Rwanda Community Performance-Based Financing Impact Evaluation 2013: Household Follow Up Survey (Rwanda)

Abstract

Since June 2006, Rwanda has implemented a national supply-side performance-based financing (PBF) program in hospitals and health centers. This ‘first generation’ PBF program provided financial rewards to health facilities in order to promote maternal, child, and HIV/AIDS healthcare. A prospective, rigorous impact evaluation (IE) was developed with the commitment of the Government of Rwanda (GOR) to assess the impacts of PBF on health outcomes and determine the possibility of scaling-up the PBF initiative nationwide. The IE showed significant positive impacts on quality of prenatal care, as well as increased utilization of institutional delivery and child preventive care services. However, it highlighted the limited effect of the supply-side intervention on other services, such as the demand for timely prenatal care and family planning.

In 2010 the ‘second generation’ of PBF in Rwanda, the Community Performance-Based Financing (CPBF) program, was initiated to tackle the remaining issue of low utilization of health services by mothers and their children. The Community PBF implemented the following three interventions: (i) demand-side in-kind incentives for women, (ii) financial rewards for community health worker (CHW) cooperatives, and (iii) combined demand-side and CHW rewards. The three CPBF interventions were introduced in October 2010 in randomly selected sectors.

The CPBF program is evaluated by a prospective, randomized impact evaluation (IE). The IE evaluates the causal effect of the CPBF interventions on maternal and child health outputs and outcomes. This report presents analysis of the IE data that assessed the program impact by comparing outcomes in sectors implementing the three different interventions and a comparison group of sectors not implementing any of the interventions. The IE relies on baseline (Q1-Q2 2010) and follow up (Q4 2013-Q2 2014) data collection, consisting of household, CHW and CHW cooperative surveys. A health facility assessment was conducted at the follow up survey.

The endline survey for the impact evaluation was fielded from November 2013 to June 2014. The University of Rwanda College of Medicine and Health Sciences School of Public Health (UR-CMHS-SPH) managed all activities related to data collection and entry. The baseline survey questionnaires were adapted to account for the modification to the sample described in the previous chapter. In addition, modules asking about the experience with the CPBF program have been added. Because the baseline survey did not include the health facility assessment, the relevant questionnaires were created by adapting the Health Results Innovation Trust Fund (HRITF) samples. Apart from the health facility assessment that was conducted in French, all other interviews were conducted in Kinyarwanda.

The household questionnaire duration was approximately 120-150 minutes after having introduced the team and explained the purpose of their visit. Each field team was composed of 8 members and led by a team supervisor. All enumerators recruited were trained in biometric and anthropometric measurement. At the end of the interview, the interviewer would give caregivers feedback about the relevant health outcomes (e.g. anemia status, child nutrition status) of the children. In the event that study personnel identified children with serious health problems, the medical professional on the team provided the caregiver with information about where and how to seek medical care (usually at the closest medical center). The study team did not provide direct medical care to participants. There was no financial payment to participants enrolled in the study. However, each household received a contribution of 3,000 RWF for the community health insurance scheme (Mutuelle de santé) for one household member.

This dataset contains the data from the follow up household survey. In each village from the baseline sample, the survey team was to re-interview the baseline women, interview the women with the most recent birth or pregnancy in the village. These women were defined as the ‘core respondents’ and their households were selected for the survey. 4,468 Households are included in the dataset, 2157 with baseline women and 2343 with a recent birth or pregnancy. Thirty-two households fall into both categories.

Because of the costs implied the large sample, the research team decided to narrow the scope of the household survey relative to that in the baseline. Some of the sections have been shortened. In addition, the survey is focused to greater extent on the core woman, defined as the woman who gave the birth that led to the selection of the household. The women questionnaire was only administered to these women (and not all women of ages 15-49 as in the baseline). The Child questionnaire only covered children of the core respondent, and not all children in the household.