp 7.0 billion from the allocated Rp 16.2 billion), and increased to 57.0 percent in 2012 (Rp 15.2 billion from the allocated Rp 26.7 billion). By the end of the second quarter of 2013, the absorption was almost Rp 6.0 billion. Table 1.7 Budget Allocation and Utilization of Jampersal in Garut, 2011–13 Jampersal budget tranches Year I II III IV Total 2011 Allocation 6,133,501,000 8,178,002,000 1.942,169,000 - 16,253,672,000 Utilization 7,076,388,000 2012 Allocation 8,017,686,000 10,690,248,000 - 8.017,686,000 26,725,620,000 Utilization 15,258,030,000 2013* Allocation 8,390,279,000 8,770.560,000 - - 17,160,839,000 Utilization 5,759,470,750 Source: Garut District Health Office 2011-–13, *Second quarter. In 2013, the fund allocation in Depok decreased significantly (table 1.8). This was because the DHO reported utilization of Jampersal in 2012 was only around Rp 500 million, although not all services had been claimed. The study could not gain access to the disbursement data. The low utilization in Depok might be related to the high use of private providers for maternal care in Depok; many private providers did not participate in Jampersal. 17 12 Table 1.8 Budget Allocation of Jampersal in Depok, 2011–13 Budget allocation tranches Year I II III IV Total 2011 Allocation n.a. n.a. — n.a. n.a. 2012 Allocation 3,377,336,000 4,503,114,000 — 3.377,336,000 11,257,786,000 2013 Allocation 875,700,000 — — — 875,700,000 Source: Depok District Health Office, Technical Implementing Unit (Unit Pelaksana Teknis, UPT) Jamkesda, 2011–13. According to the revised 2012 guidelines, disbursement of Jampersal funds should comply with the APBD (Anggaran Pendapatan dan Belanja Daerah, local government budget mechanism). The DHO prepared an annual Jampersal budget work plan (RKA, Rencana Kerja Anggaran) based on the yearly projection of the number of deliveries in the district. The amount had to be locally approved and recorded in the APBD and would be the reference for the Jampersal money available to the district for the year. The DHO could not pay for claims above the amount set in the APBD. If the total claim was higher, as experienced by Garut, the DHO could not cover the whole claim and had to delay payment until the subsequent budget year. The DHO must return any unused Jampersal funds to KPPN V, and submit a fund utilization report to MoH at the end of the budget year. 6.2.3. Verification, Claim, and Reimbursement System 6. 2. 3. 1. Primary Care Level Jampersal used verificators to check the completeness and validity of Jampersal claims. There were seven verificators in the Garut DHO and five in the Depok DHO. In addition, at the puskesmas in Garut, a team consisting of the head of the puskesmas, the Jampersal treasurer, and the midwife coordinator verified the completeness of documents for services delivered through the puskesmas network (including the polindes) before submitting the claims to the DHO. One puskesmas did spot-- checks to uncover any irregularities by conducting home visits to Jampersal users. "I had an experience finding fictitious data. The woman turned out to be an elderly. Field verification is meant to confirm the accuracy of the claim." (Puskesmas) Staff involved in the verification process received some incentives from the local government. For delivery complications, the Maternity Care Plan (Askeb, Asuhan Kebidanan) was added to the requirements for service claims. The Garut DHO went further by introducing a local policy requiring submission of the Maternity Care Plan for all deliveries as a way to improve discipline in documentation of provided care. This local initiative was considered burdensome by the midwives, who thought the partograph 13 was a sufficient tool for recording normal as well as complicated birth delivery processes. Puskesmas and private providers must submit the required documentation for fee claim processing at the DHO. “It took between two or three weeks until one or two months to finish (verification of documents for Jampersal claims). There were so many...especially...when it has piled up for several months. Maybe for private providers there are not as many ...around two to three patients per month. But, for the puskesmas, there are tens up to hundreds for every single month. So it takes a longer time.” (DHO) 12. The researcher did not get any fund utilization data from Depok DHO. 13. Partograph is a composite graphical record of key data (maternal and fetal) during labor, entered against time on a single sheet of paper. Relevant measurements might include statistics such as cervical dilation, fetal heart rate, duration of labor, and vital signs. 18 Figure 1.5 Claim and Reimbursment Mechanism at Primary Care Level Puskesmas and private providers submit patients’ claim documents: KTP/ID card, MCH book and t h f h ti t Documents submitted to Verificator at DHO Verification is conducted for each patient. For referral cases with complication, additional documents is needed: midwifery care plan After verification, the claim documents are submitted to APBN treasurer for disbursement process Private Practice Midwives (PPM) Puskesma C The amount of funds withdrawed by The funds withdrawed by APBN APBN treasurer is in accordance with Treasurer (100% of the total the local government decree funds) will be given to DHO Treasurer DHO Treasurer will make the SP2D to withdraw the money from the Local Government Treasury If the SP2D is approved, the cash can DHO Treasurer will give the be withdrawn by DHO treasurer to be cash directly individual PPM given to each puskesmas Source: Authors. The DHO verified Jampersal claims from the puskesmas and private providers (figure 1.5). Private providers at the primary level could send claims directly to the DHO verificators. Puskesmas in rural/remote areas of Garut submitted claims every two to three months instead of monthly. On the other hand, urban puskesmas came frequently to the DHO for consultation during the verification process. In 2011, after completion of the verification process, the verficators sent the documents to the DHO Jampersal management team leader for approval. The team leader forwarded the documents to the DHO treasurer of Jamkesmas/Jampersal with the claim amount to be paid. The puskesmas would receive payment in cash from the DHO verificators. In 2012, the center changed the puskesmas claim mechanism to follow MoF regulation on district financial management. The verificators submitted verified claims to the APBN treasurer, who had the authority to withdraw money from the district APBN account. The APBN treasurer transferred the money to the local government treasury at the DHO. Fund withdrawal from the local government treasury was based on a disbursement warrant letter (Surat Perintah Pencairan Dana, SP2D). The DHO treasurer would transfer the claim payment 19 to the puskesmas, while primary level private providers obtained the payment directly from the DHO treasurer (figure 1.5). The deadline for the use of Jampersal funds in each calendar year was December 20. This was because MoH must return unused funds at the end of the year as nontax payment (Penerimaan Negara Bukan Pajak, PNBP) to the KPPN. Claims submitted after December 20 would be paid in the following year. The balance in the DHO account should be zero at the end of the year when the BPK (Badan Pemeriksa Keuangan, National Finance Audit Board) audits the DHO. The DHO was also bound by Ministry of Home Affairs (MoHA) regulation no. 262/PMK.03/2010, regarding income tax, applying 6 percent income tax (Pajak Penghasilan, PPH 21) to fees received by civil servants level III and above. The tax would be deducted from the amount paid to the puskesmas. 6. 2. 3. 2. Hospital Level P2JK assigned independent verificators to verify Jampersal and also Jamkesmas claims at the hospital level. Each hospital may also assign its own staff to verify the documents before submission to the independent verificator. Hospital Jampersal claims were based on the INA-CBGs 14 tariff (figure 1.6). Required claim documents included a copy of the medical record, including supporting examination results, family card, the mother’s and the husband’s ID cards, and the mother’s MCH book.15 The hospital would issue a statement letter for entering data into the INA-CBGs software and for verification by the independent verificator. Problems in the verification process were related mainly to software changes and broken computers. 14. Indonesia case-based groups or DRGs. 15. The MCH book records information from maternal services records (pregnancy, childbirth, and postpartum) and child services records (immunization, growth monitoring) and provides relevant information about maternal and child health care. 20 Figure 1.6 Claim and Reimbursment Mechanism at the Hospital Level The hospital prepare the documents for jampersal user claim: Medical resume, ID card, and MCH book The documents will be verified by Hospital Verificator to be inputed into INA-CBGs software All claim that has been inpjuted into the software will be re-verified by Independent Verificator from MoH The claim that has been approved by the independent verificator will be inputed in a CD to be sent to P2JK The funds that is already available in the Hospital account can be used/withdrawn by the hospital without waiting for feedback from P2JK Source: Authors. Because MoH has transferred Jampersal money to the hospital in advance, the claim payment process did not take long once the verification process was complete. So far, the amount transferred by MoH to the hospitals was enough to cover the claim payments. If all Jampersal money in the hospital account has been spent, the MoH would make another transfer to the hospital. 6.2.4. Provider Payment Scheme 6. 2. 4. 1. Primary Care According to Jampersal guidelines, a minimum 75 percent of puskesmas claims were for service fees; the remainder (25 percent) could be used by the DHO for operational costs, such as program socialization, consumables, drugs, and monitoring and evaluation. Depok allocated only 15 percent for DHO operational costs. 16 Private providers were reimbursed 100 percent of their claims. The puskesmas has the flexibility to decide on the fee amount for the providers. In some puskesmas in Garut, the amount received was shared by all puskesmas staff depending on the workload; therefore, the amount received by the actual provider was quite small even though the government had increased Jampersal service fees in 2012. 6. 2. 4. 2. Hospital Level Payment to public and private hospitals was determined according to the same INA-CBG tariffs. Public hospital management reported that the service fee for medical staff was a very small proportion of the total claim per service. This was considered unfair, particularly for specialists who saw many 16. Depok Mayor Decree no. 903/99/KPTS/dinkes/Huk/2013. 21 hospital patients. One informant stated that the service fee received by a specialist from a public hospital per patient was only 5 percent. For private hospitals, the INA-CBG payment was considered far below the private hospital tariff, even after the government increased the tariff. Cesarean section for third-class patients under INA-CBG was Rp 1.4 million, while non-Jampersal patients in a private hospital paid Rp 8 million. A private hospital informant stated 50 percent of the Jampersal payment was for the obstetrician, 30 percent for the anesthesia specialist, and 20 percent for the pediatrician. Other clinical staff such as the anesthesia assistant got Rp 50,000 to 100,000 per Jampersal patient from the hospital’s own budget. Midwives and nurses in that private hospital were considered salaried workers and did not receive anything from the Jampersal claims. 6.2.5. Socialization The study found that the central level socialization workshop preceded the distribution of Jampersal guidelines to the PHO/DHO and the health providers. The guidelines were available at the implementation level around three months after the start of the implementation or mid-2011. There were indications of some confusion at the beginning of program implementation, particularly regarding implementation procedures and the benefits package. The DHO conducted socialization for all Jampersal providers through monthly workshops, meetings, and other activities. An informant from a private clinic stated there was no follow up from the DHO to discuss agreement on the Jampersal program. In puskesmas, participants of the socialization workshop were the head of the puskesmas, the midwife coordinator as the technical implementer, and the treasurer of Jampersal. The DHO adjusted the socialization material according to participants’ duties and responsibilities. Interviews during the study revealed that the general nature of the guidelines may have undermined the wide variation of implementation conditions. “It is not suitable to the field condition. The technical guidelines are general, as if all health facilities are the same: having a doctor, midwives, complete tools, and infrastructure.” (Puskesmas) The DHO reported that questions raised by the midwives indicated they did not have access to or had not read the gudeilines. “….. Many midwives still ask questions…whether pills can be claimed. In fact, the answer is stated in the technical guidelines. Some midwives did not claim for family planning services because they did not know that postpartum family planning was covered.” (DHO) Some midwives thought Jampersal information was clear except for information on payment to the providers. They also questioned the use of the 25 percent operational costs by the DHO. Puskesmas and village midwives conducted Jampersal socialization for the community through mini workshop/meetings at the village office, posyandu, or puskesmas. In rural and remote areas, they also used community activities such as religious gatherings as venues. Socialization for Jampersal beneficiaries was done mostly by midwives for women seeking antenatal care. Interviews with women and community leaders showed that community understanding about Jampersal was inadequate. Many women did not know about Jampersal. Some informants perceived that Jampersal only covered fees for birth delivery. Others thought that Jampersal required cost- sharing and patients would have to pay half of the service costs. Some women did not want to use Jampersal because they were not convinced that services were free, particularly for treatment of complications in the hospital. Some perceived Jampersal provided lower quality of service. 