Effects of results-based financing intervention on technical efficiency of health services in Afghanistan

Abstract

The government of Afghanistan has been piloting a Result-Based Financing (RBF) Intervention to improve key maternal and child health indicators and the quality of health services. Data Envelopment Analysis (DEA) was used to analyze and compare the relative technical efficiency of the RBF intervention in health facilities and their associated control facilities for delivering the Basic Package of Health Services (BPHS) in Afghanistan. The study period is one year (July 2011-June 2012) in 372 health facilities. In the following stage, regression analysis was used to determine the factors of technical efficiency in the BPHS facilities. The descriptive statistics show that with similar inputs across both treatment and control arms, average outputs in Basic Health Centers (BHC) were higher for the treatment arm compared to the control arm, with statistical significance (p < 0.05): 10% for outpatient visits, 24% for antenatal care, 21% for postnatal care and 29% for institutional deliveries. A separate DEA analysis based on the type of each health facility was conducted and showed that on average the treatment arm of district hospitals’ technical efficiency constant return to scale (TECRS) was 87%; for technical efficiency variable return to scale (TEVRS) was 95.2%; and scale efficiency (SE) was 90.7%, which indicates a 9% improvement in TECRS and 8.7% in SE compared to the control arm of district hospitals. Three district hospitals from the treatment arm and only one hospital from the control arm had full efficiency scores (50% and 16%, respectively). The findings also showed that out of 92 BHCs in the treatment arm, twenty fully efficient compare to 12 out of 92 BHCs in the control arm (22% and 13%, respectively). The quality of health services and the number of health facilities located in urban areas were significantly positive correlated with health facility technical efficiency scores, while the RBF impact was not statistically significant on technical efficiency of health facilities with 5 per cent level of significant.