Physicians in peer groups prescribe fewer opioids than physicians in solo practice

In their paper titled “The effect of organizations on physician prescribing: the case of opioids” (National Bureau of Economic Research, Working Paper 31785, October 2023), M. Kate Bundorf, Daniel Kessler, and Sahil Lalwani examine how peer effects influenced opioid-prescribing behavior among physicians. They use data of Medicare recipients in 2014 and 2018. First, they look at whether a physician worked solo or belonged to a physician group. Next, the authors look at the size of the group (large, medium, or small). They then examine the effects that membership to these groups had on opioid-prescribing behaviors, looking at not only the different opioid-prescribing behaviors (measured in how many days of opioids a prescription is allotted for a patient) of solo physicians and physicians belonging to a group but also those behaviors of physicians who practiced solo in 2014 and then became integrated into a group practice in 2018. And finally, the authors investigate how the average age of the other physicians in the group affects the opioid-prescribing behaviors of physicians in that group.

The authors find that group physicians prescribed fewer opioids than did physicians in solo practice. They also learn that solo physicians prescribed opioids more intensively and less appropriately than group physicians. To measure appropriateness of prescribing, Bundorf, Kessler, and Lalwani overlapped the prescriptions of different opioids. When comparing all physicians in solo practice in the study’s sample, the authors learn that the average length of an opioid prescription was 25.11 days. Physicians who switched from a solo practice in 2014 to a group practice of any size in 2018 reduced opioid prescriptions by 3.35 to 4.63 days (13.3 to 18.4 percent). However, the authors find that 14.4 percent of physicians in their sample did not prescribe opioids at all. Of the 85.6 percent that prescribed an opioid, the average length of an opioid prescription by a solo physician in 2014 was 27.44 days. Of these physicians, those who switched from solo practice in 2014 to an integrated group of any size in 2018 reduced their prescribing by 3.36 to 5.24 days per prescription (12.2 to 19.1 percent).

The authors show, too, that physicians who switched from solo to a group practice of any size not only reduced prescription days but also prescribed fewer overlapping prescriptions by 2.7 to 3.8 percentage points (14.3 to 20.1 percent) than did physicians who continued in a solo practice. In contrast, physicians who switched from a group of any size in 2014 to solo practice in 2018 increased their overlapping prescribing by 1.8 to 2.7 percentage points more than physicians who remained in a solo practice.

In addition to looking at physicians who belonged to peer groups, the authors look at groups of physicians who were integrated under a hospital or hospital system. The authors find that although hospital-employed group physicians prescribed fewer opioids than solo physicians, the authors determine that this reduced prescribing may be related to factors other than peer effects.

Another key finding that the authors note is regarding the average age of physician group members. Older physicians prescribed more opioids than did younger physicians. Physician groups with an average age of between 30 and 39 years old prescribed 3.66 fewer days of opioids than physicians in groups with an average age of 65 or older.

The authors point out that before they completed their paper, not much research had been done on the effects of peer pressure on opioid-prescribing behavior. These results present an “opportunity to reduce opioid prescribing in general and inappropriate opioid prescribing in particular.”