22 "… Have heard of Jampersal, but I hesitated to use, I am afraid that I still have to pay." (Mothers) "I think the service is the same, but patient who pays get faster service, gets good and expensive medicine.” (Mothers) 6.2.6. Monitoring and Evaluation Jampersal monitoring and evaluation (monev) was conducted through a cascade process. The DHO conducted visits and regular meetings with the puskesmas head and with the verificators to discuss implementation issues, such as places eligible for birth delivery, claim schedule for puskesmas, disbursement schedule of Jampersal funds, and coverage report. A DHO informant reported that implementation of monev activities were not regular. At the puskesmas level, the discussion was mostly about program coverage, barriers, and problems. Overall, information gathered during the study seemed to indicate that the focus of monev was primarily on the utilization of Jampersal funds. 6.2.7. Supply-Side Capacity 6. 2. 7. 1. Human Resources Unlike Depok, Garut had a serious shortfall in HRH numbers and distribution, especially for its rural and remote areas. The local government relied mostly on support from the provincial and central governments for HRH recruitment. However, most HRH sent by the province did not want to serve in difficult areas. Recruitment of Bidan Honorer in Garut has somewhat helped the puskesmas in solving the HRH availability issue, but the compensation of Bidan Honorer has become a burden to the puskesmas. A puskesmas reported that the monthly fee of a Bidan Honorer was only Rp 100,000 per month, although they were eligible for receiving Jampersal service fees. The turnover of Bidan Honorer was higher than that of civil servants. The DHO considered this problematic because the high turnover affected puskesmas achievements. "In puskesmas BEONC X, the PTT (contract) midwife has received BEONC training, but the building was not ready. When the building is finished, the midwife has moved out…the puskesmas recruit...volunteers…. In reality there are many Bidan Honorer in the BEONC, but they do not have BEONC competencies." (DHO) In the public hospitals, the increasing number of patients due to Jampersal was not followed by a parallel increase in HRH. The hospital was lacking specialists and midwives for maternity care. Although there were obstetric interns in the hospital, the workload of the obstetricians was quite high. “Now there are eight patients, a patient has not given birth, another one came in. The surgery has just finished; there is another one already in the waiting lists for the next surgery. Honestly we really want to refuse patients, but it is not allowed…. It makes us so tired….” (Hospital) 6. 2. 7. 2. Public Health Facilities After Jampersal, all puskesmas in Garut were instructed to provide delivery services; while before Jampersal, only puskesmas with in-patients provided the services. Despite improvements to the puskesmas, for example by increasing the number of BEONC (basic essential obstetric and neonatal care) facilities, puskesmas in the rural areas of Garut were only open from 7 am to 2 pm. The midwives considered the presence of BEONC facilities in remote areas helpful, at least for stabilizing patients with complications before referring them to the hospital. Dr. Slamet Public Hospital in Garut could not fully function as a CEONC (comprehensive essential obstetric and neonatal care) hospital because it did not have sufficient staff and facilities. The neonatal intensive care unit (NICU) was available but not yet functioning. 23 "Referred obstetric patients go to emergency room where there is only one midwife per shift, incomplete equipment, and lack of facilities. If there are three referred patients ...will be a burden to the emergency room." (Hospital) One clinical staff stated that even before the Jampersal program, the district hospital has always had a high load of referred obstetric cases. “What a pity …patients must be tired of waiting, and for us this did not make our performance optimal. If the Cesarean section was delayed, we could not go home to rest.... It was so tiring and make us angry.” (Obstetrician in a district hospital) Both study areas faced problems in referring maternal complications. Depok has one public hospital with very limited capacity to provide maternal and neonatal care; as a result, the hospital often referred severe complications to other hospitals, including to public hospitals outside Depok and sometimes to private hospitals. Garut has two public hospitals, but only one could fully function as a referral hospital, causing a real burden to that hospital. The situation was exacerbated by geographical constraints. The referral hospital was difficult to access from the southern part of Garut with at least a three-hour travel time to reach the hospital. 6. 2. 7. 3. Private Providers Private sector participation in Garut was low. Only 11 of 571 private practice midwives in Garut have signed the Jampersal memorandum of understanding (MoU) with the DHO. Jampersal guidelines stated services should be provided by competent staff. There is no accreditation process prior to enrollment of the providers. Most private practice midwives in Garut engaged in dual practice as they were public sector employees in the morning. In Depok, 67 percent of private practice midwives had Jampersal MoU with the DHO. The Indonesia Midwives Association (IBI) was instrumental in increasing the involvement of private practice midwives. The acceptance of private practice midwives to Jampersal varies according to their perceived benefits. Many private providers in the two locations were dissatisfied with the service fees; this was the main reason for refusing to become Jampersal service providers. The regular tariff for delivery in puskesmas according to Garut District regulation was Rp 75,000 to Rp 125,000, depending on the type of the delivery provider. 17 Although the Jampersal tariff was already higher (Rp 500,000 for normal delivery), it was still lower than the private provider tariff. In Depok, the fee for normal delivery by a private provider was Rp 1 million, while in Garut it was Rp 850,000. For midwives with a low visit rate or with a practice in areas where the majority of the community was of low socioeconomic status, Jampersal increased the number of deliveries in their private practice. A midwife who had few patients before Jampersal mentioned that Jampersal increased her income because she saw more Jampersal patients. On the other hand, midwives with a high number of patients paying out-of-pocket were reluctant to serve Jampersal patients. A private midwife with many patients, who did not sign up for Jampersal, reported a reduced income because Jampersal shifted some of her patients to the puskesmas or to other private practice Jampersal midwives. Other midwives reported that the additional workload due to Jampersal was not comparable to the compensation received. Regardless of the midwife’s perception of Jampersal benefits, almost all midwives interviewed consistently mentioned the problem of delayed reimbursement and troublesome claim processes. One private midwife interviewed reported she stopped providing Jampersal services for those reasons. Non-Jampersal clients paid directly out-of-pocket for services received, while in Jampersal, private providers had to submit claims, and payment was often delayed. Based on DHO information, no private clinic in Garut provided Jampersal services. One private clinic owner stated the main reason for not joining Jampersal was the low reimbursement that did not cover service fees and operational costs. The tariff rate set out by INA-CBGs was much lower than the 17. Garut Local Government regulation no. 821/Kep.0006.A/Dinkes/2013. 24 usual obstetrician fee rate. Private clinics were concerned that a low fee would result in low quality, while service quality was the trademark of the private sector. Moreover, private clinics with low bed occupancy rates (BORs) could not exercise cross-subsidy. “It does not cover at all. I have to pay the wages for my staff, and in the private sector the cost is borne by the patients.” (Private Clinic Owner) “The specialist will get nothing…even the nurses will only receive 10 to 25 percent of the usual service fee. In the end I fear that it will affect the service delivery performance.” (Private Clinic Owner) There was no private hospital in Garut. All private hospitals in Depok signed a Jampersal MoU, and demanded that each private hospital get an equal share of Jampersal patients. In practice, many private hospitals in Depok did not accept Jampersal patients, despite the MOU; or they limited the number of patients. In addition, not all facilities adhered to Jampersal portability principles. Some private hospitals rejected Jampersal patients from neighboring districts, while others refused referral from private practice midwives and demanded a referral letter from a puskesmas or public hospital. The main reason for low participation of the private hospitals was the low Jampersal reimbursement. Another study found that the low involvement of private hospitals is due to unclear information about the program itself, including about the criteria of cases covered by Jampersal, the payment system, and the benefit package (Najib et al. 2012). The DHO acknowledged difficulties in engaging the private sector to join Jampersal. There was concern in the DHO that if private clinics provided Jampersal services, most patients would prefer private clinics to public hospitals or puskesmas. 6.2.8. Commitment to Jampersal Garut DHO acted to overcome some Jampersal implementation barriers, for example: (1) advocacy to district government to increase the allocation for service fees to public providers at primary care level; (2) advocacy to increase the ceiling of the Jampersal fund allocation in the APBD through a budget amendment to ensure that all claims could be covered within the ongoing year; (3) speeding up the verification process by providing incentives to the puskesmas verificators; (4) encouraging village midwives and private practice midwives to improve the quality of their practice sites for eligibility as Jampersal delivery facilities; and (5) developing a claim submission schedule for a puskesmas to better organize the verification system and prevent delays. Village government contribution to Jampersal included the following: (1) socializing Jampersal information; (2) providing temporary ID cards or statement letters of domicile; (3) facilitating the provision of a village ambulance by encouraging a Corporate Social Responsibility Program of the private sector; (4) issuing village regulation on partnership between midwife and traditional birth attendants (TBAs); and (5) approving the construction of a village polyclinic financed by PNPM 18 as the village priority. In 2013, the Depok DHO sought to improve Jampersal coverage by (1) adding a verificator at the DHO to accelerate the verification and claim payment process; (2) requesting Jampersal providers to send monthly claims to the verificator to avoid claim accumulation; (3) increasing the partnership with private practice midwives in serving Jampersal patients by providing reimbursement for consumables and drugs. Low commitment of private hospitals to Jampersal in Depok compromised the effectiveness of the Jampersal referral system. Some hospitals refused to admit referred Jampersal patients, and the DHO often had to intervene and directly call the hospital to ensure that referred patients with government insurance were accepted by the hospital. 18. PNPM (Program Nasional Pemberdayaan Masyarakat) is a national community empowerment program for poverty reduction. 25 6.2.9. Garut’s Dilemma in Implementing Jampersal Policy In response to the lack of delivery facilities in rural and remote areas, Garut allowed skilled birth attendants to charge for home deliveries to Jampersal. This local policy was inconsistent with the Jampersal guidelines. “Deviation still occurs in the Southern region and mountainous areas, please note and take this for input…. They give birth at home and still claim as Jampersal.” (DHO) Another issue was that Garut had a decree from the DHO (No 821/Kep.0006.A/Dinkes/2013) for a Service Fee Target for Local Revenues, meaning that all puskesmas had an annual target of service fee earnings that would contribute to local government revenue. The fees were collected from in- patient services, deliveries, medical treatment, laboratories, and ambulances. The fee target was based on the number of villages, types of maternal care, and the number of pregnant women. The fee for birth delivery was Rp 75,000 to Rp 125,000 per patient depending on the type of birth attendant (midwife or doctor), and Rp 200,000 per patient for delivery in a BEONC facility. The three puskesmas visited in the study had different fee targets for delivery care. The target for a BEONC facility in an urban area was Rp 12,075,000; for a BEONC facility in a remote area, Rp 7,350,000; and for a puskesmas without in-patient care in a rural area, Rp 12,700,000 per year. The local regulation was a dilemma for providers in rural and remote areas. Although Jampersal patients did not pay for delivery services, a puskesmas still had the obligation to meet the district government’s target. Some puskesmas believed an unachieved target would be carried over to the next year and decided to request a contribution from the midwives. Each puskesmas and village midwife contributed about Rp 80,000 to Rp 150,000 per month. In contrast, puskesmas in urban areas with a high visit rate could cover the target from other services. “You can imagine, the target in my puskesmas is 9 million per year, and we only have five midwives. Every month, how much do we have to contribute? There is Jampersal, but we still have to pay….” (Puskesmas) 26 6.3. RESULTS OF THE HOUSEHOLD SURVEY 6. 3. 