by Paul Govern
Statins are drugs used to treat high cholesterol and thereby help prevent cardiovascular disease. According to guidelines, most patients with atherosclerotic cardiovascular disease, or ASCVD, should take high intensity statins (as opposed to low or medium intensity statins) but many who should aren’t receiving these drugs, studies have shown.
In a randomized trial at Veterans Health Administration clinics in Houston and Nashville, automated targeted reminders for clinicians helped increase prescribing of high intensity statins for patients with various ASCVD diagnoses, including coronary or peripheral artery disease and ischemic stroke.
In the trial’s intervention arm, which included 18,427 patients with ASCVD, for every 10 patients whose primary care doctors received a targeted reminder, high intensity statins were initiated for one patient.
The trial, reported in Circulation, was led by researchers at Baylor College of Medicine in Houston, with key collaboration from a biomedical informatics team at Vanderbilt University Medical Center.
In all, the project involved 36,641 patients with ASCVD and 245 clinicians in 27 primary care clinics.
“It was exciting to see that developing a personalized prior care summary to recommend a high intensity statin among ASCVD patients resulted in significantly improved receipt of this medication,” said Michael Matheny, MD, MS, MPH, professor of Biomedical Informatics, who led part of the informatics work at VUMC. “Projected to the health care system as a whole, the improvement in high intensity statin prescribing realized in this study might pose considerable benefits in terms of ASCVD-associated morbidity, mortality and costs.”
The team leveraged findings from their previous studies to develop a computer program to automatically generate clinician reminders tailored to each patient with ASCVD, including date of ASCVD diagnosis, history of statin use, history of statin associated side effects, and links to guidelines on high intensity statins and management of side effects.
The study concluded last November after running for 16 months, with clinicians in the intervention arm having received targeted reminders concerning 4,532 patients.
Over the course of the study, use of high intensity statins increased from 53.6% to 55.2% among patients in the intervention arm, while dropping from 55.9% to 53.7% in the usual-care arm.
“The most exciting part of this study was to see the use of machine learning and natural language processing, along with structured EHR data, generate a personalized care summary of the patient to deliver to the provider as a note in the chart,” Matheny said. “This then could reduce the cognitive burden of clinical review to determine how they wanted to address the recommendation. Although this still generated alert fatigue, with almost 33% of providers eventually opting out of receiving reminders, the impact was substantial.”
The study was led by Salim Virani, MD, PhD, with appointments at Baylor and The Aga Khan University in Karachi, Pakistan. The automated reminders were engineered by Matheny and Dax Westerman, MS, principal application developer at VUMC’s Center for Improving the Public’s Health through Informatics. Glenn Gobbel, DVM, PhD, research assistant professor of Biomedical Informatics at VUMC, contributed a natural language processing component that aided surveillance of statin side effects.
The study was supported by the Department of Veterans Affairs.