To the Editor I found the report in a recent issue of JAMA Internal Medicine by Schwartz et al1 on changes in low-value services in the first year of the Pioneer Accountable Care Organization (ACO) program interesting yet frustrating. It begs the question of whether the modest differential reductions in low-value services seen among the Pioneer ACOs included in the study vs the control group was specific to low-value services or simply reflective of a broad-based, nonspecific reduction in services that might have undesirable consequences. Just as the Rand Health Insurance Experiment,2 decades ago, showed that higher coinsurance payments led to across-the-board reductions in use that affected more and less discretionary care equally, nonspecific service reduction is a plausible outcome of Pioneer ACO participation and needs to be evaluated because ACOs are supposed to improve the value of care provided, not lower costs by indiscriminate service reduction. This needs careful assessment before the findings of Schwartz et al can be interpreted as potentially indicating success of some of the Pioneer ACO participants in successfully reducing low-value care.
Saver BG. Discretionary Interpretations of Accountable Care Organization Data. JAMA Intern Med. 2016;176(3):411-412. doi:10.1001/jamainternmed.2015.8503