Health Systems & Reform ISSN: 2328-8604 (Print) 2328-8620 (Online) Journal homepage: http://www.tandfonline.com/loi/khsr20 Myriad of Health Care Financing Reforms in Zambia: Have the Poor Benefited? Bona Mukosha Chitah, Collins Chansa, Oliver Kaonga & Netsanet Walelign Workie To cite this article: Bona Mukosha Chitah, Collins Chansa, Oliver Kaonga & Netsanet Walelign Workie (2018) Myriad of Health Care Financing Reforms in Zambia: Have the Poor Benefited?, Health Systems & Reform, 4:4, 313-323, DOI: 10.1080/23288604.2018.1510286 To link to this article: https://doi.org/10.1080/23288604.2018.1510286 Published with license by Taylor & Francis Group, LLC© 2018 International Bank for Reconstruction and Development / The World Bank Accepted author version posted online: 24 Sep 2018. Published online: 05 Nov 2018. Submit your article to this journal Article views: 266 View Crossmark data Full Terms & Conditions of access and use can be found at http://www.tandfonline.com/action/journalInformation?journalCode=khsr20 Health Systems & Reform, 4(4):313–323, 2018 Published with license by Taylor & Francis Group, LLC ISSN: 2328-8604 print / 2328-8620 online DOI: 10.1080/23288604.2018.1510286 Research Article Myriad of Health Care Financing Reforms in Zambia: Have the Poor Benefited? Bona Mukosha Chitah 1,*, Collins Chansa 2,3 , Oliver Kaonga1, and Netsanet Walelign Workie4 1 Department of Economics, University of Zambia, Lusaka, Zambia 2 Health Nutrition and Population Global Practice, World Bank Group, Lusaka, Zambia 3 Institute of Public Health, Heidelberg University, Heidelberg, Germany 4 Health Nutrition and Population Global Practice, World Bank Group, Washington, DC, USA CONTENTS Abstract—Zambia has implemented a number of financing and Introduction organizational reforms since the 1990s aimed at increasing effi- Materials and Methods ciency, enhancing equity, and improving health outcomes. This Results study reviews the distributional impact of these health reforms Discussion on enhancing equity at the regional level and for different socio- Conclusion economic groups. Data from three nationally representative household surveys were collected, and a benefit incidence ana- References lysis was conducted to determine the distributional impact over the period 2010–2015. The results show that distribution of subsidies and utilization of outpatient services at public health facilities in Zambia has consistently been in favor of urban provinces. Further, distribution of health subsidies across the ten provinces in Zambia does not correspond to reported ill- nesses in each province. The study also shows that utilization of outpatient services at public (hospitals and health centers) and private health facilities is generally in favor of the rich, and utilization of both inpatient and outpatient services at public and private health facilities benefits the rich more than the poor. And although the results show a pro-poor redistribution of benefits across income groups in 2015 compared to 2010 whereby the poorest two income groups received more than a 20% share of benefits in each quintile, the benefits were still lower than their health needs. This is contrary to the richest two income groups whose share of benefits was higher than their Keywords: benefit incidence analysis, equity, public expenditure, resource health needs in both 2010 and 2015. The study concludes that allocation, Zambia Zambia has not yet fully attained its long-term health reform Received 16 April 2018; revised 25 July 2018; accepted 7 August 2018. vision of “equity of access to quality health care” despite years *Correspondence to: Bona Mukosha Chitah; Email: bona.chitah@unza.zm Color versions of one or more of the figures in the article can be found of successive health reforms. The study calls for the Zambian online at www.tandfonline.com/khsr. government to complement strategies on financial risk protection with deliberate supply- and demand-side actions in order to © 2018 International Bank for Reconstruction and Development / The World Bank enhance equity. Improvements in long- and short-term planning This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/4.0/), which permits and regular monitoring and evaluation are critical. unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited. 313 314 Health Systems & Reform, Vol. 4 (2018), No. 4 INTRODUCTION reforms that have been implemented in Zambia between 1992 and 2017 is provided in Table 1. Though the form Globally, countries pursue socioeconomic objectives aimed and depth of these reforms have been varying, the initial at fostering human capital development as a key instrument health reform vision of “equity of access to cost-effective for sustained growth and development.1 To achieve this level quality health care as close to the family as possible (p. 2)” of development, risk factors for poverty and deprivation has remained the same.9 From Zambia’s health reform particularly in the health, education, and water and sanitation vision, intermediate health systems development objectives sectors have to be eliminated. In the health sector, this is such as equity, efficiency, access, quality, safety, and cover- particularly important because health expenditure is prone to age are prioritized. be catastrophic,2,3 and some studies show that the health One of the key health reforms that was implemented in sector is one of the major contributors to inequalities the health sector in Zambia is the decentralization of between the poor and wealthy.4 Therefore, assessing fairness health service provision in 1992–199310 and establishment in financing of health care, resource allocation, and impact of an autonomous Central Board of Health (CBoH) in of public policies on the poor is critical to monitoring and 1996 that took over the operational functions of the evaluating the attainment of health systems goals of (1) Ministry of Health (MoH).6,10 The functions of the MoH improved health status, (2) financial