Authored by: Gil Shapira, Tashrik Ahmed, Aneesa Arur, Damien de Walque, Kate Mandevill
The global knowledge on performance-based financing (PBF), an approach rewarding both quantity and quality of health services, has been expanding in recent years. Completed impact evaluations, including those funded by the Health Results Innovation Trust Fund, measured how PBF pilots in different developing countries affected quality and coverage of maternal and child health services. This global evidence base, however, is overwhelmingly based on studies form sub-Saharan Africa. In a recent working paper on the PBF pilot in Tajikistan, first results from a Central Asian country with a post-Soviet health system are presented.
In many ways, the design of the PBF pilot in Tajikistan is similar to the PBF models implemented in sub-Saharan Africa. Rural primary health facilities receive quarterly payments depending on the number of provided targeted services and on a general quality score. Up to 70% of the financial incentives can be distributed to facility staff while a minimum of 30% must be reinvested in the facilities. There is a multi-layered system of supervision, verification and counter-verification of performance indicators.
The PBF pilot in Tajikistan, however, is different from the typical African PBF intervention in several notable ways. First, performance incentives are provided solely at primary level facility that do not offer inpatient services such as labor and delivery. Second, in addition to a package of reproductive, maternal and child health services, hypertension-related services were also incentivized. Third, prior to the launch of the pilot, health providers were retrained in family medicine. More generally, the political, economic and epidemiological contexts in Tajikistan are different. For example, the baseline health services coverage rates in Tajikistan were higher than those reported in most African studies.
The evaluation relies on a difference-in-difference design, which compares changes in outcomes between the districts implementing PBF and control districts within the same regions. Coverage rates were measured at the population level by household surveys conducted before the launch of the program in 2015 and after three years of implementation. Detailed data on quality of care were collected through extensive facility-based surveys that included general facility assessments, interviews with health providers, and direct clinical observations of patient-provider interactions.
The results show that the PBF reform had a positive impact on many dimensions of quality of care in the rural health centers. The availability of equipment and supplies increased. There were significant positive impacts on indicators related to infrastructure and infection prevention and control standards, such as the presence of containers for sharps and needles in consultation rooms. Positive impacts are also found on provider competency, measured through clinical vignettes. Finally, there is evidence that the improvements in structural quality and provider knowledge also translated into better content of care. For example, providers in the PBF facilities were more likely to perform key physical exams such as measuring the height and weight of children under 5.
While the quality of the facilities improved, and the communities reported to notice these changes, there are only modest impacts on the utilization of health services. Statistically significant impacts are found on the rate of adults aged 40 and above that had their blood pressure measured in the preceding year, and on the rate of women who received timely postnatal care. However, no statistically significant impacts are found on timing and number of antenatal consultations, use of family planning methods, or coverage rates of child growth monitoring and vaccination.
The more modest impacts on health care utilization might be explained by several factors. One explanation may be that behavioral change might take time to materialize. Although we did not find increased utilization of many of the targeted services, the population in the PBF districts reported higher satisfaction with the local primary care facilities. It could be that with longer implementation, utilization of targeted services would have increased. It is also important to note that for some of the indicators, the overall coverage was already high in baseline. For example, close to 90% of aged 12-23 months received all basic vaccinations. With respect to some indicators, we find overall positive trends even if we do not find impact of the PBF pilot when we compare with the control districts. Between the baseline and follow-up surveys, the rate of women in the control districts that initiated antenatal care during the first trimester increased from 57% to 74%. The relative high baseline coverage rates and the overall positive trends might have limited the scope of the supply-side incentives to impact utilization.
How do these results compare to those from other settings? Despite the differences in the contexts in which the programs were implemented and the differences in design, the results from Tajikistan are overall in line with those from other countries. We find positive impacts on a range of measures of quality of care. Like the studies conducted in Rwanda, Afghanistan and Zambia, this study also shows that PBF can be effective not only in improving structural quality (e.g. infrastructure and equipment) but can also have positive impact on the content of care. Consistent with most of the other studies, however, we only find impacts on utilization of a few out of all incentivized services.
To improve health outcomes, health system reforms should improve effective coverage - coverage of high-quality services. Supply-side incentives in the form of PBF can increase quality of care and therefore result in more effective coverage. However, the results in Tajikistan and elsewhere suggest that the improvements achieved by these programs are not sufficient to relieve some of the demand-side barriers. Therefore, progress on effective coverage is likely to require that PBF programs are introduced together with interventions that better address these barriers. In Rwanda, for example, demand-side in-kind incentives were effective in improving rates of timely antenatal and postnatal care even though health facilities were already incentivized to improve these indicators through a PBF program. To promote health seeking behavior, several countries have been piloting PBF programs that include incentives for health providers to conduct home visits or incentives for community health workers. There is a need to pilot and evaluate more programs that combine supply-side and demand-side interventions in order to better understand the complementarities and synergies these interventions might have in achieving better population health outcomes.