The Market for Healthcare in Low Income Countries

Joint Applied Micro Seminar
Co-authors: Abhijit Chowdhury, Jishnu Das, Jeffrey Hammer, Reshmaan Hussam, and Aakash Mohpal

Abstract
Patient trust is an important driver of the demand for healthcare. But it may also impact supply: doctors who realize that patients may not trust them may adjust their behavior in response. We assemble a large dataset that assesses clinical performance using standardized-patients (akin to audit studies in economics) in low-income countries to investigate this possibility; most of these data are on healthcare providers who practice in the private sector on a fee-for-service basis. We establish that patients receive low quality of care, with a generous definition suggesting that fewer than 50% of cases are correctly managed, and between 70% and 90% of expenditures are medically unnecessary. Strikingly, and in contrast to literature suggesting that the main problem with fee-for-service provision is over-treatment, the majority of these unnecessary expenditures are incurred because patients are incorrectly rather than over-treated. We then rule-out two plausible explanations for low quality of care: low levels of medical knowledge and low market incentives to invest effort. In our data, there are many healthcare providers who know how to correctly treat the patient and could substantially increase their revenue by doing so given the price-quality gradients we estimate, but still treat the patient incorrectly. A model of the patient-provider relationship in which patients have incomplete information about the quality of providers generates predictions consistent with our findings. The theory additionally suggests that issuing a credible signal of quality should raise average quality of care among providers, even if their underlying ability remains unchanged. We assess this prediction through an evaluation of a highly-publicized training program with informal healthcare providers in West Bengal, India. The program has no impact on knowledge, yet substantially raises quality of care, leading to an increase in the likelihood of correct treatment, a 19% decline in unnecessary expenditures for patients and a 9% increase in revenues for providers. We conclude that low trust undermines clinical performance in an economically and medically significant manner.