Web app for self-prescription of statins 96% concordant with physician reviewers
A web application to help individuals self-prescribe statin therapy demonstrated 96% agreement with subsequent physician judgment, according to research published in the Journal of the American College of Cardiology.
The goal of this investigation was to assess the ability of a web-based app to accurately discern patients who would benefit from statin therapy; therefore, improving access to preventive therapies.
“There have been five efforts to bring statins over-the-counter, and they were well intentioned, but they failed pretty badly. None of them received FDA approval,” Steven E. Nissen, MD, MACC, chief academic officer of the Sydell and Arnold Miller Family Heart, Vascular & Thoracic Institute and Lewis and Patricia Dickey Chair in Cardiovascular Medicine at Cleveland Clinic, told Healio. “I publicly opposed each of the prior efforts to take statins over-the-counter. I did so because the studies showed that as many as half of the people who actually self-selected to take over-the-counter statins were ineligible due to their risk being too low or too high, or to have a contrary indication, which of course is risky.
“For this analysis, we mimicked how the web application would work in the real world, and got a 96% accuracy,” said Nissen, who is also a Cardiology Today Editorial Board Member. “If you go to a primary care doctor and that doctor uses their own judgment to decide whether you should get a statin or not, they probably wouldn’t be 96% accurate. This might be as good if not better than one could get from seeing a physician one-on-one. This is a very high level of accuracy, and it ought to settle the question with skeptics. Skeptics, like me, who were not in favor of prior efforts.”
For the study, 500 participants completed an at-home web app that evaluated the appropriateness of treatment with rosuvastatin 5 mg. The web app was programmed using the 2018 cholesterol treatment guidelines from the American Heart Association, American College of Cardiology and 10 other societies on moderate-intensity statin use as well as the warnings and precautions for rosuvastatin. The primary endpoint was the proportion of participants whose self-selected eligibility for nonprescription rosuvastatin was concordant with clinician assessment.
Nissen said rosuvastatin 5 mg was chosen due to its safety and efficacy profile.
Researchers included 83 participants with a Rapid Estimate of Adult Literacy in Medicine (REALM) test score of less than 61, corresponding to a seventh or eighth grade education.
During the self-assessment, participants responded to questions related to medical history, medication use, total cholesterol, HDL, LDL, triglycerides, BP and, if needed, waist circumference, high-sensitivity C-reactive protein and coronary artery calcium score. Based on the participant’s responses, online statin eligibility assessment concluded with one of three possible outcomes: “OK to use,” “not right for you” or “ask a doctor.”
According to the study, respondents were deemed ineligible for therapy with rosuvastatin 5 mg if their assessed risk was less than 5% or greater than 20%; a risk greater than 5% and less than 7.5% without additional risk factors; or if they had a contraindication to rosuvastatin. The results of the web app assessment were not revealed to participants.
Concordance with physician judgement
Nissen and colleagues reported that the web app self-selection for rosuvastatin therapy was concordant with clinician selection in 96.2% of participants (95% CI, 94.1-97.7). Among the cohort, 4.6% were deemed appropriate (“OK to use”) and 91.6% were deemed inappropriate for treatment (“not right for you”).
According to the study, discordance was attributed to incorrect self-selection in three cases, incorrect rejection in 14 cases and an incorrect “ask a doctor” outcome in two cases.
Nissen said the high level of statin ineligibility that was observed was due to the trial’s all-comer enrollment design.
“No program will be 100% accurate. The FDA has said it wants to see 90% or greater accuracy. If you think about other over-the-counter medications, the accuracy is probably not 90%, but, for this group of drugs, we sought to achieve a program that could get to that 90% level of accuracy,” Nissen said. “We got 96%, we were very pleased and we’re going to now go on and do another study of about 1,000 patients to verify that we’re getting the right people on statin therapy.”
‘Before we leap’ to self-directed care
In a related editorial, Neha J. Pagidipati, MD, MPH, assistant professor of medicine at Duke University School of Medicine and member of the Duke Clinical Research Institute, and Eric D. Peterson, MD, MPH, Adelyn and Edmund M. Hoffman Distinguished Chair in Medical Science at University of Texas Southwestern Medical Center, wrote there are factors to consider when looking at technology-assisted self-selection for drug therapies.
“Before we leap into this new world of technology-aided self-directed care, however, we need both optimized digital decision-aids and larger implementation studies. Although Nissen et al found that patients were reasonable at reporting their health histories, the accuracy, completeness and ease of collection of health information could be improved if the tools had been digitally integrated with the participants’ electronic medical record,” Pagidipati and Peterson wrote. “Future implementation studies should also assess for adverse effects resulting from using self-guided patient decision aids (such as overuse, safety concerns, and patient anxiety),” they wrote. “Although such studies will take resources, the potential upside gained from better cardiovascular prevention could be enormous. The time has come for us to learn how to safely empower patients to self-prescribe preventive therapies.”
- Grundy SM, et al. Circulation. 2018;doi:10.1161/CIR.0000000000000625.
- Pagidipati NJ, et al. J Am Coll Cardiol. 2021;doi:10.1016/j.jacc.2021.07.020.
For more information:
Steven E. Nissen, MD, MACC, can be reached at firstname.lastname@example.org.