1 Respondent Characteristics The household survey enrolled a total of 1,839 respondents, who are women of childbearing age (15 to 49 years): 921 sampled from Depok Municipality and 918 from Garut District. The sample size difference between the two locations was due to the rounding of samples per census block. The sample size per district for before and after Jampersal met the required minimum sample size. Table 1.9 presents the respondent characteristics. The majority of the sample in both districts was in the two age groups: 21-to-30 and 31-to-40 years. The proportion of respondents less than 20-years old showed a different pattern: in Garut the percentage was much higher (3.9 percent among pre- Jampersal samples and 15.6 percent among post-Jampersal samples) than in Depok (0.4 percent and 3.0 percent, respectively). Table 1.9 Respondent Characteristics Characteristics Garut Depok Before After Before After (n=462) (n=456) (n=460) (n=461) % % % % Age ≤ 20 years 3.9 15.6 0.4 3.0 21–30 year 50.4 50.7 45.0 56.6 31–40 year 36.8 30.5 47.2 36.9 >40 year 8.9 3.3 7.4 3.5 Minimum (years) 18 17 20 14 Maximum (years) 49 49 45 45 Mean (years) 30.3 28.1 31.6 29.7 Pregnancy 2–3 times pregnancy 48.9 46.9 52.6 54.9 1 time pregnancy 31.6 35.3 35.2 32.3 > 3 time pregnancy 19.5 17.8 12.2 12.8 Minimum 1 1 1 1 Maximum 10 11 7 9 Mean 2.5 2.3 2.2 2.2 Completed No school/some primary 3.0 1.8 1.3 0.9 education Primary 46.5 43.4 14.3 12.6 Secondary 47.0 49.3 72.8 73.8 Academy/university 3.5 5.5 11.5 12.8 Occupation status Working 25.1 14.0 26.7 20.4 Not working 74.9 86.0 73.3 79.6 Urban-rural Urban 45.0 45.6 100.0 100.0 Rural 55.0 54.4 0.0 0.0 Health insurance No insurance 48.3 56.1 55.4 57.5 Insurance for the poor 46.5 38.4 15.9 14.3 Other insurance 5.2 5.5 28.7 28.2 Most before and after Jampersal implementation respondents in Depok and Garut had two to three pregnancies. After Jampersal was initiated, the proportion of those with a first pregnancy was higher in Garut, but lower in Depok. For women pregnant more than three times, the proportion after Jampersal was lower in Garut, but was equal before and after Jampersal in Depok. The educational background of respondents in the two locations was also different. In Depok, almost 75 percent of the respondents had completed secondary school and less than 15 percent attended only primary school, while in Garut about the same proportions attended primary and secondary schools. In both locations, only a quarter of respondents were working. By design, Depok Municipality was selected to represent urban characteristics, while Garut Distrct was selected to represent rural characteristics, although the district had some urban areas. The percentage of poor people in Garut 27 was higher than in Depok as indicated by the higher percentage of those having the insurance for the poor in Garut. 6. 3. 2 Community Understanding about Jampersal Program The study found a gap in women’s exposure to information about Jampersal as it entered its third year of implementation. Approximately one-third (32 percent) and one-fourth (24 percent) of respondents in Depok and Garut, respectively, stated that they had never heard of Jampersal (figure 1.7). Figure 1.7 Proportion of Women Who Have Heard of Jampersal Garut Depok "I have heard that, but I did not pay attention, free No No delivery...to be 24% 32% honest...I don’t know...I rarely go to Yes the Posyandu." Yes 76% 68% (Focus Group Discussion Mothers, Non-Jampersal Among those who had heard of Jampersal, sources of information included health facilities and personnel, nonhealth providers, and mass media. Different sources provided the information in Depok and Garut. In Depok, a large proportion of women got Jampersal information from the health providers (38.3 percent) and/or the health facilities (25.2 percent); while in Garut the majority of women received information from nonhealth providers (60.3 percent), and only 8.3 percent and 17.4 percent got it from the health providers and the health facilities, respectively (figure 1.8). Figure 1.8 Source of Jampersal Information in Depok and Garut 70,0% 60,0% 50,0% 40,0% 30,0% 20,0% 10,0% 0,0% Garut Depok Health facilities 17,4% 25,2% Health provider 8,3% 38,3% Non-health provider 60,3% 41,5% Mass media 29,9% 21,4% Community activities 0,5% 0,7% For those who had heard of Jampersal, the study also measured their understanding about the program. Most women who had heard of Jampersal mentioned the program provided free delivery (mentioned by 81.8 percent of respondents in Depok, and 84.8 percent in Garut). The second most frequently mentioned benefit was free antenatal care. Only a few women mentioned the other benefits. Only 6.9 percent of women in Garut mentioned free Cesarean section as a Jampersal 28 benefit compared to 27.8 percent in Depok. Overall, the proportion of women who mentioned other Jampersal benefits apart from free delivery was lower in Garut than in Depok (figure 1.9). Figure 1.9 Perception about Jampersal 90 80 70 60 50 40 30 20 10 0 Free Free C- Free Free Free Post Do Not Free ANC Third Class Newborn Section w/ Contacepti Delivery Natal Care Know Care Indication ve Method Garut 17 84.8 6.4 5.9 6.7 6.9 4.8 6.1 Depok 18.6 81.8 10.9 10.9 11.9 27.8 9.1 10.1 Women were asked about health facilities eligible for Jampersal services. In both districts, the main response was public facilities. A small proportion of women mentioned that services at home were also free of charge under Jampersal. 29 6. 3. 3 Jampersal Utilization Jampersal utilization for the last delivery post-Jampersal was 28.1 percent in Garut and 10.3 percent in Depok. Jampersal was used mostly by less-educated women (26.2 percent), who resided in rural areas (33.5 percent), and who were in the lowest and second-lowest wealth quintile (21.8 and 23.1 percent, respectively). Jampersal use was also high among women with delivery complications (29.1 percent) (see table 1.10). Table 1.10 Utilization of Jampersal for the Last Delivery, by Respondents’ Characteristics, Depok and Garut, Post-Jampersal (after 2011) Jampersal utilization for last delivery Women's characteristics Garut Depok Garut and Depok n % p-value n % p-value n % p-value All 456 28.1 458 10.3 914 19.1 District/municipality Garut 456 28.1 0.000 Depok 458 10.3 Women's age 20–34 years 325 27.4 352 10.2 677 18.5 <20 years 43 30.2 0.874 9 22.2 0.472 52 28.8 0.185 >=35 years 88 29.5 97 9.3 185 18.9 Women’s education level completed No school/primary 206 30.6 61 11.5 267 26.2 Secondary 225 25.3 0.434 338 11.2 0.176 563 16.9 0.001 Academy/univ 25 32 59 3.4 84 11.9 Women’s occupation status Not working 392 28.6 365 11.2 757 20.2 0.555 0.175 0.072 Working 64 25 93 6.5 157 14 Number of pregnancies 2–3 pregnancies 214 27.6 251 10.4 465 18.3 First pregnancy 161 26.1 0.484 148 9.5 0.873 309 18.1 0.244 Multiple pregnancies 81 33.3 59 11.9 140 24.3 Antenatal care to health provider No 9 0 2 0 11 0 0.058 0.632 0.104 Yes 447 28.6 456 10.3 903 19.4 Complication during delivery No complication 367 22.1 286 6.3 653 15.2 0.000 0 0.000 Any complication 89 52.8 172 16.9 261 29.1 Residence Urban 208 21.6 458 10.3 n.a. 666 13.8 0.005 0.000 Rural 248 33.5 248 33.5 Insurance Other insurance 25 16 130 3.1 155 5.2 Insurance for the poor 175 32 0.180 65 13.8 0.006 240 27.1 0.000 No insurance 256 26.6 263 12.9 519 19.7 Wealth quintile (total) Lowest 90 26.7 89 16.9 179 21.8 Second 90 33.3 92 13 182 23.1 Middle 96 24 0.002 89 9 0.084 185 16.8 0.007 Fourth 89 41.6 93 6.5 182 23.6 Highest 91 15.4 95 6.3 186 10.8 30 6. 3. 4 Utilization of Maternal Health Services 1.3.4.1. Antenatal Care The total number of ANC visits and the MoH-recommended 1-1-2 visit pattern (at least one visit in the first trimester, one visit in the second trimester, and two visits in the third trimester) were the same before and after Jampersal implementation. This finding was consistent across the two study areas (table 1.11 and figure 1.10). Table 1.11 Number of ANC Visits Number of ANC Garut Depok Before After Before After (n=462) (n=456) (n=460) (n=461) 4 times or more 93.7 93.9 97.4 96.3 3 times 2.8 2.4 1.3 1.5 1–2 times 2.2 1.8 0.4 1.8 No ANC 1.3 2.0 0.9 0.4 Figure 1.10 ANC 1-1-2 Visit Pattern 100% 13.40% 16.10% 90% 21.10% 20.40% 80% 70% 60% 50% 86.60% 40% 78.90% 79.60% 83.90% 30% 20% 10% 0% BEFORE AFTER BEFORE AFTER GARUT DEPOK 1-1-2- visits Not meeting 1-1-2 visits 1-1-2- visits Not meeting 1-1-2 visits 31 1.3.4.2. Delivery After Jampersal, institutional deliveries increased by 8.4 percentage points or 13.7 percent compared to before Jampersal (table 1.12). Table 1.12 Place of Delivery by Time and by Women’s Characteristics, Total Samples (Depok and Garut) Variables Place of delivery Total p-value Home Facility n=633 n=1206 n=1839 Period Pre-Jampersal 356 (38.6) 566 (61.4) 922 (100) 0.000 Post-Jampersal 277 (30.2) 640 (69.8) 917 (100) Figure 1.11 shows that institutional deliveries in Garut increased 16.6 percentage points after Jampersal, from 30.5 percent to 47.1 percent. In Depok, institutional deliveries before Jampersal were already high (92.3 percent), and remained at relatively the same level after Jampersal (92.2 percent). However, there was a slight shift of place of delivery in Depok from private to public facilities. The change in choice of place of delivery before and after Jampersal was more apparent in Garut. Home deliveries declined by 16.6 percentage points after Jampersal, while delivery at public and private facilities increased by 5.2 percentage points and 11.4 percentage points, respectively. Figure 1.11 Place of Delivery Garut District Before 9,7% 20,8% 69,5% “Free delivery is After 14,9% 32,2% 52,9% only for delivery in PHC, maybe it’s 0,0% 20,0% 40,0% 60,0% 80,0% 100,0% called Jampersal” Public Facility Private Facility Home (FGD) Depok Municipality Before 14,3% 78,0% 7,6% After 17,6% 74,6% 7,8% 0,0% 20,0% 40,0% 60,0% 80,0% 100,0% Public Facility Private Facility Home Respondents who delivered at home were asked the reason/s for choosing home delivery. The reason most frequently mentioned was cost (34 percent), more convenient/more comfortable to deliver at home (27 percent), and delivery process started before the mother could be transported to a facility (20 percent). This pattern was the same for Garut and Depok, although in Garut a large proportion of respondents (21 percent) also expressed transport and distance to facility as reasons for choosing home delivery. 32 Table 1.13 shows the breakdown of preferred delivery facilities in the two study locations. In Garut, delivery at midwife/nurse practice increased from 19.9 percent before Jampersal to 31.8 percent after Jampersal. There was also an increase in delivery at a hospital from 8.0 percent to 11.2 percent. The proportion of deliveries at the puskesmas in Garut was very low (2.2 percent) and only increased slightly after Jampersal to 3.9 percent. The preferred place of delivery in Depok before Jampersal was the midwife/nurse practice (43.9 percent) and hospital (36.1 percent). After Jampersal, birth delivery at midwife/nurse practice in Depok decreased slightly (40.6 percent), while delivery at hospital increased slightly (37.5 percent). Interestingly, delivery at private clinics/doctors/specialist practices in Depok after Jampersal increased by almost 4 percentage points from 6.3 percent to 10.0 percent, while delivery at the puskesmas declined from 6.1 percent to 4.1 percent. Table 1.13 Breakdown of Preferred Birth Delivery Facilities Garut Depok Birth delivery facilities Before After Before After % % % % n=462 n=456 n=460 n=461 Home 69.3 52.9 7.6 7.8 Puskesmas 2.2 3.9 6.1 4.1 Midwife/nurse practice 19.9 31.8 43.9 40.6 Clinic/doctor/specialist practice 0.4 0.2 6.3 10.0 Hospital 8.0 11.2 36.1 37.5 Other 0.2 0.0 0.0 0.0 Figure 1.12 shows the proportion of deliveries by skilled birth attendants) in Depok and Garut. More than 95 percent of deliveries in Depok were by skilled birth attendants before and after Jampersal. In Garut the proportion of deliveries by SBAs was only around 60.0 percent before Jampersal, but increased to 71.7 percent after Jampersal. Figure 1.12 Skilled Birth Attendance before and after Jampersal The increased proportion of deliveries by skilled birth attendants in Garut was higher in rural than in urban areas. The increase in rural areas was around 14.5 percentage points, while in urban areas the increase was only around 9.1 percentage points (table 1.14). 33 Table 1.14 Skilled Birth Attendance before and after Jampersal, by Urban and Rural Areas, in Garut District Before After Urban n=208 n=208 (n=416) 69.7% 78.8% Rural n=254 n=248 (n=502) 51.2% 65.7% 1.3.4.3. Cesarean Section In Garut, the percentage of Cesarean sections more than doubled after Jampersal from 1.1 to 2.6 percent, although the increase is not statistically significant (p-value 0.092), and the level remains low. In Depok, the percentage of Cesarean sections after Jampersal also increased from 20.2 to 24.1 percent, which is above the WHO standard for 5 to 15 percent, but as in Garut, the increase was not statistically significant (p-value 0.159) (see table 1.15). Table 1.15 Results of Bivariate Analysis on Association between Jampersal and Cesarean Sections Variables Garut Depok Cesarian section p-value Cesarian section p-value No Yes No Yes n=901 n=17 n=717 n=204 Before Jampersal 457 (98.9) 5 (1.1) 0.092 367 (79.8) 93 (20.2) 0.159 After Jampersal 444 (97.4) 12 (2.6) 350 (75.9) 111 (24.1) 1.3.4.4. Postnatal Care As with antenatal care, there was almost no change in the proportion of postnatal care before and after Jampersal (figure 1.13). The coverage of postnatal care was low in both areas, but the coverage in Garut was much lower (around 5 percent), while in Depok the coverage was around 22 percent. It should be noted that those percentages reflect the coverage of four visits of postnatal care, according to Jampersal guidelines introduced in 2012 (see table 1.1). 34 Figure 1.13 Percentage of Postnatal Care Yes No Yes No Before 5.5% 94.5% Before 22.2% 77.8% Depok Garut After 5.3% 94.7% After 21.7% 78.3% 1.3.4.5. Family Planning Family planning services endorsed by Jampersal were the long-term contraceptive methods (IUD, 19 implant, vasectomy, tubectomy ), although short-term methods were also eligible. Figure 1.14 below shows low coverage of long-term contraceptive methods in both locations, although there was a higher increase in Depok compared to Garut after Jampersal. The graphs also indicated that the increase was due to substitution rather than to new contraceptive users. The focus group discussion (FGD) with mothers revealed that they did not like long-term contraceptives because they were afraid of the side effects. Figure 1.14 Percentage of Family Planning, before and after Jampersal, by Method Long Term Short Term Long Term Short Term "…Most of the people are still afraid, especially in using spiral 81.3% (IUD);” “…People 91.5% 89.4% 89.7% are not free to choose.” 