The impact of user fee exemption on maternal health care utilisation and health outcomes at mission health care facilities in Malawi

The Government of Malawi has entered into agreements with Christian health association of Malawi (CHAM) health care facilities in order to exempt their catchment populations from paying user fees. These agreements are called service level agreements (SLAs). Government in turn reimburses the CHAM health care facilities for the health services that they provide. The agreements started in 2006 with 28 out of 166 CHAM health care facilities and increased to 68 in 2010. The aim of the exemption policy is to guarantee universal access to a basic package of health care services. Although the agreements were designed to cover every health service in the basic health care package, only maternal and neonatal health services are included due to limited resources. The main objective of this thesis was to evaluate the impact of the health care financing change on health care utilisation and health. The specific objectives were as follows: first, to examine whether health care facility visits for maternal health care changed due to user fee exemption; second, to evaluate whether user fee exemption affected the choice of the health care provider where women living in the catchment areas of CHAM health care facilities with user fee exemption sought maternal health care; third, to analyse the effect of user fee exemption on birth weight and; fourth, to explore and apply novel methods in the evaluation of user fee exemption. The gradual uptake of service level agreements by CHAM health care facilities provided a natural experiment with treated and control health care facilities. An additional control group comprised of other demographic groups apart from pregnant women and neonates at CHAM health care facilities with service level agreements. In household survey data, individuals were assigned to treatment and control groups based on their proximity to either a CHAM health care facility with SLA or a CHAM health care facility without SLA. This proffered the unique opportunity to estimate the effect of a single treatment on multiple outcomes. The difference-in-differences (DiD) approach was used to obtain causal effects of user fee exemption. It was implemented in the context of fixed effects, switching regression and multinomial logit models across different chapters. Health care facility level panel data for utmost 146 health care facilities for a maximum of 8 years, 2003-2010, were used. The data were obtained from the Malawi health management xiii information system (HMIS). Linked survey data were also used. Malawi demographic and survey data for 2004 and 2010 were linked to health care facility data and then merged. Analyses that utilised health care facility data showed that user fee exemption had led to increases in first antenatal care visits in the first trimester, first antenatal care visits in any trimester, average antenatal care visits and deliveries at CHAM health care facilities with SLAs. Results from survey data showed that the probability of using a CHAM health care facility with user fee exemption for antenatal care increased, the probability of using home antenatal care declined and the probability of not using antenatal care also declined due to user fee exemption. The probability of delivering at a CHAM health care facility with SLA also increased while the probability of delivering at home declined. User fee exemption did not affect the choice of where to go for postpartum care. Results of the effect of user fee exemption on birth weight were not reported because of potential endogeneity bias arising from lack of instrumental variables for antenatal care. The key policy messages from this thesis are that the user fee exemption policy is an important intervention for increasing the utilisation of maternal health care and needs to be extended to as many CHAM health care facilities as necessary. User fee exemption is not enough, however. Other factors such as education of the woman and her husband/partner, wealth status and cultural factors are also important. This thesis has contributed to the body of knowledge in the following ways. First, it has generated evidence on the impact of user fee exemption on maternal health care utilisation and birth weight in Malawi. Second, with respect to maternal health care utilisation, the thesis has looked at variables that capture the whole maternal health care process from early pregnancy to postpartum care and in a policy relevant way. Third, the thesis has evaluated the effect of user fee exemption on a variable that have not been looked at before, first antenatal care visits in the first trimester. Fourth, the thesis has examined the effect of a single treatment on multiple outcomes in a methodologically unique way. Treatment effects, which were the changes in the probabilities of using different alternatives summed up to zero, thus showing where any increase in the probability of using the outcome of interest came from. Fifth, this thesis is first to use disequilibrium theory of demand and supply in health economics. Application of this theory entailed using switching regression models with unknown sample separation, a seldom used estimation method in health economics. This was an important contribution to the methods xiv of analysing aggregate health care utilisation. Sixth, the STATA program that was written for the estimation of the disequilibrium models was itself a very important contribution to the methods for estimating aggregate supply and demand.