risk protection, (3) were restricted to policy development, norm setting, and responsiveness to needs, and (4) client satisfaction. Since regulation.11 As an implementing agency, the CBoH con- the 1990s, the government in Zambia has been implementing tracted hospital and district health boards at provincial and a number of financing and organizational reforms aimed at district levels to deliver health services through a achieving these goals.5,6 A summary of the key health Period Organization Finance Financing Modality 1992–1993 Devolution of health services Pooling of government and donor Sector-wide approach programming funds for districtsMedical user fees introduced with exemptions for the poor 1995–1996 Provider–purchaser split Basic health care package Country-wide CBoH created as an autonomous institution responsible Population-based resource allocation performance-based for purchasing health services formula contracting Functions of Medical Stores Limited restricted to storage and distribution 1998–1999 Functions of CBoH and MoH streamlined Medical Stores Limited contracted out under a lease agreement 2003–2004 Medical Stores Limited contracted out under a Medium-term expenditure framework management contract Pooled funding extended to all levels Reorganization of sector-wide approach programming Needs-based resource allocation formula coordination mechanisms Introduction of medical levy 2006–2007 Dissolution of CBoH Some donors transition from pooled Performance-based MoH assumes role of provider, purchaser, and regulator funding at the MoH to general budget contracting support discontinued Medical user fees removed in all rural areas (2006) and peri-urban areas (2007) 2011–2013 Transfer of the primary health care function from the Medical user fees removed at the entire RBF in 11 districts MoH to the Ministry of Community Development primary health care level Medical levy abolished 2015–2017 Remerger of the primary health care function to the MoH RBF in 53 districts in (2015) five out of ten Structural reorganization of the MoH (2016–2017) provinces Data sources.5,7,8 TABLE 1. Key Health Reform Areas and Elements, Zambia: 1992–2017 Chitah et al.: Myriad of Health Care Financing Reforms in Zambia 315 nationwide performance-based contracting (PBC) arrange- Though Zambia has implemented several health reforms ment that covered the entire public health system. Under and has a fiscal redistributive system including social expen- this system, the CBoH used to make direct budget trans- ditures and taxes, the impact of these reforms and policies on fers to districts after satisfactory approval of quarterly poverty reduction and shared prosperity have not been ade- performance audits and financial reports by a district bas- quately evaluated, especially in the health sector. One of the ket steering committee that used to manage pooled govern- analytical techniques that can be used to assess how public ment and donor funds.6 The rationale for implementing health spending is allocated and utilized across socioeco- PBC was to enhance value for money and results focus nomic groups is benefit incidence analysis (BIA).4 In other given that Zambia was receiving a lot of donor support words, BIA tracks the distribution of public resources across during the 1990s and early 2000s.12 PBC was abandoned different socioeconomic groups and the extent to which in 2006 but reintroduced as results-based financing (RBF) different groups are utilizing or benefiting from public in 11 districts between 2011 and 2014 and in 53 districts services.19,22 BIA is important for developing a pattern for in 2018 with support from the World Bank. total health expenditure among different providers in both The other major health financing reform that has been the public and private sectors.23 Despite its importance in implemented in Zambia was the introduction of user fees in establishing the impact of fiscal policies on addressing 199210 and abolition of user fees in rural areas, peri-urban inequities, most developing countries do not conduct BIA areas, and all primary health care facilities in 2006, 2007, studies.20 This makes it difficult to establish a benchmark and 2012, respectively.13–15 Further, a needs-based resource benefit incidence pattern, which is important for assessing allocation formula for allocating operational grants from the past policies and designing and implementing remedial stra- MoH headquarters to the districts has been in implementa- tegies if the intended goals are not met.4,20 In a few devel- tion since 2004.16 The process of developing a needs-based oping countries where BIA studies have been conducted, resource allocation formula has evolved since the 1990s. they are often limited in scope. For instance, a study by de Before the health reforms of 1992, allocation of resources la Fuente and others18 uses the commitment to equity meth- in the health sector in Zambia was based on historically odology to evaluate the fiscal system in Zambia but does not adjusted budgets.17 In 1994, the MoH implemented a for- compare the health benefits received with need for health mula for allocating financial resources to districts based on care. Secondly, dominance tests were not conducted, and this district population size and density and hospital beds at makes it difficult to understand whether the distribution of different levels of care.10 This formula was later revised to the concentration curve is entirely pro-rich or pro-poor. In include other factors such as the presence of a commercial addition, the study uses a single year to evaluate the impact bank and a second- or third-level hospital in the district, of fiscal policies and does not examine changes overtime. price of fuel, and proneness of a district to diseases such as The other study that has examined beneficiary incidence of cholera or dysentery.17 However, this formula did not com- health care utilization in Zambia presents the situation for prehensively