18.7% 8.5% 10.6% 10.3% (FGD Mothers, Before After Before After Non-Jampersal Users) Garut Depok 19. Vasectomy is a surgical procedure for male sterilization and/or permanent birth control. Tubectomy is a surgical procedure for sterilization in which a woman's fallopian tubes are clamped and blocked, or severed and sealed; both method prevents eggs from reaching the uterus for fertilization. 35 1.3.4.6. Source of Payment for Last Delivery Table 1.16 Source of Payment for the Last Delivery Source of payment for Garut (n=918) Depok (n=921) delivery Before After Before After n=462 n=456 n=460 n=461 Out-of-pocket 93.3 81.6 92.2 86.3 Company 0.6 1.5 11.7 13.4 Private insurance 0.4 0.4 2.2 2.4 Askes20 0.4 1.1 1.5 0.9 Jamsostek21 0.0 0.7 3.0 4.3 Jamkesmas22 8.2 3.1 0.9 1.3 Jamkesda23 1.3 0.7 1.1 1.5 Jampersal 0.0 28.1 0.0 10.2 Others 8.4 10.3 16.5 12.8 Respondents were asked if they made any payment during the last delivery. If they made payments, they were asked to list all payment sources. Table 1.16 shows the percentage of out-of-pocket payment for delivery decreased in Garut after Jampersal from 93.3 percent to 81.6 percent, while in Depok the decline was from 92.2 percent to 86.3 percent. In Garut, the use of Jamkesda and Jamkesmas declined after Jampersal. Payment by Jampersal was higher in Garut (28.1 percent) than in Depok (10.2 percent). Among respondents who used Jampersal, almost 60 percent reported additional out-of-pocket payments. This proportion was almost the same in Garut and in Depok. From those spending additional funds, more than 55 percent and around 30 percent of respondents in Garut and Depok, respectively, could not give detailed information on the purpose of the additional payment (table 1.16). Additional payment for drugs and injections was 13.5 percent in Garut and 22.2 percent in Depok, while that for delivery services was 6.8 percent in Garut and 3.7 percent in Depok (table 1.17). 20. Askes is health insurance for civil servants and retired armed forces personnel. Active and retired civil servants, retired military and police personnel, veterans, and national patriots, and their dependents are covered by this compulsory health insurance scheme managed by PT Askes. 21. Jamsostek is social insurance for private sector workers, health insurance for formal workers and social insurance for workers in large companies, providing four programs: employment injury, death, health insurance, and a provident fund–type old-age benefit. 22. Jamkesmas is a national tax-funded health insurance plan that targets the poor and near-poor through a proxy means test targeting method. The scheme provides beneficiaries with free health services in puskesmas and third-class wards in public and designated private hospitals. 23. Jamkesda is social health insurance provided by provincial or district governments. Jamkesda typically targets people identified by the local authorities as poor but not covered by Jamkesmas (because of mistargeting or because they recently became poor due to illness, etc), with some provinces (such as Bali and Aceh) heading toward universal health insurance. Schemes vary between provinces/districts, and benefits are normally only provided through health care providers in their respective provinces. 36 Table 1.17 Type of Other Payments during Last Delivery, among Jampersal Users Garut Depok Total Utilities of payment for last delivery % % % n=74 n=27 n=101 Registration .0 3.7 1.0 Delivery service 6.8 3.7 5.9 Drug and injection 13.5 22.2 15.8 In-patient care 1.4 .0 1.0 Birth certificate 13.5 3.7 10.9 Volunteer 5.4 .0 4.0 Overall cost (could not be detailed) 55.4 29.6 48.5 Others (e.g., copying document) 5.4 25.9 10.9 6. 3. 5 Impact of Jampersal The impact of Jampersal was estimated using logistic regression, controlling for other factors that may contribute to increased institutional deliveries. The distribution of the potential contributing factors was assessed before and after Jampersal in both locations to ensure that the two populations were comparable. The logistic regression was conducted with the total samples from both locations, and for the samples from each location. 6.3.5.1. Impact of Jampersal on Institutional Deliveries Bivariate analysis of Garut showed significant association between Jampersal and institutional deliveries. The result remained significant after controlling for other contributing factors. The adjusted Odds Ratio (OR) was 2.40 (95% CI=1.74 – 3.33) (table 1.18 and annex 1). The result implied that women who delivered after the introduction of Jampersal had a 2.4 times higher chance of giving birth in health facilities compared to women who delivered before Jampersal. In other words, women who gave birth after the introduction of Jampersal had 87.7 percent probability of giving birth in health facilities. For Garut, factors contributing to institutional deliveries other than Jampersal were delivery complications and residence. Women with complications around delivery were 5.3 times more likely to give birth in health facilities compared to those without complications. Furthermore, women living in rural areas had 15 percent lower probability of giving birth in health facilities compared to those living in urban areas. For Depok, Jampersal introduction was not associated with institutional deliveries in bivariate as well as in multivariate analysis. The adjusted OR was 0.93 (95% CI=0.55 – 1.57). After controlling for other contributing factors, the strong predictor for institutional deliveries in Depok was a delivery complication (OR 3.22, 95% CI=1.59 – 6.51) (see annex 1). Women who had any complication around delivery had an approximately 3.2 times higher chance of giving birth in health facilities as opposed to those who did not have any complication. 37 Table 1.18 Association between Jampersal and Institutional Deliveries Garut District Depok Municipality Variables Crude OR p- Adjusted p- Crude OR p- Adjusted p- (95% CI) value OR value (95% CI) value OR value (95% CI) (95% CI) Time of delivery* Pre-Jampersal Reference Reference Reference Reference (2009/2010) Post-Jampersal 2.03 0.000 2.404 0.000 0.97 0.909 0.93 0.788 (2011 to 2013) (1.55–2.66) (1.74–3.33) (0.60–1.58) (0.55–1.57) * Adjusted by age, education level completed, occupation status, gravida, complication during delivery, residence, insurance, and wealth quintile. Results of the bivariate analysis also showed an increase in institutional deliveries in Garut after Jampersal among women who were least educated, poor, and lived in rural areas; while in Depok there was very little change in institutional deliveries before and after Jampersal across the same variables (table 1.19). Table 1.19 Change in Institutional Deliveries by Mother’s Education, Socioeconomic Status, Insurance Ownership, and Residence before and after Jampersal in Garut and Depok (percent) Garut Depok Women's characteristics Before After p- Before After Change Change p-value n=462 n=456 value n=460 n=461 Institutional delivery 30.5 47.1 54.4 0.000 92.4 92.2 -0.2 0.909 Women’s education level completed No school/primary 16.2 37.9 134.0 0.000 83.3 79.0 -5.2 0.525 Secondary 43.8 52.9 20.8 0.056 93.1 93.8 0.8 0.717 Academy/univ 56.3 72.0 27.9 0.303 100.0 96.6 -3.4 Residence Urban 48.1 65.4 36.0 0.000 92.4 92.2 -0.2 0.909 Rural 16.1 31.9 98.1 0.000 n.a. n.a. n.a. n.a. Insurance Other insurance 37.5 84.0 124.0 0.002 95.5 98.5 3.1 0.177 Insurance for the poor 28.4 41.7 46.8 0.006 84.9 83.3 -1.9 0.797 No insurance 31.8 47.3 48.7 0.001 92.9 91.3 -1.7 0.494 Wealth quintile (total) Lowest 10.9 24.4 123.9 0.019 86.7 88.0 1.5 0.78 Second 19.4 34.4 77.3 0.023 89.4 91.3 2.1 0.655 Middle 31.9 53.1 66.5 0.004 96.7 95.5 -1.2 0.679 Fourth 38.3 57.3 49.6 0.011 95.9 93.5 -2.5 0.476 Highest 52.2 65.9 26.2 0.059 93.2 92.6 -0.6 0.885 In Garut, after Jampersal, institutional deliveries among women with elementary school education or lower increased by 134 percent, or by more than five times compared to the increase among women of other education levels. A very high increase (77.3 percent) was also observed among the second- lowest wealth quintile (Q2). Although institutional deliveries in the lowest wealth quintile (Q1) also increased by 123.9 percent, the increase was much lower than the increase in Q2. Institutional deliveries in rural areas showed an increase by 98.1 percent compared to only 36.0 percent in urban areas. Contrary to the purpose of Jampersal, although institutional deliveries among those without 38 insurance increased after Jampersal, the change was the lowest (48.7 percent) — compared to those covered by insurance for the poor (46.8 percent) — and other insurance (124 percent). 6.3.5.2. Impact of Jampersal on Cesarian Section Bivariate analysis for Garut showed the percentage of Cesarean sections more than doubled after Jampersal implementation from 1.1 percent to 2.6 percent, but the increase was not statistically significant (p-value 0.092) (table 1.20). Other contributing factors did not show association except for education level and occupation status. Result of the multivariate analysis showed no significant correlation among the variables (table 1.21). This might be due to the small number of Cesarean sections in Garut. Table 1.20 Change in Cesarean Sections by Women’s Characteristics before and after Jampersal in Garut and Depok (percent) Garut Depok p- p- Women’s characteristics Before After Change Before After Change value value n=462 n=456 n=460 n=461 Cesarian section 1.1 2.6 136.4 0.092 20.2 24.1 19.3 0.159 Women's age 20–34 years 0.6 2.5 316.7 0.073 19.4 22.0 13.4 0.406 <20 years 0.0 0.0 n.a. n.a. — 11.1 n.a. n.a. >=35 years 2.5 4.5 80.0 0.419 22.2 33.0 48.6 0.069 Women’s education level completed No school/primary 0.4 4.4 1000.0 0.027 6.9 16.1 133.3 0.102 Secondary 1.4 0.9 -35.7 0.627 19.7 22.1 12.2 0.451 Academy/univ 6.3 4.0 -36.5 0.746 41.5 44.1 6.3 0.785 Women’s occupation status Not working 0.3 3.1 933.3 0.022 19.3 21.3 10.4 0.781 Working 3.4 0.0 -100.0 n.a. 22.8 35.1 53.9 0.046 Gravida 2–3 gravida 0.4 2.8 600.0 0.085 18.2 23.3 28.0 0.160 Primigravida 0.7 1.9 171.4 0.383 23.5 24.2 3.0 0.884 Multigravida 3.3 3.7 12.1 0.895 19.6 27.1 38.3 0.346 Antenatal care to health provider No 0.0 0.0 n.a. n.a. 0.0 0.0 n.a. n.a. Yes 1.1 2.7 145.5 0.089 20.4 24.2 18.6 0.169 Complication during delivery No complication 0.5 0.8 60.0 0.639 5.5 5.9 7.3 0.840 Any complication 3.4 10.1 197.1 0.094 50.0 54.3 8.6 0.435 Residence Urban 1.9 3.4 78.9 0.365 20.2 24.1 19.3 0.159 Rural 0.4 2.0 400.0 0.133 n.a. n.a. n.a. n.a. 39 Insurance Other insurance 4.2 4.0 -4.8 0.976 26.5 32.3 21.9 0.304 Insurance for the poor 1.4 1.1 -21.4 0.826 13.7 24.2 76.6 0.115 No insurance 0.4 3.5 775.0 0.048 18.8 20.0 6.4 — Wealth quintile (total) Lowest 0.0 0.0 n.a. n.a. 16.7 13.0 -22.2 0.493 Second 1.1 5.6 409.1 0.127 16.0 27.2 70.0 0.065 Middle 0.0 3.1 n.a. n.a. 20.9 24.7 18.2 0.539 Fourth 2.1 3.4 61.9 0.609 15.5 34.4 121.9 0.003 Highest 2.2 1.1 -50.0 0.574 33.0 21.1 -36.1 0.071 Cesarean section percentage in Depok increased after Jampersal implementation from 20 to 24 percent, but as in Garut, the increase was not statistically significant (p-value 0.159) (table 1.20). For the multivariate analysis, after controlling for other contributing factors, the result did not show significant association between Jampersal implementation and Cesarean sections. The adjusted Odds Ratio (OR) was 1.15 (95% CI=0.78 – 1.69), meaning the probability of delivery with Cesarean section was not different before and after Jampersal implementation (table 1.21). Table 1.21 Association between Jampersal and Cesarean Sections Garut Depok Variables Crude OR p- Adjusted p- Crude OR p- Adjusted OR p- (95% CI) value OR value (95% CI) value (95% CI) value (95% CI) Period of delivery* Pre-Jampersal Reference Reference Reference Reference (2009/2010) Post-Jampersal 2.470 0.092 2.45 - 1.252 0.159 1.15 0.479 (2011 s/d 2013) (0.863–7.069) (0.77–7.78) (0.916–1.710) (0.78–1.69) * Adjusted by age, education level completed, occupation status, pregnancy, ANC, complication during delivery, residence, insurance, and wealth quintile (see annex 2 for more detailed information). Table 1.22 shows the result of multivariate analysis on association between wealth quintiles and Cesarean section. Results from both locations did not show significant association between wealth quintiles and Cesarean section. The difference among the wealth quintiles was also not significant. However, the OR for the second quintile to the highest quintile was more than one, while the OR for the lowest quintile was less than one. ta 40 Table 1.22 Results of Multivariate Analysis on Association between Wealth Quintiles and Cesarean Section Garut Depok Variables Crude OR p- Adjust OR p- Crude OR p- Adjust OR p- (95% CI) value (95% CI) value (95% CI) value (95% CI) value Wealth quintile Lowest 0.000 0.995 0.000 0.995 0.495 0.008 0.524 0.043 (0.293–0.835) (0.280–0.980) Second 1.676 0.484 2.854 0.175 0.795 0.353 1.179 0.589 (0.395–7.118) (0.627–12.992) (0.491–1.289) (0.648–2.148) Middle 1.333 0.709 1.841 0.443 1.029 0.902 1.120 0.693 (0.294–6.043) (0.387–8.765) (0.655–1.616) (0.640–1.959) Fourth 1.676 0.484 1.999 0.362 0.827 0.453 0.770 0.394 (0.395–7.118) (0.451– (0.504–1.357) (0.422–1.404) 8.854) Highest Reference 0.959 Reference 0.761 Reference 0.055 Reference 0.080 * Adjusted by Jampersal period, age, education level completed, occupation status, pregnancy, ANC, complication during delivery, residence, and insurance. 