Hospital Payment Incentives and Opioid Prescribing

Over 216,000 individuals have died from overdoses related to opioid pain relievers between 1999-2015, making this the deadliest drug-related epidemic in U.S. history. The opioid crisis has in part been fueled by the prescribing behavior of physicians, with the amount of opioids prescribed per person in 2015 being three times higher (640 MME – morphine milligram equivalent) than in 1999. Prescribing of opioids within the hospital setting and at time of discharge may be an important, albeit understudied, pathway through which new initiates of opioid therapy may develop longer term use and dependence. Some 56 million patients receive opioid medications following surgery each year, and persistent opioid use after surgery among previously opioid-naïve patients has become increasingly common. To address the opioid epidemic, it is imperative that all potential factors be examined and studied. There is growing concern that one of these may be provider payment incentives. Under the Affordable Care Act, the Centers for Medicare and Medicaid Services (CMS) implemented the Hospital Value-Based Purchasing Program (HVBP), directly tying incentive payments to Medicare patients’ experience of care, of which pain management is a component. Many have raised the concern that providers may be prescribing opioids unnecessarily for the express purpose of improving their patient experience metrics, concerns which were voiced by Senators Dianne Feinstein and Charles Grassley in their letter urging CMS to consider, study, and address this link between patient satisfaction-based reimbursement and opioid prescribing and also voiced by 26 senators in a similar missive to the Secretary of Health and Human Services. CMS acknowledged the lack of a rigorous evidence base on this issue, a point recently echoed in the July 2017 interim report of the White House opioid commission that underscored the need for a thorough examination of the role played by patient satisfaction and pain level as criteria through which providers are evaluated and paid.

The proposed research will address this knowledge gap and provide some of the first empirical evidence on whether, and to what extent, linking hospital reimbursement to patient experience of care metrics affects opioid prescribing during inpatient care and post-discharge, and how it affects longer-term opioid use. The project will link Medicare Part A claims and Medicare Part D event data, and exploit the natural experiment provided by the introduction of Medicare’s HVBP in October 2012 to most, but not all hospitals, to estimate plausibly causal effects. The project will also exploit the removal of the pain management questions from HVBP starting in FY2018 to further disentangle these effects. With the private sector following Medicare’s suit in giving patient experience of care a larger role in calculating payments, the effects can propagate beyond Medicare, and the proposed study has broader implications in understanding any potential unintended consequences of tying patient experience measures to payment incentives.