Ricki Lewis, PhD
March 30, 2018
Most emergency department (ED) physicians underestimated the number of opioid prescriptions they wrote, but began prescribing fewer upon learning of their behavior, according to results of a study published in Academic Emergency Medicine.
Most ED physicians lack data on their opioid prescribing habits, and there are few benchmarks that would allow them to compare their habits with an accepted norm. In fact, only 17 states have published guidelines for opioid prescribing in the ED, and most that do rely on a physician's knowledge of the opioid crisis and common sense to minimize prescribing rather than assess the behavior of individual clinicians.
In this prospective, multicenter, randomized trial, Sean S. Michael, MD, MBA, from the department of emergency medicine, University of Colorado School of Medicine, in Aurora, and colleagues investigated how providers perceive their opioid prescribing habits and assessed clinicians' response to learning how their actual prescribing practices compared with group tendencies. The participants were then monitored for self-correction over the course of a year.
The goal of the strategy, which the researchers term "query-reveal intervention," was to personalize how practitioners view the opioid crisis in their own workplaces. The researchers surveyed attending physicians, residents, and advanced practice providers (APPs) at four hospital EDs: an urban tertiary academic center, an urban acute care hospital/nonprimary teaching site, a small suburban community hospital, and a small rural community hospital.
The investigators randomly assigned 51 clinicians (34 attendings, 15 residents, and 2 APPs) to the intervention arm and 58 (31 attendings, 21 residents, and 6 APPs) to the control arm.
To estimate degree of opioid prescribing, a member of the study team showed each participant in the intervention group bar graphs representing all providers at their ED and asking "Each of these bars represents one provider in the group, including you. Which do you think is you?" Immediately after that, each clinician was provided with his/her true prescribing profile in absolute and relative to peers. Physicians in the control arm did not receive information on individual or group prescribing habits.
The primary outcome was change in percentage of patients discharged with an opioid prescription at 6 and 12 months.
During the year of the study, the 109 practitioners altogether discharged 119,428 patients and wrote 75,203 total prescriptions, including 15,124 (20.1%) for opioids.
Of the participants receiving the intervention, 73% of the attending physicians and advanced practice providers and 27% of the residents underestimated their prescribing rank compared with their peers by more than one decile in one metric or more. Just five providers (three residents and two attendings) overestimated their opioid prescribing rank.
Clinicians in the intervention group who had underestimated their prescribing showed larger declines in the number of patients discharged with an opioid prescription compared with clinicians in the control group. Specifically, compared with the control group, they had a median of 2.1 (95% confidence interval [CI], 0.5 - 3.9; P = .007) fewer opioid prescriptions per 100 patients at 6 months and 2.2 (95% CI, 0.01 - 4.8; P = .05) fewer prescriptions per 100 patients at 12 months.
Meanwhile, those in the intervention group who had not underestimated their prescribing at baseline prescribed 1.3 more prescriptions per 100 patients at 6 months compared with controls, and 1.2 more per 100 patients at 12 months.
"In the multivariable mixed-effects model, intervention allocation with underestimation of one's prescribing relative to peers, physician level of training (compared to APP), and fewer years of experience were significant predictors of larger-magnitude six and twelve-month decreases in the proportion of patients discharged with an opioid prescription," the authors write.
Providers with accurate self-perception or who overestimated their opioid prescribing practices didn't decrease opioid prescribing compared with controls.
The four EDs served different communities but were part of the same healthcare system. Prescribing behavior did not significantly differ among them.
The researchers compare the lack of self-awareness about opioid prescribing to the observation that a majority of drivers think they are above average. They suggest that the disconnect between perceived behavior and being confronted with a different reality creates cognitive dissonance that drives behavior change.
"This randomized trial exposes important gaps in providers' self-perceptions of opioid prescribing and demonstrates that a simple, data-driven intervention using query-reveal methodology may decrease future prescribing, particularly among providers who underestimate their own prescribing practices," the researchers conclude.
Limitations of the study include the fact that only 5% to 10% of opioid prescriptions come from EDs, lack of information on the indications for which opioids were prescribed or whether alternative treatments were suggested, and implementation of a state law limiting opioid prescription during the eighth month of the investigation.
The researchers have disclosed no relevant financial relationships.