41 PART 7: DISCUSSION Over the past few years, there has been a global movement to provide universal health coverage (UHC) to reduce financial barriers to health care. In 2005, the 58th World Health Assembly encouraged countries to plan health financing systems to achieve the goal of UHC such that “all people have access to services and do not suffer financial hardship paying for them” (WHO 2010). Indonesia started on the path to UHC in 2005 when Askeskin was introduced as a health insurance for the poor. Two years later, Askeskin was expanded into Jamkesmas, covering not only the poor, but also the near-poor (Harimurti et al. 2013). The local government contributed by providing Jamkesda not long after Jamkesmas was introduced to increase the coverage of insurance protection for the poor and near-poor. Substantial challenges in meeting MDGs 4 and 5 targets have led to the initiation of Jampersal policy in 2011. The formulation of Jampersal policy started in 2010 involving Bappenas (National Development Planning Agency), the Office of the Vice President, Ministry of Health, Ministry of Finance, and Ministry of People’s Welfare. The policy was launched nationwide in 2011 and aimed to cover uninsured pregnant women and newborns, irrespective of their economic strata. Thus by design, Jampersal could be considered universal coverage for maternal and newborn care. The program ended in December 2013, and the National Health Insurance Program (JKN) was launched in January 2014. The Three Dimensions of Universal Health Coverage The challenge for countries in moving toward universal health coverage is to expand the breadth, depth, and height of coverage (figure 1.15). The first dimension highlights the importance to progressively expand the coverage to include all uninsured populations, thus leading to equity across wealth, education, age, place of residence, and other population attributes. The second dimension is about expanding the range of essential health services to meet the health needs of the population, taking into consideration demand and expectations. The last dimension points out that the health care cost coverage should increase to reduce out-of-pocket copayment at service points (WHO 2010). Figure 1.15 Three Dimentsions of Universal Health Coverage Source: WHO 2010. 42 7.1. BREADTH OF THE COVERAGE: WHO IS INSURED? Who benefits the most from Jampersal. Overall, Jampersal utilization was higher among marginalized women — those with low education (26.2 percent of no school/primary school compared to 11.9 percent of academy/university graduates), poor (around 22 percent and 23 percent of the lowest and second-lowest wealth quintiles, respectively, compared to only around 11 percent of the highest wealth quintile), and from rural areas (33.5 percent of rural respondents compared to 13.8 percent of urban rspondents). Furthermore, women experiencing complications during delivery were more likely to use Jampersal. The study data show an inverse association between Jampersal use and education, as well as between Jampesal use and socioeconomic status: Jampersal use increased as the level of education decreased; similarly, the poorer the woman, the higher the Jampersal utilization. This indicates the positive influence of Jampersal in alleviating barriers to care for the most vulnerable group of women (table 1.19). Complications during pregnancy appear to be a significant factor affecting the use of Jampersal, as 29 percent of those who reported complications used Jampersal, compared to 15 percent who did not report any complication. This finding provides evidence of the positive influence of Jampersal in improving access to complication management. Another study in West Java Province, including Garut District, found perceived need was one reason for institutional delivery, meaning that women would be transferred to health facilities only if they perceived they had problems (Titaley et al. 2010), although it is not clear whether they were first told they had a complication, and hence were transferred. Jampersal targets uninsured deliveries estimated at 41.5 percent of total deliveries or, according to MoH, around 2.8 million deliveries in 2011 and in 2012. An interesting finding was the higher use of Jampersal among those who were covered by insurance for the poor, compared to uninsured women. Among Jampersal users, approximately 32 percent already had other insurance protection, indicating an overlap between Jampersal and other insurance schemes. In practice, health providers tended to claim deliveries to Jampersal, perhaps because Jampersal was specifically covering deliveries. On the demand side, Jampersal requirements were considered easy. Unlike Jamkesmas, women could obtain services without any eligibility card. They only had to bring their ID card and the MCH book. Jampersal also provided wider access to services, as benefiaries could obtain services not just from public facilities, but also from any private providers/facilities that had signed an agreement with the DHO to accept Jampersal clients. This finding was consistent with another study reporting that most Jampersal users had previously been covered by another insurance scheme, such as Jamkesda, and there was a tendency to charge Jampersal rather than Jamkesda for maternal care at district/municipality level (Rachmawati et al. 2012). The low awareness about Jampersal may have contributed to the relatively low utilization of Jampersal for the last delivery after three years of program implementation. Results of the household survey showed that around a third of women had not heard of Jampersal (32 percent in Depok and 24 percent in Garut). Most women who were aware of Jampersal did not understand the continuum of care that is covered through Jampersal. Many only knew that Jampersal was for delivery. There was also a gap in understanding about facilities providing Jampersal services. Women who did not use Jampersal expressed they were concerned that the service was not really free. Some perceived that a visit to the health facility might result in additional cost for examination or treatment. Another concern was the notion that the Jampersal service would be of lesser quality than service to patients who pay. Fear of coercion to use long-term contraception was also mentioned. Low awareness about Jampersal may be due in part to the low participation of other nonhealth sectors in socializing Jampersal during its implementation. 43 The qualitative study showed that most women who used Jampersal were pleased with the Jampersal program, and this seemed due primarily to the free nature of the service. Some reported they received less attention from health providers compared to non-Jampersal patients, but did not mind because the service was free. Some were aware that Jampersal also covered Cesarean sections, while others considered eligibility requirements for Jampersal were easy to meet. An interesting finding was Garut local government’s decision to claim home deliveries to Jampersal in areas with low availability of health workers and difficult access to delivery facilities as long as the delivery was assisted by a skilled birth attendant. Another study reported similar findings and suggested a special policy for districts with geographical challenges and limited facilities (Febriany et al. 2011). 7.2. DEPTH OF THE COVERAGE: THE BENEFIT PACKAGE The services covered within a health insurance scheme is fundamental to how the scheme affects health outcomes and financial protection (Lagomarsino et al, 2012). The success of UHC requires expansion of services, at least to meet essential care. Jampersal covers essential services within the continuum of maternity and newborn care, that is, antenatal care, delivery care, postpartum and postnatal care for mother and newborn, and family planning. MoH expanded Jampersal service coverage further by (1) increasing antenatal and postpartum care frequency for complication cases to more than the regular four times; (2) including treatment for nonobstetric, life-threatening maternal complications; (3) covering in-patient care in BEONC facilities for pregnancies with complications; (4) covering in-patient care for sick newborns; and (5) covering the management of sick newborns. Service coverage may have been lower than expected because not all providers understood the scope of Jampersal services, indicating the need to improve the clarity of the Jampersal guidelines and to encourage the use of the guidelines as a reference point. Some health officers have a negative perception of Jampersal and considered the program a threat to the success of family planning. According to them the free services might encourage people to have more children. Service availability. The capacity to deliver Jampersal primary and secondary health service in Garut may have been constrained by the relatively low availability of delivery facilities and the fact that only one hospital could provide comprehensive maternal health services. The situation in Depok was entirely different, as the city had a large number of private providers, but they were unwilling to join the Jampersal scheme. Nevertheless, the low private provider participation in Jampersal at primary level did not affect the level of institutional delivery in Depok, which was already very high before Jampersal — suggesting Jampersal did not have an impact in Depok. There were reports of difficulties in gaining access to referral care in Depok because most hospitals in the city were private and their participation in Jampersal was low, but since there were participating facilities in nearby Jakarta, access to referral care appeared not to be an issue for Depok. Coverage of MCH services. The household survey showed minimal change in the utilization of antenatal and postpartum care in the two study areas before and after Jampersal implementation. The minimal change in antenatal care coverage after Jampersal was not surprising as the coverage was already high in both locations before Jampersal. The policy to have four instances of postpartum care was relatively new, and this might explain the low coverage even after Jampersal (5.5 percent in Garut and 22.2 percent in Depok). The use of long-term contraceptives, on the other hand, increased after Jampersal in both locations, consistent with Jampersal endorsement for long-term contraceptive use. The proportion of skilled birth attendants and institutional deliveries in Garut increased after Jampersal, but there was relatively no change to both indicators in Depok. Despite the low number of midwives in rural areas, the increase in use of SBAs was higher in rural than in urban areas, suggesting Jampersal impoved access to skilled birth attendants in rural areas because the service was free. Despite Jampersal, around 30 percent of birth deliveries in Garut were still assisted by TBAs. Alleviating the risks for adverse maternal outcome may require reducing deliveries by the TBAs, among other efforts (Titaley et al. 2010). A study in West Java Province including Garut reported five main factors influencing the preference for home delivery and having the TBA as the delivery 44 assistant: economic, trust and tradition, perceived need, access to service, and the community member’s perception of the knowledge and skill of the care provider. The findings of the qualitative study in Garut confirmed the five factors. Tradition, belief in the skill of the TBAs, previous experience, fear of health care (for example, injection), difficulty in finding transportation, perception that a midwife was only for emergencies, practicality and comfort of giving birth at home, and costs were the reasons expressed by women. Jampersal may have removed the economic barrier, but there are behavioral and cultural factors that are beyond Jampersal. Moreover, there has to be trust that the delivery is really free of charge, and further that, although free, the quality of care is the same. Institutional deliveries. The multivariate analysis confirmed the impact of Jampersal in Garut after controlling for other contributing factors. The two-fold increase in institutional deliveries was statistically significant (OR 2.4; 95 % CI=1.74–3.33). This finding was consistent with the evaluation of the Delivery Fee Exemption Policy in Ghana, Senegal, and Nepal, although the latter two studies used secondary data. A population-based survey in two regions in Ghana reported evidence of institutional delivery increase by 11.9 and 5.0 percentage points after implementation of the fee exemption policy. In Senegal, facility data showed an increase in institutional deliveries from 40 to 44 percent (p-value <0.001) after one year of implementation of the policy. Health facility registers in Nepal showed a 19 percent increase of institutional deliveries after one year of implementing free delivery (Penfold et al. 2007; Witter et al. 2008; Witter et al. 2011). On the contrary, Jampersal did not have an impact on institutional deliveries in Depok (OR 0.93; 95% CI=0.55–1.57). This was because even before Jampersal, institutional delivery in Depok was already very high (92.4 percent). Nevertheless, it was interesting to find some shift of place of delivery from the private to the public sector; and a shift from delivery with private practice midwives to a higher level of care with hospitals and the private clinics/doctors/obstetricians. For policy makers, the findings in Depok were important as they suggest in urban areas with high availability of and demand for the private sector, introducing Jampersal might be irrelevant and inefficient. Many local governments, including in cities like Depok, have responded to Jampersal policy by investing in public health facilities. For example, despite the large number of hospitals in Depok, the city invested in converting three puskesmas into BEONC facilities since the introduction of Jampersal, even though there was little information on the use of these. Moreover, the shift to a higher level of care, like the specialist clinics during the last two years in Depok, indicates that for the city, service cost was not a barrier to accessing institutional delivery. Cesarean section coverage. There was a slight increase in incidence of Cesarean sections in Garut and Depok after Jampersal implementation, although the effect was not statistically significant (Garut OR 2.45; 95% CI=0.77–7.78; Depok OR 1.15; 95%=CI 0.78–1.69). Interestingly, the odds of Cesarean sections were much higher among women in Depok, compared to Garut (OR 11.257; 95% CI=6.555-19.333), suggesting a higher access to Cesarean sections in Depok compared to Garut. It should be noted that Cesarean sections in Depok were already high before Jampersal — 20 percent of all deliveries — and increased to 24 percent after Jampersal. The study could not convincingly confirm that all Cesarean sections were due to pregnancy or delivery complications, although most respondents perceived their Cesarean section was due to a complication. The low use of Cesarean sections in Garut may be related to the limited availability of Cesarean section services in the district. Another explanation for limited impact is the short implementation period of the program. 