Using EHRs + snail mail nudges up statin prescriptions

Using EHRs and letters mailed to at-risk patients, physicians doubled the proportion of patients who received new lipid-lowering drug prescriptions. However, study results failed to show an improvement in levels of low density lipoprotein (LDL) cholesterol at nine months and only modest improvements at 18 months.

Results from the cluster-randomized trial were published online Nov. 11 in the Journal of General Internal Medicine.

Stephen D. Persell, MD, MPH, of the Feinberg School of Medicine at Northwestern University in Chicago, and colleagues designed the study to test if providing individualized information to patients at high to moderate risk of cardiovascular disease would increase the frequency of them being prescribed recommended treatments. They also wanted to explore the effect of the intervention on outcomes, particularly whether the approach would lead to lowering of LDL cholesterol levels in at-risk patients.    

The study enrolled 29 primary care physicians (PCPs) in one large practice that had used the same EHR for more than 13 years. Physicians were randomized to either an intervention or control group. Study staff mailed risk messages directly to intervention physicians’ patients whom the EHR identified as high to moderate risk of cardiovascular disease between February and November 2011.

To be included in the study, patients had be in the care of a physician in the study; be between 40 and 79 years old; not prescribed lipid-lowering medications; have no diagnosis of coronary heart disease, heart failure, stroke, diabetes or peripheral artery disease; and a combination of LDL cholesterol test results and Framingham Risk Scores that showed them at high to moderate 10-year risk for a cardiovascular event. There were 218 patients seen by 14 physicians in the intervention group and 217 patients and 15 physicians in the control group.

Patients in the intervention group received a letter addressed from their PCP with personal cardiovascular disease information, an estimate of the reduction in risk if they used a statin and encouragement to discuss options with their doctor. The primary outcome was a repeat LDL cholesterol level of at least 30 mg/dl below the baseline level. Secondary outcomes included prescription of lipid-lowering drugs.

Persell and colleagues collected data through the EHR and performed analyses at nine months. They also conducted post hoc analyses on data collected within 18 months after the launch of the intervention.

At nine months, they found no difference between the intervention and control patients in the primary outcome. But 11.9 percent of the invention group patients received lipid-lowering drug prescriptions compared with 6 percent of the control group patients. At 18 months, EHR data showed 22.5 percent of the intervention group patients’ LDL cholesterol had been lowered by 30 mg/dl or more compared with 16.1 percent in the control group. At 18 months, the percentage of patients to receive prescriptions remained modest, at 17.4 percent for the intervention group and 11.1 percent for the control group.

“Even though we eventually observed differences in cholesterol lowering, the intervention effect was small,” Persell and colleagues wrote. But they also observed that the intervention did not demand many resources.  

“If a measurable positive effect of a more limited intervention like ours can be accomplished at a sufficiently low cost, it may be a worthwhile approach to adopt,” they wrote. “If the process of delivering these messages can be fully automated (such as using a patient portal connected to the EHR), this may reduce the cost of providing this service further.”

The authors pointed out that the study involved only one center and may not be replicable in other settings. They offered possible reasons for the modest benefits seen in the study, including that the message was sent only once and patients may have perceived their risk as low.

Their study showed that their approach is feasible, they concluded, but added that repeated personalized messaging may be more effective.