EHRs produce too many false positives in medication adherence program
Researchers, led by Thomas S. Rector, PharmD, PhD, investigator at the Center for Chronic Disease Outcomes Research for the VA, developed an automated system to perform biweekly reviews of electronic VA medication records.
From February to July of 2010, they used this system to determine how many of 21,935 statin recipients were failing to adhere to this specific medication. The system identified 1,000 patients as past due for a statin resupply and researchers sent these patients surveys to determine how many had been accurately identified.
Out of the 1,000 patients, 824 were accurately identified for discontinuation. The other 176 had been inaccurately identified and had adhered to their medication. Out of the 824 accurately identified, 95 were deceased, 17 were long-term care residents, 302 reported good reasons for stopping statin treatment and 208 eventually got another supply.
Based on the low yield of patients accurately identified, researchers concluded that it is not currently viable to use EHRs to screen patients who may be appropriate candidates for efforts to increase medication adherence.
“The actionable yield of this particular effort to screen the VA’s EHRs of patients that were filling prescriptions for a long-term medication was quite low, due to that was not placed in a timely manner or that would be difficult to extract without a much more sophisticated program including text searches,” Rector concluded. “Furthermore, supplemental information is necessary to determine the patients’ undocumented reasons for discontinuing a medication.”