7.3. THE HEIGHT OF THE COVERAGE: FINANCIAL PROTECTION Almost 60 percent of respondents who used Jampersal in Garut and Depok reported paying for other delivery-related costs. About half of those who paid additional out-of-pocket expenses could not explain the reason for having to pay extra. Around 16 percent reported paying for drugs and injections, and 11 percent reported paying for a birth certificate (birth certificates were not a part of the benefit package). If the needed drug was not available in the health facility, the patient had to buy it from outside the facility. Some hospitals charged the patient for blood transfusion. Private practice midwives sometimes requested extra payment for long-term contraceptives because the price of the contraceptives was higher than the Jampersal fees. 45 PART 8: CONCLUSION 8.1. POLICY FORMULATION AND IMPLEMENTATION The formulation of Jampersal policy involved not only MoH but other sectors such as Bappenas, Office of the Vice President, Ministry of Finance, and Ministry of People’s Welfare. During implementation, MoH and the local governments implemented Jampersal with little engagement with the other sectors. The study revealed that some local government regulations (Perda) were not supportive of Jampersal implementation. Ministry of Home Affairs’ regulations regarding local government planning and budgeting, for example, affected Jampersal implementation, as payment for Jampersal claims could not exceed the amount allocated in the local government budget (APBD), despite the notional amount allocated by the central government. This regulation has contributed to delayed claim payments. 8.2. HEALTH SERVICE PROVISION AND PERFORMANCE After more than two years of implementation, the supply side was not entirely ready to respond to an increase in the demand for services. Limited facilities in rural and remote areas may have influenced the coverage of institutional deliveries despite Jampersal. Their condition, for example, triggered the Garut DHO policy to allow claiming home deliveries to Jampersal in areas with difficult access to birth delivery facilities as long as there was an urgent need of assistance and the delivery was by a SBA. Access to referral care in rural/remote areas was an issue in Garut because the district only had one fully functioning hospital and the hospital’s workload increased after Jampersal. The number of private facilities in Garut was very small, although most public sector midwives also had private practice office hours (dual practice). In an urban area like Depok, although there were many delivery facilities such as maternity clinics and hospitals, most of them were run by the private sector, and they were reluctant to participate in Jampersal. Although the government increased the amount of service fee in the second year of implementation, the private providers were still dissatisfied with Jampersal fees. Low hospital participation in Depok resulted in frequent referral to public hospitals outside Depok. The study noted the number of BEONC facilities in Depok increased from one to four after Jampersal. There has been no evaluation of the efficiency of investing in BEONCs compared to introducing financing policies to increase private sector participation, which was dominating health care in Depok. Other sources of provider dissatisfaction were the lengthy reimbursement process and confusion about the benefit package, for example about family planning services and coverage of complications. The issues found were not specific to this study as they were also reported in other studies conducted during the earlier period of Jampersal. There seems little change to those issues after more than two years of implementation. Moreover, Jampersal has resulted in a workload increase for public sector providers, and this has affected the income of midwives, particularly those doing dual practice. Jampersal has negatively influenced dual practice midwives who had a high number of visits to their private practice before Jampersal. On the other hand, private practice midwives with fewer patients before Jampersal have positively benefited from the program. Both Depok and Garut health offices were supportive of Jampersal implementation as indicated by the provision of consumables and medicines for private practice midwives in Garut, and by the increase in the share of provider fees from puskesmas’ claims from 75 to 90 percent, among other factors. 46 8.3. COMMUNITY ACCEPTANCE TO JAMPERSAL POLICY The community has not fully understood Jampersal. The study found that a third of women respondents had not heard of Jampersal. Among women who knew about Jampersal, most perceived Jampersal was only for free deliveries, and service was limited to puskesmas and public hospitals. Utilization was affected by uncertainty that the services were indeed free. There was also perception that “free” meant lower quality of service compared to that received by clients paying out-of-pocket. Low awareness about Jampersal may be due in part to the low participation of other nonhealth sectors in Jamperal implementation. 8.4. IMPACT OF JAMPERSAL ON MCH SERVICE COVERAGE There has been no change in antenatal and pospartum care utilization in the two study areas before and after Jampersal implementation. However, Jampersal resulted in a relative increase in the use of long-term contraception. The study observed increased use of skilled birth attendance (SBA) in Garut. The increase of SBA was higher in rural than in urban areas of Garut. There was almost no change in SBA use in Depok. However, the study found a shift of the type of preferred providers in Depok from the private to public sector providers; and from lower to higher type of providers. The study found Jampersal had a statistically significant impact in increasing institutional deliveries in Garut but not in Depok. In Garut, there was approximately 2.4 times higher likelihood of institutional deliveries after Jampersal. Nevertheless, in Garut the proportion of home deliveries was also higher than in Depok. Women’s preference for home-based delivery in Garut was influenced by values, practicality, and the comfort of home delivery, and geographical as well as transportation barriers in reaching delivery facilities. There was the perception that free service meant lower quality service, although others felt quality did not matter that much as long as the service was free. Increase in deliveries by Cesarean section was not statistically significant. Nevertheless, the likelihood to deliver by Cesarean section among women in Depok was higher than Garut (OR=11.257; 95% CI=6.555–19.333). This might be due to the easy access to hospitals in Depok. Jampersal showed higher impact in areas with low coverage and greater challenges for accessing institutional deliveries. The challenges shown in Garut were geographical and socioeconomic constraints, low availability of health facilities, and low insurance coverage. In urban areas such as Depok with easy access to health facilities — albeit private facilities — the impact of Jampersal was minimal. 47 PART 9: RECOMMENDATIONS A program like Jampersal requires strong support from stakeholders beyond the health sector. Jampersal encompassed horizontal (across sectors within the same level) and vertical (between the central and local government) collaboration. Jampersal has brought sectors together during policy formulation, but sector collaboration is important for building the momentum during implementation as well. For example: Given the decentralized system in Indonesia, MoHA’s involvement could help in building local government commitment to the program. A strong buy-in from local governments may increase subdistrict and village government support in increasing community awareness and potentially in removing other barriers to care. Joint collaboration between MoH, MoHA, and MoF in reviewing existing regulations would be helpful in finding ways to reduce the complexity of Jampersal (or other insurance) fund management and the reimbursement process. Multisector involvement is required in planning and implementing long-term investment for improving road infrastructure, transportation, and health facilities in geographically difficult areas to improve access to institutional deliveries. Jampersal could help to increase institutional deliveries while JKN voluntary participation is still low. The study has shown the potential of Jampersal in increasing institutional deliveries where coverage was low, as in Garut. The government may want to consider this finding and reevaluate the policy to terminate Jampersal implementation with the launching of the National Health Insurance Program (JKN) on January 1, 2014. Nevertheless, before continuing the implementation, it is important to analyze the cost-effectiveness of the program. Addressing “nonservice” cost is important. The study reported families still pay additional OOP costs for referral transport, and this might be a barrier to accessing care. Building linkages with other programs such as PNPM GSC (he National Program for Community Empowerment, Smart and Health Generation) might be an option for addressing this issue. Moreover, reducing OOP would also require better understanding about the reason for the OOP; for example, whether it is an issue of supply chain and drug shortages, an issue of exclusion of certain drugs from the basic benefits package, or an issue of provider prescription behavior. Regarding provider behavior, it will also be interesting to monitor the possibility of cream skimming.24 “Free” is not enough; service readiness and quality also matter. Women who experienced low quality of care tended to stop using the care, or even dissuaded other women from using care (Wairimu 2013). In Kenya, women’s refusal to use free maternity care was due to poor quality of the facilities and the rude attitude of health providers, among other factors. Continuous quality improvements should follow any effort to improve access to health care. Complementary work shows that maternal health supply-side readiness problems remain, especially in some parts of the country, and this could deter patients from utilizing care despite removal of financial barriers via programs such as Jampersal. Getting private providers on board requires a carefully designed provider payment system. In areas where private providers are dominant and demand for private provision is high, buying services from the private sector to improve access might be more efficient than expanding public sector investment. Obviously, this would require reasonable fees and quality assurance of service provision. The study also showed that support from professional associations like the Indonesia Midwives Association (IBI) could facilitate private midwife participation in Jampersal. 24. Midwives working in dual practice may decide to ask patients to come to their private practice as they can use the whole claim reimbursement for their own benefit, while in puskesmas they have to share with other staff. 48 There is a need to explore more options to improve human resources for health (HRH) availability and distribution in remote areas. Skilled birth attendance in rural Garut increased after Jampersal despite the low presence of midwives, suggesting adding more midwives could increase skilled birth attendance even more if paying for service is not an issue. Experience shows that monetary incentives are often not enough for deploying HRH to “difficult” areas. Other attractive features such as housing, children’s education, continuing education, more definitive and specific terms of services period need to be considered. In developing HRH policies, Indonesia will benefit from a labor market analysis to understand labor market dynamics influencing HRH supply and demand. In areas with difficult access to hospitals, the presence of BEONC facilities is essential. Investment in BEONCs should be followed by close monitoring and consistent support to ensure continuity of care by avoiding high turnover of trained staff and by increasing utilization. Implementation of Jampersal (or UHC) policy may have to be adjusted according to the utilization pattern for efficiency and effectiveness. The household survey showed Jampersal only had an impact in an area where institutional delivery coverage was still low, such as Garut. Further assessment is needed to translate this finding into future policy changes. 49 REFERENCES BPS (Badan Pusat Statistik, Central Bureau of Statistics). 2010. Statistical Yearbook of Indonesia. Jakarta: BPS. BPS (Badan Pusat Statistik, Central Bureau of Statistics) and Macro International. 2008. Indonesia Demographic and Health Survey 2007. Calverton, MD: BPS and Macro International. BPS (Badan Pusat Statistik, Central Bureau of Statistics), National Population and Family Planning Board (BKKBN), and Kementerian Kesehatan (Kemenkes — MOH), and ICF International. 2013. Indonesia Demographic and Health Survey 2012. Jakarta, Indonesia: BPS, BKKBN, Kemenkes, and ICF International. Cheema, S., and D. Rondinelli. 1983. Decentralization and Development: Policy Implementation in Developing Countries. Beverly Hills, CA: SAGE. DHO (District Health Office). 2011. District Health Office Profile. Depok, Indonesia: Dinas Kesehatan Depok. Dr. Slamet District Hospital Profile. 2010. Garut Health Profile. Garut, Indonesia: Rumah Sakit Umum Daerah. ———.2011. Garut Health Profile. Garut, Indonesia: Rumah Sakit Umum Daerah. ———.2012. Garut Health Profile. Garut, Indonesia: Rumah Sakit Umum Daerah. E. Febriany, J. Wijayanto, W. Ngatman, and S. Nathya.. 2011. Report of Jampersal Program Review in 6 Districts. Cipta Cara Padu Foundation, Indonesia.. Haas, Sherri, Laurel Hatt, Anthony Leegwater, Marianne El-Khoury, and Wendy Wong. 2012. “Indicators for Measuring Universal Health Coverage: A Five-Country Analysis.” (Draft) Bethesda, MD: Health Systems 20/20 Project, Abt Associates Inc. Harimurti, P., E. Pambudi, A. Pigazzini, and A. Tandon. 