incorporate measures of “health need.” The 2003 and 2007, and a trend analysis was not conducted.24 latter was achieved in 2004 when a material deprivation This study assesses the distributional impact of the health index incorporating demographic, health, poverty, and other reforms in Zambia on public spending on health care and socioeconomic measures of deprivation or need was devel- equity across regions and income groups by using various oped and integrated into the formula. The 2004 formula was nationally representative household surveys. By using multi- revised in 2009 to make it more equity focused. ple surveys, the study evaluates changes in the distributional In addition to the health sector, the Zambian government impact of the health reforms over a period of time. The study has been implementing a number of fiscal policies and hypothesizes that the various health reforms that have been instruments aimed at targeting public spending to poor and implemented in Zambia have benefited the poor more than vulnerable households to achieve a broader distribution of the non-poor. The study does not look at each individual the benefits of economic growth. This is against the back- health reform but examines changes in benefit incidence drop of strong economic growth between 2004 and 2014 that across different income groups over time. The ultimate has not been commensurate with reduced poverty and objective of the package of health reforms in Zambia has inequality, particularly in rural areas.18 Social spending in been to improve financial risk protection and equity of Zambia includes public spending on health and education, access to health care, particularly for the poor, as outlined social cash transfers, and subsidies for farming inputs, fuel, in the health vision.9,25 In this regard, the series of health and electricity.18 reforms was planned to be integrated and progressive to 316 Health Systems & Reform, Vol. 4 (2018), No. 4 ensure a continuous process toward the desired effect. Si ¼ unit subsidy of funding health subsidy at level i: However, considering that enhancing financial protection Hi and equity in access to health care services requires radical changes in health care financing, the Discussion section The share of total health subsidy (S) accruing to the groups provides more details on the abolition of user fees and is given by the formula below: introduction of a needs-based resource allocation formula X n   X n for district-level services. Hij Si xj ¼ ; hij Si (2) i¼1 Hi S i¼1 MATERIALS AND METHODS From this equation, the share of total health subsidy to each group is determined by two factors: (1) share of the group This study uses a repeated cross-sectional survey design and within the context of the total health visits at each level of applies the traditional BIA methodology21,22 to assess the care (hij) and (2) the share for each level of care in total distribution of public subsidies and service benefits (utiliza- health subsidy (si). tion of health services). Listed below are the key steps and From the first equation, the provincial or regional analysis activities that were undertaken: is derived as follows: 1. Using household expenditure as a measure of socio- Xn X N   X n X N economic status, quintiles were constructed and used Eijk Sik xj ¼ ; eijk sik ; to rank the population by wealth. k¼1 i¼1 Ej S k¼1 i¼1 2. Data on the utilization of health services were disag- gregated by provider, level of health care, outpatient/ in which k refers to the region specified in the unit subsidy, inpatient, and socioeconomic status. and n depicts the number of provinces (regions) under con- 3. Unit costs for outpatient and inpatient services were sideration, which in this case is ten. An assumption made in calculated by using expenditure data, population, and the literature is that the unit subsidy S ij is constant across all utilization rates. units of type i. 4. “Benefits” were calculated by expressing utilization of health services in monetary terms by multiplying utiliza- tion rates by unit costs for each socioeconomic group. Data The benefits were then aggregated across different types The primary data sets (sources) used to perform the BIA were of health services for each socioeconomic group. the 2010 and 2015 Living Conditions and Monitoring Surveys 5. Comparing the distribution of health expenditures (LCMS) and the 2014 Zambia Household Health Expenditure (subsidies) and benefits by province, providers, type and Utilization Survey (ZHHEUS). The LCMS is a repeated of health services and for the different socioeconomic nationally representative cross-sectional household survey that groups in order to determine differences in benefit uses a two-stage stratified cluster sampling method to generate incidence and with respect to need. household and individual-level information.26,27 The LCMS is designed to provide data on living conditions and welfare This study uses constant unit subsidies21 and adapts the (including poverty estimates) over time, and each survey generic formula below.19,22,23 includes modules on health, education, agriculture, household consumption and expenditure, economic and labor market X n Si X n Hij Xj ; Hij ; Si (1) activity, and so forth. The 2010 LCMS was administered to i¼1 Hi i¼1 Hi about 20,000 households and the 2015 LCMS was adminis- tered to approximately 12,260 households. Considering that where Xj is the value of total health subsidy imputed to the health modules in the LCMS are too general and do not socioeconomic group j; Hij is number of health visits of contain adequate data on health choices and spending, Zambia group j to health facilities at level i (with i = health facility conducted a nationally representative health sector–specific type); Hi is the total number of visits by different levels of household survey in 2014 that generated comprehensive data health care by different income groups; and Si is government on health expenditure and utilization. This