Taking Care: The Impacts and Limitations of Financial Incentives in Health

This piece was first published on World Bank’s Let’s Talk Development blog.

A frustrating healthcare situation exists in many low-income countries: deaths are often due to poor quality care rather than the lack of access to care. Despite sustained investments in health service delivery and increases in the use of health services health outcomes have remained poor. This frustrating state of affairs shapes the big-picture motivation behind our new World Bank Policy Research Report, Improving Effective Coverage in Health: Do Financial Incentives Work? The report draws on research spanning 15 years and 40 countries to better understand the gap between health coverage and outcomes and points a way forward.

Healthcare Financing Experiences a Transformation

Nearly two decades ago, many low- and middle-income countries (LMICs) began a transformation in healthcare financing. Centralized budgets and low levels of autonomy for facilities and workers were replaced with frontline autonomy, flexible operating budgets, performance pay, and greater transparency and accountability for results. This package of reforms came to be known as performance-based financing (PBF). PBF, which includes performance pay, among other critical features, appealed to development agencies and donors because of its explicit links to transparency and accountability.

Since the late 2000s, more than US$2.5 billion has been invested in PBF projects in primary health service delivery in low-income countries—a significant departure from previous financing models, which didn’t link financing to outcomes and results. Unfortunately, although health coverage did expand dramatically over the past two decades, this expansion was not accompanied by the expected improvements in health outcomes.

Understanding data through metrics

To understand the drivers behind the frustrating situation, we used a metric called effective health coverage. The metric measures primary health coverage with a minimum content and quality of care. The data show that the gap between coverage and effective coverage can be extreme in some cases. For example, coverage for prenatal care in several sub-Saharan African countries is high, but the quality is often lacking (Figure 1).

Figure 1: Coverage versus Quality of Antenatal Care

Note: Coverage: percent of women giving birth who had 1+ antenatal care visits. Quality: of those covered, the percent who had 4+ visits, 1+ visits with a skilled provider, blood pressure taken, and blood and urine samples taken (correct treatment). EC = effective coverage. The isoquant lines demarcate comparable levels of effective coverage. See page 21 in the report for more details.

In many LMICs the disparity between coverage and quality means that maternal deaths remain unacceptably high, despite widespread access to prenatal care. A study of maternal deaths in 137 LMICs found that out of an estimated 207,000 excess deaths in 2016, 104,000 patients had sought health care, but from that number, 57,000 patient deaths were likely due to poor quality of care, and an estimated 47,000 deaths that were due to a lack of access to care. This pattern holds not only for maternal health but many other health services as well. For this reason, understanding the drivers of the gap between quantity and quality of care is crucial to figure out why more health services don’t necessarily translate into improved health outcomes.

The “know-can-do” framework

The first step in improving quality is understanding what the constraints are. To identify the relative importance of various constraints, the report relies on a “know-can-do” framework that examines structural gaps (do healthcare facilities have drugs and equipment in stock?); knowledge gaps (have healthcare workers received adequate training on using the correct treatment in a specific situation?); and gaps in worker effort.

The report found that two-thirds of poor-quality care cannot be attributed to low worker effort, suggesting that performance pay alone would have a limited impact. This is not surprising given the high percentage of health facilities that lack even basic supplies (Figure 2). This finding suggests that effective policies need to address not only worker effort but other factors like structural capacities and medical training.

Figure 2: Structural Capacity Gaps

Is performance pay a silver bullet?

The rollout of PBF reforms was intended to address the types of poor outcomes such as those described above. However, the report details that while PBF achieved better outcomes than business-as-usual in low-income settings, performance pay to incentivize worker effort, in particular, proved to only marginally improve outcomes while being costly to implement and monitor. In fact, in one case, verifying performance pay alone consumed two-thirds of project administrative costs.  

While financial incentives for health workers can increase coverage, these incentives are marginal – meaning that closing large gaps in effective coverage remains an elusive goal. Performance pay might make sense in decentralized, high-quality health systems that already support facility financing and autonomy, as well as accountability and transparency. In reality, its potential may be more limited in centralized, under-resourced health systems with critical gaps. Further, the report also provides cautionary primary evidence that performance pay can incentivize the provision of inappropriate, unnecessary, or irrelevant care.

Pivoting to impactful health financing reform

PBF projects produced gains in health outcomes compared to business-as-usual, but these gains did not necessarily result from the specific financial incentives and associated monitoring components of projects. Instead, impactful health financing reform may mean pivoting from performance pay while retaining other essential aspects of PBF projects—like transparency, accountability, and direct frontline financing. We recommend that policymakers assess not only coverage but also effective coverage to ensure they address the most critical gaps in care. Before introducing performance pay as part of a reform package, policymakers should identify whether the constraints in a healthcare system are within the control of frontline healthcare workers or lie elsewhere. Finally, the sequencing of interventions is key.  Rather than starting with across-the-board performance pay, we recommend 18 – 24 months of decentralized facility financing and demand-side interventions, followed by a diagnostics exercise to assess progress before introducing performance pay for indicators with sizable gaps in effective coverage.

Lastly, the measurement of healthcare quality and efficiency should be built into health systems reform. This will allow local policymakers to ensure that the context and time are right and that the constraints to quality can be addressed by performance pay. If not, scarce resources will continue to be directed inefficiently.