2013. “The Nuts and Bolts of Jamkesmas, Indonesia’s Government-Financed Health Coverage Program for the Poor and Near-Poor.” Universal Health Coverage Studies Series (UNICO) 8, World Bank, Washington, DC. Indonesia, MoH (Ministry of Health). 2010. Indonesia Health Profile 2010. Jakarta: MoH. ———. 2011. MoH regulation no.631/Menkes/PER/III/2011: Jampersal Technical Guideline 2011. Jakarta: MoH ———. 2012. MoH regulation no. 2562/MENKES/PER/XII/2011: Jampersal Technical Guideline 2012. Jakarta: MoH ———. 2012. MoH regulation no. 40 year 2012: Jamkesmas Guideline 2012. Jakarta: MoH Joint Learning Network. http://programs.jointlearningnetwork.org/content/jamkesmas. Lagomarsino, Gina, A. Garabrant, A. Adyas, R. Muga, and N. Otoo. 2012. “Moving towards Universal Health Coverage: Health Insurance Reforms in Nine Developing Countries in Africa and Asia.” The Lancet 380: 933–43. 50 Lemeshow, Stanley, David W. Hosmer, Janellae Klar, and Stephen Kaggwa Lwanga, 1990. “Adequacy of Sample Size in Health Studies.” Geneva: World Health Organization M. Najib, H. Thabrany, and F. Nandiaty. 2012. ”Analysis of the Situation to Minimize Financial Barriers to Access and Utilization of Maternal and Neonatal Health Services by the Poor.” Jakarta, Indonesia: FKM UI (Fakultas Kesehatan Masyarakat — Public Health Faculty University of Indonesia) and Expanding Maternal and Neonatal Survival, USAID, Washington, DC. Najam, A. 1995. “Learning from the Literature on Policy Implementation: A Synthesis Perspective.” Laxenburg, Austria: International Institute for Applied System Analysis. Penfold, S. E. Harrison, J. Bell, and A. Fitzmaurice. 2007. “Evaluation of the Delivery Fee Exemption Policy in Ghana: Population Estimates of Changes in Delivery Service Utilization in Two Regions.” Ghana Med J 41 (3): 100–109. PHO (Province Health Office). 2010. Province Health Office Profile. Depok, Indonesia: Dinas Kesehatan Provinsi West Java. ———. 2011. Province Health Office Profile. Depok, Indonesia: Dinas Kesehatan Provinsi West Java. ———. 2012. Province Health Office Profile. Depok, Indonesia: Dinas Kesehatan Provinsi West Java. Rachmawati, Tety et al. 2012. Riset Evaluasi Implementasi Jaminan Persalinan. Kementerian Kesehatan RI. Badan Penelitian dan Pengembangan Kesehatan. Pusat Humaniora, Kebijakan Kesehatan dan Pemberdayaan Masyarakat. Surabaya, Indonesia. RSUD (Rumah Sakit Umum Daerah Depok, Depok District Hospital). 2010. Depok Health Profile. Depok, Indonesia: RSUD. ———. 2011. Depok Health Profile. Depok, Indonesia: RSUD. ———. 2012. Depok Health Profile. Depok, Indonesia: RSUD. Titaley, C. R., C. L. Hunter, M. J. Dibley, and P. Heywood. 2010. “Why Do Some Women Still Prefer Traditional Birth Attendants and Home Delivery? A Qualitative Study on Delivery Care Services in West Java Province, Indonesia.” BMC Pregnancy and Childbirth 10: 43. Wairimu, M. 2013. “Despite Newly Free Deliveries in Kenya, Some Mothers Opt for Traditional Birth Attendants.” Global Press Institute Journal, July 24. Witter, S., M. A. Klemesu, and T. Dieng. 2008. “National Fee Exemption Schemes for Deliveries: Comparing the Recent Experiences of Ghana and Senegal.” Studies in Health Services Organisation and Policy 24, Antwerp, Belgium. Witter, S., S. Adjei, M. Armar-Kiemesu, and W. Graham. 2009. “Providing Free Maternal Health Care: Ten Lessons from an Evaluation of the National Delivery Exemption Policy in Ghana.” Global Health Action 2: 10. Witter, S., S. Khadka, H. Nath, and S. Tiwari. 2011. “The National Free Delivery Policy in Nepal: Early Evidence of Its Effects on Health Facilities.” Health Policy and Planning 26: ii84–ii91. WHO (World Health Organization). 2007–2011. WHO Country Cooperation Strategy. Jakarta, Indonesia: WHO. 51 ———. 2010. Health System Financing: The Path to Universal Coverage. The World Health Report. 12, Geneva. ———. 2013. Research for Universal Health Coverage. WHO, Geneva. World Bank. 2012.Overview of Public Expenditures on Social Assistance. Jakarta, Indonesia: World Bank. 52 ANNEXES ANNEX 1: ASSOCIATION BETWEEN JAMPERSAL AND INSTITUTIONAL DELIVERIES Table 1A.1 Association between Jampersal and Institutional Deliveries Garut Depok Adjusted p- CI 95% Adjusted p- CI 95% OR value Lower Upper OR value Lower Upper After Jampersal 2.40 0.000 1.74 3.33 0.93 0.788 0.55 1.57 Women's age 20–34 years Ref <20 years 2.03 0.046 1.01 4.06 1.06 0.957 0.11 9.95 >=35 years 1.16 0.542 0.72 1.85 1.66 0.176 0.80 3.47 Women’s education level completed No school/primary Ref Secondary 1.72 0.003 1.21 2.44 2.53 0.002 1.39 4.57 Academy/univ 2.83 0.015 1.22 6.54 5.92 0.027 1.22 28.64 Women’s occupation status Not working Ref Working 1.00 0.983 0.66 1.51 0.86 0.657 0.44 1.67 Gravida 2–3 gravidas Ref Primigravida 0.97 0.855 0.66 1.41 1.30 0.407 0.70 2.40 Multigravida 0.78 0.334 0.48 1.29 0.80 0.596 0.35 1.83 Complication during delivery No complication Ref Any complication 5.34 0.000 3.54 8.04 3.22 0.001 1.59 6.51 Residence Urban Ref Rural 0.31 0.000 0.23 0.44 1.00 n.a. n.a. n.a. Insurance Other insurance Ref Insurance for the poor 0.97 0.927 0.46 2.03 0.30 0.01 0.12 0.75 No insurance 0.79 0.513 0.38 1.61 0.53 0.125 0.24 1.19 Wealth quintile (total) Lowest Ref Second 1.36 0.287 0.77 2.38 0.98 0.946 0.48 1.98 Middle 2.01 0.013 1.16 3.48 2.20 0.091 0.88 5.52 Fourth 2.82 0.000 1.62 4.92 1.56 0.299 0.67 3.61 Highest 3.57 0.000 1.98 6.43 1.03 0.937 0.47 2.26 Obs 918 921 Pseudo R2 0.22 0.11 Source: Household Survey. 53 ANNEX 2: ASSOCIATION BETWEEN JAMPERSAL AND CESAREAN SECTION Table 2A.1 Association between Jampersal and Cesarean Section Garut Depok Adjusted p- Adjusted p- OR value Lower Upper OR value Lower Upper After Jampersal 2.45 0.128 0.77 7.78 1.15 0.479 0.78 1.69 Women's age 20–34 years Ref <20 years 1.00 0.85 0.898 0.08 9.60 >=35 years 2.61 0.159 0.69 9.96 1.32 0.265 0.81 2.17 Women’s education level completed No school/primary Ref Secondary 0.41 0.182 0.11 1.51 1.82 0.083 0.93 3.57 Academy/univ 2.08 0.517 0.23 19.11 3.92 0.002 1.66 9.29 Women’s occupation status Not working Ref Working 1.17 0.826 0.30 4.56 1.22 0.413 0.75 1.99 Gravida 2–3 gravida Ref Primigravida 1.01 0.990 0.26 3.98 1.00 0.993 0.64 1.55 Multigravida 1.34 0.682 0.33 5.43 1.00 0.990 0.52 1.91 Complication during delivery No complication Ref Any complication 11.93 0.000 3.80 37.41 18.35 0.000 12.00 28.06 Residence Urban Ref Rural 0.46 0.193 0.14 1.49 1.00 n.a. n.a. n.a. Insurance Other insurance Ref Insurance for the poor 0.22 0.139 0.03 1.64 0.85 0.627 0.44 1.65 No insurance 0.35 0.267 0.05 2.24 0.87 0.550 0.56 1.37 Wealth quintile (total) Lowest n.a. n.a. n.a. n.a. Second n.a. n.a. n.a. n.a. 1.97 0.043 1.02 3.80 Middle n.a. n.a. n.a. n.a. 1.44 0.273 0.75 2.78 Fourth n.a. n.a. n.a. n.a. 1.71 0.107 0.89 3.29 Highest n.a. n.a. n.a. n.a. 1.68 0.120 0.87 3.21 Obs 865.00 921.00 Pseudo-R2 0.26 0.30 Source: Household Survey. 54 ANNEX 3: METHODOLOGY The study applied cross-sectional design with pre/post assessment, using both qualitative and quantitative approaches. The qualitative approach provided information about the overall financing, payment, and organizational policies of the Jampersal program, and specifically the impact of the policies on HRH performance; while the quantitative approach examined whether Jampersal implementation has impact on improving coverage of services, particularly on institutional deliveries (primary level) and Cesarean section coverage (secondary level). 1. Qualitative Study The qualitative approach used in-depth interviews and focus group discussions, collecting information from central, provincial, district/municipality, and community levels. At each level stakeholders were asked particular questions about the following aspects: 1. Central and provincial levels: role in policy formulation, implementation, and evaluation, including analysis on the sustainability 2. District/municipality level: process of implementation, its barriers and enabling factors 3. Community: community leaders’ involvement; satisfaction, barriers, and enabling factors in utilizing Jampersal Table 3A.1 Informants in the Qualitative Study Institution Method Informants Note Central level MoH In-depth interview • Directorate General for Nutrition and MCH • Directorate General for Health Effort, include primary health and referral • Center for Health Financing and Insurance/P2JK • Planning Bureau • Board for Development and Empowerment of Human Resources in Health/BPPSDMK MoHA • Directorate General for Regional Autonomy/Ditjen Otoda MoF • Ministry of Financing Legislative Commission IX, which handles health issues Commission D, DPRD Garut PHO In-depth interview • MCH Unit • Unit for Health Insurance/Administration Unit/Tata Usaha Garut District DHO In-depth interview • MCH Unit • Primary Health Care Unit • Referral Health Care Unit • Unit for Health Insurance/Administration Unit/Tata Usaha Board for Regional In-depth interview Unit for Community Welfare 55 Institution Method Informants Note Development Planning (Bappeda) District hospitals In-depth interview • Vice Director/Head of Health Service Unit • Vice Director of Finance /Head of Finance Unit • Ob-Gyn • Head of delivery room Maternity clinic with In-depth interview • Ob-Gyn/ Owner of maternity clinic Cesarean section facility Health center In-depth interview • Head of Health Center/Coordinator • Tarogong PHC Midwife • Padaawas PHC • Administration Unit/JKM Unit • Peundeuy PHC Village midwives Focus Group Midwife who has worked for at least 2 • Tarogong PHC Discussion years and currently still works as village • Padaawas PHC midwife. • Peundeuy PHC Midwives will be formed to represent urban-rural setting in Garut Private midwife/dual In-depth interview Midwife who has practiced as private • Jati Village practice practice midwife for at least 2 years and • Karyamekar Village currently is still in practice. This will cover pure private and dual practice midwives. Head of village, In-depth interview Lurah or TOMA per village • Jati Village community leaders • Padaawas Village (TOMA) • Pangrumasan Village Community (group of Focus Group • Mothers of at least 2 children who • Jati Village mothers who gave Discussion born before and after Jampersal • Padaawas Village birth after the • Pangrumasan Village implementation of Jampersal) In-depth interview • Mothers of children born after • Jati Village Jampersal and USED the Jampersal • Padaawas Village scheme • Pangrumasan Village • Mothers of children born after Jampersal and DID NOT USE the Jampersal scheme Depok Municipality DHO In-depth interview • MCH Unit • Primary Health Care Unit • Referral Health Care Unit • Unit for Health Insurance • Financial Treasurer Board for Regional In-depth interview Unit for Community Welfare Development Planning (Bappeda) District hospitals In-depth interview • Vice Director/Head of Health Service Unit • Vice Director of Finance /Head of Finance Unit • Ob-Gyn 56 Institution Method Informants Note • Head of delivery room Private hospitals In-depth interview • Vice Director/Head of Health Service • Sentra Medika Unit Hospital • Vice Director/Head of Finance Unit • Hasanah Graha • Ob-Gyn/ Head of Obstetric Ward Afiah (HGA) Hospital Health center In-depth interview • Head of Health Center/Coordinator • Sukmajaya PHC Midwife • Duren Seribu PHC • Administration Unit/JKM unit Village midwives In-depth interview Midwife who has worked for at least 2 • Mekarjaya Village years and currently still works as village midwife. Midwives will be formed to represent urban-rural setting in Depok Private midwife/dual In-depth interview Midwife who has practiced as Private • Mekarjaya Village practice Practice Midwife for at least 2 years and currently is still in practice. This will cover pure private and dual practice midwives. Head of village, In-depth interview Lurah or TOMA per village • Mekarjaya Village community leaders • Bojongsari Lama (Tokoh Masyarakat — Village TOMA) Community (group of Focus group • Mothers of at least one of whom • Mekarjaya Village mothers who gave discussion was born before and one after • Tirtajaya Village birth after the Jampersal • Bojongsari Lama implementation of Jampersal) Village • Duren Seribu Village In-depth interview • Mothers of children born after • Mekarjaya Village Jampersal and USED the Jampersal • Bojongsari Lama scheme Village • Mothers of children born after Jampersal and DID NOT USE the Jampersal scheme 2. Quantitative Study Sample Size The population of this study were mothers and children as beneficiaries of the Jampersal package, that is, antenatal, delivery, postnatal care, and family planning in the two areas. Women who delivered two years before Jampersal implementation (this cut-off point is defined to reduce recall bias) were considered part of the sample for the baseline/pre, and women who delivered after Jampersal implementation will be considered as the sample for the endline/post. The sample size is calculated using a sample size formula for hypothesis testing between two population proportions for each group (Lemeshow et al. 1990). Note: n = number of sample size 57 Z1-α = Z-score for significance level of α in one-sided hypothesis testing Z1-β = Z-score for 1-β power of statistical test P = average of P1 and P2 p1 = estimated proportion at baseline survey p2 = estimated proportion at evaluation survey deff = design effect Sample was selected using cluster sampling, instead of Simple Random Sampling (SRS), mainly due to financial and time limitations. Thus, calculation of sample size uses correction of design effect. We have anticipated that there is a wide variation of number of households per village within and between study areas. Population size in Depok Municipality is around 1.7 million located in 63 villages; while population in Garut District is approximately 2.4 million located in 424 villages. This condition results in un-comparability between the study areas. Thus we used census block as the cluster. Census block (CB) is the enumeration area developed by Indonesia Central Bureau of Statistics (BPS); the latest was used for the 2010 national census. In average, each census block consists of 80 to 120 households. Since all main indicators are measured through the household survey, the sample size is