study (the recurrent net spending (less all private payments). ZHHEUS) used a two-stage stratified sampling approach Chitah et al.: Myriad of Health Care Financing Reforms in Zambia 317 (similar to the LCMS) and gathered individual- and house- health care. Lastly, comparison of benefits to “need” for health hold-level information from 11,927 households. Because the care has been recommended for BIA studies by McIntyre and 2010 and 2015 LCMS only reported outpatient visits, the Ataguba,21 who base their argument on the standard definition of ZHHEUS was used to examine beneficiary incidence for health equity, which requires access to health services to be both inpatient and outpatient visits. aligned to health needs.28 Health need is defined as a desire for The three primary data sets were complemented by data health services by an individual that he or she is prepared to from national health accounts and the Health Management acknowledge.29 This definition of need assumes two conditions Information System from which expenditure and utilization that are necessary for need to exist:(1) health services desired are data were obtained. For example, data on utilization of health necessary to attain an individual’s goal for better health and (2) services were derived from the Health Management the goal is sufficiently meritorious.30 Therefore, to estimate need, Information System and weighted by the number of users this study uses self-rated illness conditions from the 2010 and captured in the 2010 and 2015 LCMS and the 2014 2015 LCMS and the 2014 ZHHEUS. ZHHEUS data sets. Population figures were obtained from the national census report for Zambia for 2010. RESULTS Computing Unit Subsidies and Need Distribution of Health Subsidies and Outpatient Visits at Unit subsidies were estimated from actual government recurrent Public Health Facilities by Province expenditure data as reported in the 2010, 2014, and 2015 national Figure 1 (panel A) shows that four provinces (Luapula, Southern, health accounts surveys that have been conducted in country by Copperbelt, and Eastern) recorded a reduction in their share of the Ministry of Health. To analyze benefit incidence by different total health subsidies in 2015 in comparison to 2010. The largest socioeconomic groups, households were classified by quintile reduction in the share of health subsidies was in Copperbelt and based on household expenditure levels. Expenditure was selected Southern provinces at 7% and 5%, respectively. Eastern province over income due to the reported unreliability associated with received the highest share of total health subsidies from the income measures in household surveys.22 The relative share of government in 2010 and 2015 despite a two percentage point benefits received by each socioeconomic group was then calcu- reduction between 2010 and 2015. This is followed by Lusaka lated by using the convenient regression approach22 from which and Copperbelt provinces, which ranked second and third overall, concentration indices were generated to describe the distribution respectively.a Outpatient visits at public health facilities also of benefits. In addition, the multiple comparison approach was show a reduction in four provinces (Southern, Luapula, used to determine the dominant concentration curve and level of Copperbelt, and Eastern) in 2015 compared to 2010 (Figure 1, statistical significance. In this case, a positive concentration panel B). These four provinces had ranked highest in outpatient index would signify a pro-rich distribution of health care and a visits in 2010. The largest reduction in outpatient visits was negative concentration index suggests a pro-poor distribution of observed in Eastern and Copperbelt provinces and the highest FIGURE 1. Distribution of Health Subsidies and Outpatient Visits at Public Health Facilities by Province 318 Health Systems & Reform, Vol. 4 (2018), No. 4 gain of 3% was recorded in Lusaka and Central provinces. public hospitals and private health facilities were pro-rich with Eastern province ranked first in the overall share of outpatient concentration indices of 0.058 (p < 0.01) and 0.324 (p < 0.05), visits for 2010 and 2015 and Lusaka and Southern provinces respectively. In 2015, results from both the LCMS and ZHHEUS ranked second and third, respectively. show that the distribution of benefits was pro-rich at public hospitals and private health facilities. For mission health facil- ities, the 2015 LCMS shows a pro-rich distribution of benefits Distribution of Total Subsidies in Comparison to with a concentration index of 0.093 (p < 0.1) and results from the Reported Illnesses at the Provincial Level ZHHEUS are statistically insignificant. However, at a 10% level To assess whether health subsidies are distributed in line with of significance, results from the LCMS are barely statistically reported illnesses for health care at the provincial level in Zambia, significant. On the other hand, the overall distribution of benefits the share of health subsidies for each province was compared at all public health facilities (hospitals and health centers) was with the share of the population reporting illnesses for each even in 2010 with a concentration index of 0.014 (p < 0.01) but province. Data for this exercise were drawn from the 2014 became pro-rich in 2014 with a concentration index of 0.046 ZHHEUS. The results show that distribution of health subsidies (p < 0.05). at the provincial level is not in line with reported illnesses in each Using results from the 2014 ZHHEUS, health facilities province in Zambia (Figure 2). Specifically, Eastern, Lusaka, and are further broken down by provider and facility type and Copperbelt provinces received a greater share of the subsidies by outpatient and inpatient care. The benefit incidence test even though the percentage shares of the population reporting results are shown in Table 3. The results show that the illnesses were significantly lower. All of the