Abstract
Background/Aims: Harvard Pilgrim Health Care (HPHC) is one of 28 Attorney General Consumer and Prescriber Grant Program grantees. Four HMORN CERTs sites, including HPHC, received grants. An overall goal of the Program is to identify best practices for educating clinicians about the influences on their prescribing and for changing their prescribing behaviors. At HPHC, our aim was to test the effectiveness of computerized clinical decision support (alerts), alone and in combination with group academic detailing, to reduce the use of heavily marketed hypnotic medications at Harvard Vanguard Medical Associates (HVMA), a 14-site group practice that uses the Epic electronic health record.
Methods: We randomly allocated the 14 practices to three study groups:
Alerts Only (n=5 practices),
Alerts plus Detailing (n=5), and
Usual Care (n=4).
Internal medicine clinicians in both the Alerts groups received prescribing alerts when initiating new prescriptions for eszopiclone (Lunesta), zaleplon (Sonata), ramelteon (Rozerem), and controlled-release zolpidem (Ambien CR) between March and December 2007. We carried out group academic detailing sessions in June and July 2007 followed by an educational mailing in the fall.
Results: A pre-planned interim data analysis (August 2007) demonstrated that hypnotic study alerts were triggered on 98 separate occasions. After receiving an alert, 30% (95% CI, 21%–40%) of clinicians changed the prescription to an alternative recommended medication. We plan to extract post-intervention data in January 2008 and will present findings of the intervention at the HMORN meeting. In addition, we carried out pre-intervention and post-intervention surveys of all clinicians at HVMA in 2006 and 2007 which assessed clinicians’ perceptions of the influences on their prescribing, with special attention to the role of direct-to-consumer marketing. We will present the results of these surveys, as well.
Conclusions: Developing interventions using health information technology and group detailing to improve the use of medications is a high priority for many HMORN organizations. The results of this study should be generalizable to many HMORN sites, including but not limited to the many organizations that have deployed the Epic electronic health record.
- Received September 11, 2008.
