An electronic health record-based intervention can help combat the opioid epidemic.
Juan Carlos Montoy, MD, PhD
Default electronic health record (EHR) settings could influence the quantity of prescribed opioids.
The findings of a recent study indicated that an EHR-based intervention could have implications for opioid prescribing and could be used to help combat the opioid epidemic.
Juan Carlos Montoy, MD, PhD, and a team of investigators sought to determine whether and the extend that changes in default settings in the EHR were associated with opioid prescriptions for patients leaving the emergency department. The team collected data and found that a lower EHR default setting was associated with a lower number of prescribed pills.
Montoy, from the Department of Emergency Medicine at University of California, San Francisco, and colleagues changed default quantities for opioids from typical quantities of 12 or 20 pills to either none, 5, 10, or 15.The provider ultimately decided who to prescribe the opioids to and could modify the quantity prescribed without being restricted.
The investigators’ primary outcome was the number of opioid-containing tablets were prescribed under each EHR default setting. Additional measures included the distribution of quantity prescribed, the proportion of prescriptions for > 12 tablets under each default setting, and the amount of prescriptions written for the default quantity.
Over the course of 5, four-week time frames, Montoy and the team altered the dispense quantities for discharge prescriptions for commonly prescribed opioids. Emergency department providers were not made aware of the changes, or even that the study was being conducted. The sites for the study included University of California, San Francisco, Medical Center, an academic, urban, tertiary-care center, and Highland Hospital, an urban, level 1 trauma center and safety-net teaching hospitals.
The team randomly assigned default dispense quantities—null, 5, 10, and 15—to four-week periods for each opioid from November 28, 2016—July 9, 2017. For the null quantity, the provider needed to enter a dispense quantity into the EHR.
The study opioids were oxycodone hydrochloride, combined oxycodone and acetaminophen, and combined hydrocodone bitartrate and acetaminophen.
Overall, 4333 prescriptions were issued by 104 emergency department healthcare providers, and the sample size included 4320 prescriptions due to missing quantity data. The mean quantity prescribed in the preintervention period was 14.8 tablets.
A majority of the prescriptions (78.6%) were for combined acetaminophen (325 mg) and hydrocodone (5 mg).
In a regression model, the default quantity was .19 (95% CI, .15—.22)—for each one-tablet increase in default quantity, an increase of .19 tablets was prescribed. In a multivariable model, there was a decrease to .08 (95% CI, .04–.12).
Older patients were given more pills (.01 more per year of age; 95% CI, 0—.03) and women received lower quantities (.93 fewer tablets than men; 95% CI, -1.34 to -.53).
In 8 comparisons of default EHR settings, when the prescription was auto populated with a lower default quantity of opioid tablets, fewer were prescribed. Only in 1 case did a lower default have a higher number of pills prescribed (difference of -1.4; 95% CI, -2.3 to -.4).
The default of 5 yielded lower quantities than 10 (difference of 1.8; 95% CI, .8—2.7), 15 (difference of 1.8; 95% CI, .8–2.9), and 20 (difference of 2.9; 95% XI, 2.1–3.8). The 12-tablet setting had a higher amount of prescriptions (76.4%) for 12 or fewer than any other default.
Auto populating default prescription quantities for opioids influenced the number of pills prescribed to a discharged emergency department patient. The findings indicated that a low-cost, easily implementable EHR nudge could improve prescription practices and maintain prescriber autonomy.
The study, “Association of Default Electronic Medical Record Settings With Health Care Professional Patterns of Opioid Prescribing in Emergency Departments,” was published online in JAMA Internal Medicine.