other seven pro- distribution of benefits at all public health facilities (all vinces, which are predominantly rural, received a lower share of types of hospitals and health centers) is generally pro-rich health subsidies despite having a larger share of the population for both inpatient and outpatient services except for district reporting illnesses. hospitals and health centers, which are pro-poor for inpa- tient services with concentration indices of −0.09 (p < 0.1) and −0.179 (p < 0.01), respectively. Furthermore, though the Distribution of Outpatient and Inpatient Health Care results for beneficiary incidence for outpatient services at Benefits by Income Groups mission health facilities are statistically insignificant, Table 2 shows dominance test results for utilization of health results for inpatient services are pro-poor with a concentra- services (or distribution of health benefits) across the various tion index of −0.158 (p < 0.1). In addition, the distribution health providers and facilities. In 2010, mission health facilities of benefits for both outpatient and inpatient services at were pro-poor with a concentration index of −0.114 (p < 0.05) and private health facilities is pro-rich. FIGURE 2. Distribution of Total Subsidies in Comparison to Reported Illnesses by Province Chitah et al.: Myriad of Health Care Financing Reforms in Zambia 319 LCMS ZHHEUS 2010 2015 2014 Provider/facility type CI SE DT CI SE DT CI SE DT Public All hospitals 0.058*** 0.028 − 0.048* 0.030 − 0.214*** 0.024 − Health centers −0.0486 0.011 + −0.023 0.017 + 0.013 0.018 n-Dom All health facilities (hospitals and health centers) 0.014*** 0.009 n-Dom 0.002 0.007 n-Dom 0.046** 0.018 − Mission health facilities −0.114** 0.031 + 0.093* 0.054 − −0.106 0.068 + Private health facilities 0.324** 0.050 − 0.597*** 0.063 − 0.686*** 0.027 − CI = Concentration Index; DT = dominance test; n-Dom = nondominance; − = the 45 degree line dominates (pro-rich); + = the concentration curve dominates (pro-poor). *p < 0.1. **p < 0.05. ***p < 0.01. TABLE 2. Benefit Incidence Test Results—Outpatient Services Outpatient Inpatient Provider/Facility Type CI SE DT CI SE DT Public Tertiary (level three) hospitals 0.523*** 0.065 − 0.528*** 0.044 − General (level two) hospitals 0.385*** 0.032 − 0.222*** 0.033 − District (level one) hospitals 0.091** 0.037 − −0.090* 0.052 + Health centers 0.013 0.018 n-Dom −0.179*** 0.022 + All hospitals (levels one, two, and three) 0.214*** 0.024 − 0.243*** 0.015 − All health facilities (hospitals and health centers) 0.046** 0.018 − 0.160*** 0.017 − All health facilities (inpatient and outpatient) 0.059*** 0.018 − Mission health facilities −0.106 0.068 + −0.158* 0.091 + Private health facilities 0.686*** 0.027 − 0.804*** 0.071 − CI = Concentration Index; DT = dominance test; n-Dom = nondominance; − = the 45 degree line dominates (pro-rich); + = the concentration curve dominates (pro-poor). *p < 0.1. **p < 0.05. ***p < 0.01. TABLE 3. Benefit Incidence Test Results—Outpatient and Inpatient Services, 2014 Distribution of Total Benefits in Comparison to Need by share of benefits in 2015 (17.5%) compared to 2010 Income Groups when they received 18% of the benefits. This suggests To further assess the distribution of health care benefits, that inequities have reduced between 2010 and 2015. we compared the need for health care with the benefits Even though there has a pro-poor redistribution of ben- received by wealth quintile (Figure 3). Overall, there has efits in 2015 whereby the bottom 20% and 40% of the been an improvement in the cumulative proportion of the population received more than a 20% share of benefits population receiving benefits relative to their need in each quintile, the distribution of benefits is still inap- (Figure 3). The lowest or poorest 60% of the population propriate because the lowest two income groups have received a lower share of benefits relative to their share higher health needs. For instance, the poorest 20% of of need in 2010, but the situation improved in 2015 with the population only received 17% of the benefits in 2010 only the poorest 40% of the population receiving a lower despite having an 18.7% share of health need. In 2015, share of benefits relative to their share of need the percentage share of benefits for the poorest 20% of (Figure 3). Furthermore, the poorest 20% of the popula- the population increased but the benefits (22.7%) were tion received a much higher percentage share of benefits still less than the health need (23.6%). In addition, for in 2015 (22.7%) compared to 2010 when they received the richest 20% of the population, the share of benefits 17% of the benefits. On the other hand, the richest 20% received was relatively higher than their health needs in of the population received a much lower percentage both 2010 and 2015. 320 Health Systems & Reform, Vol. 4 (2018), No. 4 FIGURE 3. Distribution of Total Benefits in Comparison to Need for Health Care: 2010 versus 2015 DISCUSSION the effectiveness of the user fee removal policy that was designed to increase access and utilization of quality health The aim of this study was to assess equity in the distribution care. Some studies find no evidence that removal of user fees of public subsidies and health benefits across the ten pro- has increased utilization of health care in Zambia, particularly vinces (regions) and income groups. The study also estab- for the poor.33 Further, service quality—a key factor in boost- lishes a baseline of results that can be used to benchmark the ing utilization of health services—is low in Zambia32 and impact of future policy interventions. In particular, the varied across provinces.34 This probably explains why the results show a redistribution of public health subsidies and user fees removal policy has had minimal or no impact on utilization of outpatient services across the ten provinces of increased utilization of health services in Zambia. Provision Zambia between 2010 and 2015. Despite the