Research spotlight: Evaluating Hospital Quality

Measures of hospital quality that matter: evidence from natural experiments

Researcher: Joseph Doyle

With increasing calls to “pay for quality” rather than for volume to transform economic incentives in healthcare, a key question is whether we can measure quality. Existing quality scores are controversial: those who score poorly point to the fact that they treat different types of patients that tend to have worse health outcomes.  In this project, Joseph Doyle and co-authors have tested whether existing quality measures, including hospital outcomes, hospital inputs in the form of process improvements, and patient satisfaction, are related to patient health.  To overcome concerns over patient selection, the research team used a “natural experiment”:  the random assignment of patients transported by 2,500 ambulance companies that shift the choice of hospital where patients are treated (across roughly 3,000 hospitals in their data). The three categories of quality measures did distinguish quality of hospitals when patients visited hospitals in a way that was effectively random, akin to a clinical trial.  The strongest predictor of health outcomes turns out to be the mortality rate of the hospital over the prior three years; going from a low- to a high-scoring hospital results in a 10–15 percent reduction in mortality and hospital readmissions.

1. “Uncovering Waste in US Healthcare: Evidence from Ambulance Referral Patterns.” Doyle, Joseph J., John A. Graves, and Jonathan Gruber. Journal of Health Economics Vol. 54 (2017): 25-39. 
2. “Evaluating Measures of Hospital Quality.” Doyle, Joseph J., John A. Graves, and Jonathan Gruber. The Review of Economics and Statistics. Vol. 101, No. 5 (2019): 841-852.