calculated for each main indicator to be assessed. The largest sample size was chosen. Estimated proportion used in the sample size calculation is based on data from the 2010 National Basic Health Research (Riskesdas) using data for West Java Province. Estimation of difference coverage between pre-Jampesal and post-Jampersal implementation is 15 percent for all the main indicators, except for Cesarean section, which is estimated at 10 percent. This assumption is based on the increase of coverage data between the 2002–03 and 2007 IDHS (where no massive program is implemented), inflated by estimation of increase due to Jampersal program. Table 3A.2. Sample Size Calculation for Each Group (pre/post) for Each District Indicators Estimated Estimat Sample size Total Total % ed for group sample sample at % at indicated with 10% size for preinterve postinte under nonrespon women ntion rvention “indicators” se rate with 10% column* nonrespon se rate Delivery at facility 53.7 68.7 179 269 296 Skilled birth attendants 78.3 93.3 91 137 151 Cesarean section 15.1 25.1 274 411 453 Antenatal care according to the 67.2 82.2 142 213 235 standard** Postnatal care 75.1 90.1 108 162 179 Source: Authors’ calculation. *Significance level at 5 percent, one-side hypothesis testing; power of test 0.9; design effect=1.5. **1 time during 1st trimester, 1 time during 2nd trimester, 2 times during last trimester. According to the above sample size calculation, the highest sample size is 453 for each group (pre and post) in each study area. In total there were 906 samples for each area, equal to 1,812 samples in the two study areas. The study has collected 921 respondents in Depok and 918 respondents in Garut, a total of 1,839 respondents. Sampling Procedure In each district, a number of census blocks/CBs (here, this refers to clusters) were selected randomly as the Primary Sampling Units/PSUs. The sampling process was followed by selection of households with the designated criteria (mothers of children born either before or after Jampersal 58 implementation). The table below presents the estimated number of under-five children within each CB (born either before or after Jampersal implementation). Table 3A.3 Estimated Number of Eligible Population per Census Block # of household per # of Estimated # of households census block populationα with at least one U-5 in each CB (assumption 30%) 80–120 320–480 30 (24–36) Source: Authors’ calculation. α The number of population in each household is estimated at four per household. Based on the above table, we expect that there will be at least 24 under-five children within each CB. To control variation between CBs, the target is to enroll 18 samples in each CB. The sample has been targeted in 61 CBs per districts. In situations where the number of eligible population is the same as the needed sample size per block, all the population were enrolled as the samples. When the number of eligible population is more than the designated sample size, random selection was applied. If the eligible population is less than the sample size per cluster, additional sample was taken from the neighboring census blocks (within the same village); thus, in such situations more than 61 clusters were visited. Samples of the pre-Jampersal and post-Jampersal may come from the same household, for example, a household with a child born before Jampersal and a child born after Jampersal implementation. In each CB, the study team listed all eligible population residing within the CB area. The listing process used information from either cadres or head of the neighboring unit, who is most knowledgeable and can help prepare a list of children in the area. The sample is mother with child/children who has lived in the study area at least since 2009. 3. Data Collection Qualitative Study Qualitative data collections were conducted from May to August 2013 at district, province, and national levels. All informants were successfully interviewed except two informants in Depok — the regional representative council (DPRD) member and the Ob-Gyn of a private hospital. In Garut District, since there is no private hospital, we have replaced informants from private hospital with informants from a private maternity clinic that performed Cesarian sections. Quantitative Study Data collections for the quantitative survey were conducted from April 20, 2013, to May 25, 2013, in Depok and Garut. There were 921 respondents in Depok and 918 respondents in Garut who were interviewed. The study team visited 53 CBs in Garut and 52 CBs in Depok. The distribution of CBs in Garut was 28 in rural areas, and 25 CBs in urban area; while in Depok, all CBs were in urban area. Data collection processes of quantitative study were as follows: 1. In each district, the team consisted of a field coordinator, assistant coordinator, data collectors (nine in Garut and eight in Depok), and two data entry personnel. 2. Before data collections were started, census blocks (CBs) were already selected. Field coordinator and assistant coordinators worked on respondent sampling using household listings. In each CB, 18 respondents were selected, consisting of 9 mothers with babies born by year 2009 to year 2010, and 9 mothers with babies born by year 2011 until the study time. 59 3. In conditions where the number of informants was not enough in one CB, the field coordinator and assistant coordinator went to the nearest CB for the household listing. 4. In conditions where selected respondents were not available during home visits, a second home visit was made by data collectors. If selected respondents were not available on the second visit or refused the interview, data collectors replaced the respondents with alternate respondents who had been identified prior to data collection. 5. Spot-checks to a sample of respondents (10 percent) were randomly conducted by the field coordinator and assistant coordinator. 60 Study Limitations 1. It is still too early in the program to evaluate and assess its impact. 2. The study is only conducted in two districts, so it is difficult to represent real conditions in Indonesia. 4. Data Analysis Qualitative Study All in-depth interviews and FGDs were audio-recorded and transcripted. All transcripts were collated into matrices to identify the main findings, which were then grouped based on the themes of interest. Quantitative Study Data were entered in the field on a daily basis: (1) to identify and correct any inconsistencies of data and other problems; (2) to minimize the risk of losing data during data management. Data were entered using EPI-Info program, and transferred to SPSS for further analysis. Characteristics of the study participants were assessed to see whether the population of the preintervention and of the postintervention were comparable. Furthermore, this study examined factors and variables, either quantitatively or qualitatively, that may influence/confound the relationship between Jampersal implementation and the outcomes assessed, such as access to health services, urban-rural setting, sociodemographic characteristics of user, and existence of other related programs. Socioeconomic status was assessed using the wealth index, developed based on household assets ownership. Quantitative analysis is aimed to assess the impact of Jampersal on the coverage of institutional deliveries and Cesarean section, as well as other maternal services within the framework of the continuum of care. Changes in the outcome were examined by comparing preintervention and postintervention population. The changes were also examined by comparing Depok Municipality and Garut District to explore whether there is difference on the impact of Jampersal between districts and municipalities, in relation to urban and rural settings as well as policy implementation. 5. Data Quality Initial Test and Revision of Questionnaire Questionnaires were tested for format, wording, sequence, time needed for each questionnaire, sensitive questions, difficult/unclear questions, flow of questions, use of local terms, and problems with responses (unexpected answers, inconsistencies). Any modifications and corrections were immediately applied to adjust the questionnaire for suitability for pilot testing. Development of Manuals Manuals were developed to assist the standardization of the data collection procedures in the field. a. The General Manual consisted of the following: • The background, purpose, and rationale of the study to make sure each member of the study team had a good understanding about the study, including the questions in the questionnaires. • Explanation of the design of the study and sample selection. 61 • Organization and job description of each member of the study team, to help each member understand his or her role and responsibility in the study, especially during data collection. • Guidelines for the interviewers in finding and approaching respondents, and a description of the supervisory mechanism. • Interview techniques (quantitative and qualitative) were also included in the manual. The technique includes how to ask questions, including ethical considerations in how to avoid bias and the like. b. Questionnaires/Instrument Manual, which explains each question in the questionnaires. Training All field staff received training about the questionnaire and sampling procedures so they would have a good understanding about the study being conducted. At the beginning, all field staff participated in the general training covering all procedures of data collection in the field. Subsequently, all field staff were trained in their specific duties on the basis of their skills, experience, and educational background. The training took place over two days in classes that encompassed explanations on questionnaires, guidelines, interview techniques, and one day for pilot testing in the field. Pilot Test The pilot test was conducted in the area identical to the main study area, with the following objectives: a. To assess whether the sampling mechanism was working well, including the process of finding the households, finding and selecting respondents, and understanding the pattern of community activities. b. To identify problems and find solutions, including in the process of administering the questionnaires, and other logistical arrangements. Supervision Mechanism during Data Collection: a. Daily review of all questionnaires and reinterview of some selected respondents were done by the supervisor to secure reliability of questions filled out by interviewers. b. Regular meeting of the team took place in the base camp to check the consistency and the completeness of the questionnaire, and to prepare activities for the next day. Any problem will be identified and solved on a daily basis. c. Spot-checking or supervision visit by supervisor. Data Management a. Testing of data entry template before and after pilot test, including data entry quality check. b. Cross entry: 10 percent of sample data were reentered by different data entry staff. Consistencies were checked against the two data entries, and compared with the filled-in questionnaires. c. Data cleaning. 6. Ethical Considerations The research proposal and all related documents, including instruments, were submitted to the Institutional Review Board (IRB) of Public Health Faculty, University of Indonesia for ethical approval and clearance. Consent has been sought from selected institutions for qualitative study and from all respondents of the household survey. All key informants have been provided with information about the study. Names of key informants and respondents were kept confidential. 62 Indonesia launched Jampersal in 2011, a nationwide program to accelerate the reduction of maternal and newborn deaths. The program was financed by central government revenues and provided free and comprehensive maternal and neonatal care with an emphasis on promoting institutional deliveries. Jampersal providers were public and enlisted private facilities at the primary and secondary levels. In 2013, the World Bank and the Center for Family Welfare, University of Indonesia conducted a qualitative and quantitative study to assess the implementation and impact of the program in Garut District and Depok Municipality in West Java Province. The study found that Jampersal utilization was highest among women who were least educated, poor, and resided in rural areas. Utilization was also high among women with delivery complications. The study showed Jampersal only had an impact where institutional delivery coverage was still low such as in Garut District. In this district, women were 2.4 times more likely to have institutional deliveries after Jampersal. The finding suggests implementation of Jampersal policy may have to be adjusted according to the utilization pattern for efficiency and effectiveness. The government discontinued Jampersal with the launching of the National Health Insurance Program (JKN) on January 1, 2014. The study’s findings indicate the merit in reevaluating the policy to terminate the program, given that Jampersal helped increase institutional deliveries while voluntary participation in JKN remains low. ABOUT THIS SERIES: This series is produced by the Health, Nutrition, and Population Global Practice of the World Bank. The papers in this series aim to provide a vehicle for publishing preliminary results on HNP topics to encourage discussion and debate. The findings, interpretations, and conclusions expressed in this paper are entirely those of the author(s) and should not be attributed in any manner to the World Bank, to its affiliated organizations or to members of its Board of Executive Directors or the countries they represent. Citation and the use of material presented in this series should take into account this provisional character. For free copies of papers in this series please contact the individual author/s whose name appears on the paper. Enquiries about the series and submissions should be made directly to the Editor Martin Lutalo (mlutalo@ worldbank.org) or HNP Advisory Service (healthpop@worldbank.org, tel 202 473-2256). For more information, see also www.worldbank.org/hnppublications. 1818 H Street, NW Washington, DC USA 20433 Telephone: 202 473 1000 Facsimile: 202 477 6391 Internet: www.worldbank.org E-mail: feedback@worldbank.org