gains and losses of low-quality antenatal services signifies low value for money across the country, Southern, Copperbelt, Lusaka, and and high opportunity costs for mothers who have to pay for Eastern provinces remained the top four ranked provinces transport or walk to the health facility for an incomplete service with regards to receipt of health subsidies and utilization of and a missed opportunity for the government to provide com- public health services. Furthermore, the results show that the prehensive and quality antenatal care services. distribution of subsidies and utilization of outpatient services The provincial-level analysis also suggests minimal at public health facilities in Zambia has consistently been in impact of the needs-based resource allocation formula in favor of urban provinces, with the exception of Eastern addressing inequities in the distribution of operational province, which is rural. These results are similar to findings grants at the district level in Zambia. Apparently, disease in Zimbabwe where utilization of outpatient services by burden, poverty levels, and population size and density geographical regions is inconsistent and concentrated are some of the key elements of the district-level resource among the rich for central and provincial health services.31 allocation formula that has been in use since 2004 (with Though Eastern province had the highest percentage share modifications in 2009) in the public health sector in of the subsidies and outpatient utilization of health services, Zambia.35,36 Empirical studies that have evaluated the health outcomes in this province are among the poorest in implementation of the formula have concluded that the Zambia as highlighted in subsequent demographic and health formula has not been fully applied.16,17 This could surveys.32 This suggests poor quality of health services in be one of the reasons for the variations in the distribution Eastern provinces and other rural areas, particularly for mater- of public health subsidies by province in Zambia. And nal health services. For example, urban women are more likely since 2011, it is not clear how the Ministry of Health has than rural women to be provided information about pregnancy been allocating public funds to districts. This is because a complications, to be weighed, to have their blood pressure new formula should have been developed and applied in measured, and to have urine and blood samples taken during 2011 following an increase in the number of districts in antenatal care.32 Gaps in service coverage raise questions on Zambia from 72 in 2011 to about 110 in 2016.37 Chitah et al.: Myriad of Health Care Financing Reforms in Zambia 321 This study also shows that the distribution of outpatient population were much higher than those received by the benefits at private health facilities and public hospitals has richest 20% of the population. Further, the bottom 20% continually been in favor of the rich over the period and 40% of the population received more than a 20% share 2010–2015. The distribution of benefits by mission health of benefits in each quintile, which further shows that there providers, which was pro-poor in 2010, appears to have has been an inequality-reducing effect across the socioeco- changed to being pro-rich, but this result has weak statistical nomic groups in 2015 compared to 2010. Despite these significance. Nevertheless, this may be indicative of a pos- positive gains, the distribution of benefits across different sible deterioration in access to health services by the poor socioeconomic groups in 2015 is still inappropriate because over the years at mission health facilities. This is despite the the lowest two income groups still receive lower benefits in fact that mission health facilities are funded by governmentb comparison to their health needs. This is contrary to the top and are located in rural areas where most of the poor people two income groups (richest 40% of the population) where reside. This trend is similar to the distribution of overall the share of benefits received was higher than their health benefits at all public health facilities (hospitals and health needs in both 2010 and 2015. centers) that favored the rich in 2014 despite being evenly The findings above could be attributed to the user fees distributed in 2010. Further analysis of the distribution of removal policy. Though studies show that the removal of user benefits for both inpatient and outpatient services by provi- fees had no effect on increasing utilization of health services in ders shows that the rich benefit more than the poor at all Zambia, the reform has contributed to enhanced financial public health facilities (level one, two, and three hospitals protection through a 90% reduction in out-of-pocket expendi- and health centers) and private health facilities. However, tures in the population.33 However, the richest 50% of the inpatient services for public district hospitals, public health population benefit more from income transfers that have been centers, and mission health facilities are pro-poor. triggered by the user fees removal policy.33 Furthermore, Other studies that have been conducted in Zambia and in patients and family members in Zambia still incur indirect the African region show similar results. A study by the costs when accessing health care, such as transport, food, World Bank shows that total health care utilization (inpatient accommodation for family members taking care of patients, and outpatient) and outpatient utilization in Zambia were and purchase of medicines not available at the health facility.34 pro-poor in 2003 even though user fees were in place at Notwithstanding the above, there is potential for more gains this time. However, a follow-up World Bank study that was in equity in the future considering that access to health care conducted in 2017 shows that overall per capita benefits at was highly inequitable when the first set of health