Academic Detailing to Enhance Adoption of Clinical Decision Support for Patients at Risk of Opioid Overdose.

BACKGROUND

Not all patients at risk for opioid overdose are prescribed naloxone when discharged from the emergency department or hospital. Clinical decision support (CDS) can be used to promote clinical best practices, such as naloxone prescribing; however, it may be ignored due to knowledge deficiencies or alert fatigue.

OBJECTIVES

Assess the effect of academic detailing on responses to a CDS alert recommending a naloxone prescription for patients at risk of opioid overdose.

METHODS

A pre/postquality improvement study of 2,161 active providers at a 400-bed academic medical center. The first intervention was an educational email to all providers. The second intervention was individual emails to 150 providers who infrequently ordered naloxone in response to the alert. The main outcome measure was prescription-to-alert ratios, defined as the number of naloxone prescriptions signed in response to the alert divided by the number of times the alert fired.

RESULTS

The first academic detailing intervention resulted in a prescription-to-alert ratio increase from 32.6 to 51.7%, a 19.1% absolute increase when comparing the approximately 8 months before and after the email was sent (95% confidence interval [CI]: 16.3-21.9%,  < 0.001). The second intervention resulted in an increased prescription-to-alert ratio from 9.3 to 50.6%, an absolute increase of 41.3% when comparing the nearly 8 months before and after the emails were sent (95% CI: 36.9-45.7%,  < 0.001). Improvements were seen across all services and all provider roles, particularly for advanced practice providers, and were sustained for 8 months.

CONCLUSION

Academic detailing can be used to augment responses to CDS for patients with opioid dependence. Further study is needed to see if this effect can be replicated with CDS for other high priority conditions, and whether academic detailing with one alert might improve responses to other alerts as well, potentially decreasing alert fatigue.