reforms was all public health facilities in Zambia were higher for richer implemented in 1992 but the situation has been improving households in 2015.18 During this period, user fees were not over the years.39 However, strategies on financial risk protec- in existence and one would have expected the poor to have tion need to be complemented by both supply- and demand- benefited more than the non-poor. In neighboring Tanzania, side actions in order to have more impact. But inadequate richer households benefit more from outpatient services at supply of key inputs (i.e., drugs, health workers, infrastructure regional and referral public hospitals than poorer and equipment) and insufficient demand creation limit cover- households.38 A study of child health outcomes in 56 low- age and access to health services in Zambia.39 Another way of and middle-income countries also shows persistence of evaluating the impact of financing and organizational reforms inequalities between the poor and the rich over a period of is to look at how well Zambia has been implementing succes- time rather than contraction in the gap.4 This could be sive five-year National Health Strategic Plans and annual attributed to systematic constraints and long-term failures plans and budgets. Some studies show that translation of in the health sector, limited understanding of key constraints policies and national plans into annual action plans at all levels and failures, and inadequate policy response.4 of the health system is weak.39,40 And though the annual The study also compared health benefits with need for planning process at the district level is the main vehicle for health care by province and across socioeconomic groups. translating policies into action, the quality of the annual action The results show that the distribution of health subsidies by plans in some districts is poor.40 In some districts, priority province is not in line with the share of the population areas are not adequately planned for and implementation reporting illnesses for each province. At the household capacities are low.40 level, results show a pro-poor redistribution of benefits The main limitation of this study is that BIA does not across socioeconomic groups in 2015 compared to 2010. In take into consideration opportunities at household, facility, other words, the benefits received by the poorest 20% of the and district levels. By focusing on recurrent expenditure 322 Health Systems & Reform, Vol. 4 (2018), No. 4 data, the study also overlooks differences in the availability ACKNOWLEDGMENTS of key health service delivery inputs such as human This study is part of a broader public expenditure review that resources, medicines and other essential commodities, was undertaken by the World Bank with financial support infrastructure, and equipment. Secondly, by assuming that from the Department for International Development (DfID), health services are homogeneous across all beneficiaries, Zambia. The authors acknowledge comments and contribu- the study ignores the fact that quality of health services tions from Pia Schneider, Patrick Hoang-Vu Eozenou, Reem often varies between different geographical areas and Hafez, Ellen Van De Poel, and Laura di Giorgio. between rural and urban areas. Thirdly, by using constant unit subsidies, the study overlooks differences in costs of service provision at various levels of the health system and FUNDING between rural and urban areas. As such, the study does not This work was supported by the World Bank Group (Project take into account an assessment of the efficacy or effi- Reference Number P162287). ciency of the health services. Lastly, the study uses self- reported illness as a proxy for need but this measure could be inadequate. ORCID Bona Mukosha Chitah http://orcid.org/0000-0002-7311- 5023 CONCLUSION Collins Chansa http://orcid.org/0000-0003-0982-5087 The study concludes that Zambia has not fully attained its long- term health reform vision of equity of access to quality health care as close to the family as possible despite years of succes- REFERENCES sive health reforms. The study calls for the Zambian govern- 1. Culyer AJ, Newhouse JP. Handbook of health economics. ment to complement strategies on financial risk protection with Amsterdam (Netherlands): Elsevier Science B.V.; 2000. deliberate supply- and demand-side actions in order to enhance 2. Saito E, Gilmour S, Rahman MM, Gautam GS, Shrestha PK, equity. The government should also ensure that policies and Shibuya K. Catastrophic household expenditure on health in Nepal: a cross-sectional survey. Bull World Health Organ. strategies are adequately translated and effectively implemen- 2014;92(10):760–767. doi:10.2471/BLT.13.126615. ted through successive five-year National Health Strategic 3. Van Doorslaer E, O’Donnell O, Rannan-Eliya RP, Somanathan Plans and annual action plans and budgets. Improvements in A, Adhikari SR, Garg CC, Harbianto D, Herrin AN, Huq MN, the annual planning process at the district level and regular Ibragimova S, et al. Effect of payments for health care on monitoring and evaluation of the effectiveness of policies and poverty estimates in 11 countries in Asia: an analysis of house- plans at all levels are also critical. hold survey data. The Lancet. 2006;368(9544):1357–1364. doi:10.1016/S0140-6736(06)69560-3. 4. Yazbeck A. Attacking inequality in the health sector: a synth- esis of evidence and tools. Washington (DC): World Bank Publications; 2009. NOTES 5. Gilson L, Doherty J, Lake S, McIntyre D, Mwikisa C, Thomas S. The SAZA study: implementing health financing reform in South [a] Lusaka province recorded a three percentage point increase between Africa and Zambia. Health Policy Plan. 2003;18(1):31–46. 2010 and 2015. 6. Lake S, Musumali C. Zambia: the role of aid management in [b] The Zambian government funds all mission health facilities under the sustaining visionary reform. Health Policy Plan. 1999;14 Churches Health Association of Zambia through a monthly operational (3):254–263. grant and salaries for health workers. Churches Health Association of 7. Chansa C Evolution of the Zambia Health Sector – Key Reforms, Zambia health facilities enjoy the same privileges as government health Legislature, & Major Events: 1992-2016. forthcoming. facilities. 8. Chansa C. Zambia’s health sector wide approach (SWAp) revisited. Köln (Germany): Lambert Academic Publishing; 2009. 9. Ministry of Health. National health policies and strategies. DISCLOSURE OF POTENTIAL CONFLICTS OF Lusaka (Zambia): Ministry of Health; 1991. INTEREST 10. Bossert T, Chitah MB, Bowser D. Decentralization in Zambia: resource allocation and district performance. Health Policy The authors declare no conflict of interest. Plan. 2003;18(4):357–369. Chitah et al.: Myriad of Health Care Financing Reforms in Zambia 323 11. Foltz A-M. Policy Analysis. In Comprehensive Review of the 26. Central Statistical Office [Zambia]. Living conditions monitor- Zambian Health Reforms Volume II: Technical Reports. ing survey Report 2006 and 2010. Lusaka (Zambia): Central Lusaka (Zambia): Ministry of Health; 1997. Statistical Office; 2011. 12. Chansa C, Sundewall J, McIntyre D, Tomson G, Forsberg BC. 27. Central Statistical Office [Zambia]. 2015 living conditions Exploring SWAp’s contribution to the efficient allocation and monitoring survey report. Lusaka (Zambia): Central use of resources in the health sector in Zambia. Health Policy Statistical Office; 2016. Plan. 2008;23(4):244–251. doi:10.1093/heapol/czn013. 28. Gwatkin DR, Bhuiya A, Victora CG. Making health systems 13. Masiye F, Chitah BM, McIntyre D. From targeted exemptions more equitable. The Lancet. 2004;364(9441):1273–1280. to user fee abolition in health care: experience from rural doi:10.1016/S0140-6736(04)17145-6. Zambia. Soc Sci Med. 2010;71(4):743–750. doi:10.1016/j. 29. World Health Organisation. Health Systems Strengthening socscimed.2010.04.029. Glossary; 2011. [accessed 2018 Feb 20]. http://www.who.int/ 14. Masiye F, Kaonga O, Kirigia JM. Does user fee removal policy healthsystems/Glossary_January2011.pdf. provide financial protection from catastrophic health care pay- 30. Culyer AJ, Wagstaff A. Need, equity, and equality in health ments? Evidence from Zambia. PLoS One. 2016;11(1): and health care. York (UK): University of York; 1992. Report e0146508. doi:10.1371/journal.pone.0146508. No. 95. 15. Carasso BS, Lagarde M, Cheelo C, Chansa C, Palmer N. 2012. 31. Shamu S, January J, Rusakaniko S. Who benefits from public Health worker perspectives on user fee removal in Zambia. health financing in Zimbabwe? Towards universal health cov- Hum Resour Health. 10:40. doi:10.1186/1478-4491-10-40. erage. Global Public Health. 2017;12(9):1169–1182. 16. Chileshe L. An assessment of the financial resource allocation doi:10.1080/17441692.2015.1121283. criteria for districts health services in Zambia: from an Equity 32. Central Statistical Office [Zambia], Ministry of Health Perspective [Masters thesis]. Baifa Resources. 2013. Royal [Zambia], ICF International. Zambia Demographic and Tropical Institute, Vrije Universiteit Amsterdam. Health Survey 2013-14. Lusaka (Zambia): Central Statistical 17. Chitah B, Masiye F Deprivation-based resource allocation Office. 2014. criteria in the Zambian health service: A review of the imple- 33. Lépine A, Lagarde M, Le Nestour A. How effective and fair is mentation process. Harare (Zimbabwe): EQUINET; 2007. user fee removal? Evidence from Zambia using a pooled Report No. 51. synthetic control. Health Econ. 2017;1:16. 18. De La Fuente A, Rosales M, Jellema JR The impact of fiscal 34. Chama-Chiliba CM, Koch SF. An assessment of the effect of policy on inequality and poverty in Zambia. Washington (DC): user fee policy reform on facility-based deliveries in rural World Bank; 2017. Report No. WPS8246. Zambia. BMC Research Notes. 2016;9(1):504. doi:10.1186/ 19. Castrol–Leal F, Dayton J, Demery L, Mehra K. Public spend- s13104-016-1938-1. ing on health in Africa: do the poor benefit? Bull World Health 35. Mushota K, Tembo S. Revised resource allocation formula for Organisation. 2000;1(78):66–74. district-level services in Zambia. Lusaka (Zambia): Ministry of 20. Davoodi HR, Tiongson ER, Asawanuchit SS How useful are Health; 2009. benefit incidence analyses of public education and health 36. Mushota K, Tembo S. Resource allocation formula for district- spending? IMF Working Paper; 2003 [accessed 2017 Nov level services in Zambia. Lusaka (Zambia): Ministry of 24] http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1. Health; 2004. 558.4225&rep=rep1&type=pdf. 37. World Bank. Zambia health sector public expenditure review. 21. McIntyre D, Ataguba JE. How to do (or not to do) … a benefit Washington (DC): World Bank; 2018. incidence analysis. Health Policy Plan. 2011;26(2):174–182. 38. Mtei G, Makawia S, Ally M, Kuwawenaruwa A, Meheus F, doi:10.1093/heapol/czq031. Borghi J. Who pays and who benefits from health care? An 22. O’Donnell O, Doorslaer EV, Wagstaff A, Lindelow M. assessment of equity in health care financing and benefit Analyzing health equity using household survey data: A distribution in Tanzania. Health Policy Plan. 2012;27 guide to techniques and their implementation. Washington (suppl_1):i23–i34. doi:10.1093/heapol/czs018. (DC): World Bank; 2008. 39. Ministry of Health. Mid term review of the Zambia national 23. Demery L Benefit incidence: a practitioner’s guide. health strategic plan (NHSP) IV: 2006–2010. Lusaka Washington (DC): World Bank; 2000. Report No. 35117. (Zambia): Ministry of Health; 2008. 24. World Bank. Health equity and financial protection report – 40. Sikapande BM Evaluation of public health care spending: A Zambia. Washington (DC): World Bank; 2012. Report No. 71260. case study of three districts in Zambia Arusha. 2014. Eastern 25. Ministry of Health. National health policy. Lusaka (Zambia): & Southern Africa Management Institute and Maastricht Ministry of Health; 2012. School of Management.