Tailored Alerts Up High-Intensity Statin Prescribing: PCDS Statin

The informatics-guided alerts helped optimize statin use in VA clinics, but notable, too, was the signal of alert fatigue.

Tailored Alerts Up High-Intensity Statin Prescribing: PCDS Statin

Email reminders to physicians about proper dosing of high-intensity statins, individualized to specific patients and often timed around their appointments, significantly improved the likelihood of these medications being given and adhered to within the Veteran’s Affairs (VA) Health System over a 15-month period.

For every 10 reminders sent, one additional patient would end up being prescribed the correct statin dose, investigators for the Personalized Clinical Decision Support (PCDS) Statin trial concluded. By contrast, for patients treated at the clinics not randomized to these personalized physician reminders, high-intensity statin use actually declined.

“We believe that our study results inform how informatics driven interventions can improve evidence-based care delivery in large health care systems,” said Salim S. Virani, MD, PhD (Texas Heart Institute, Houston, TX), who presented the PCDS Statin results at the American College of Cardiology/World Congress of Cardiology (ACC/WCC) 2023 meeting.

His team spent 4 years doing qualitative work to design the optimal email reminder that would identify patients who would be eligible for higher-intensity statin prescriptions and would motivate clinicians to improve their statin prescribing while seamlessly fitting into work flows and physician preferences.

Even so, 31.6% of the clinicians who were randomized to the tailored intervention opted out of the study. “This could be related to competing demands on their time due to iterative waves of COVID-19 that occurred during the conduct of the study, as well as alert fatigue,” Virani said.

While high-intensity statin therapy is supported by guidelines for patients with atherosclerotic cardiovascular disease (ASCVD), many patients don’t end up being prescribed these medicines for reasons related to statin-associated side effects or other clinician logistical issues or beliefs. Mobile apps and other electronic health record interventions have shown improvements in both prescribing and adherence, but more-comprehensive solutions are still needed.

Still, the intervention was a success, statistically speaking, especially among the 53% of eligible patients whose physicians ended up receiving a reminder over the course of the study.

High-intensity statins are “one of the most important things we have in cardiovascular medicine,” commented Karol Watson, MD, PhD (UCLA Health, Los Angeles, CA), following Virani’s presentation. “We know what works. We know how to get it. We just don't know how to get it implemented in all the patients that need it.”

Intervention Performance

For the study, which was simultaneously published in Circulation, Virani and colleagues randomized 27 primary care clinics within the VA Health System to their intervention (14 clinics, 117 clinicians, and 18,427 patients) or usual care (13 clinics, 128 clinicians, and 18,214 patients) between August 2021 and November 2022. Mean patient age was 71.1 years, and 96.5% were men. Ischemic heart disease (77.5%) was the most common form of ASCVD, while 22.1% had PAD and 27.5% had ischemic cerebrovascular disease.

The intervention consisted of monthly data processing, synchronous and asynchronous email reminders about patients not meeting high intensity statin guidelines, and guideline resources on the clinic intranet, while usual care involved a patient dashboard displaying clinician compliance with statin therapy. Of note, the intervention algorithm would not send email reminders to clinicians with more than three open alerts.

Among patients included in the intervention arm, 41.6% were identified to have a signal of statin-associated side effects. A total of 4,928 reminders were sent regarding 4,532 patients throughout the course of the study, representing 53% of patients not on high-intensity statins at baseline in the intervention arm. Of these reminders, 73% were sent asynchronously and 27% were sent within 2 to 7 days of a patient appointment. Over the course of the study, 37 clinicians opted out of receiving these reminders.

Over the 15 months that the study was active, high-intensity statin prescribing dropped by 2.2% in the usual-care arm and rose by 1.6% in the intervention arm for a significant between-group difference of 3.8% (OR 1.06; 95% CI 1.02-1.11). When the analysis was limited to only the 53% of eligible patients for whom reminders were sent, high-intensity statin prescriptions went up even more, to 10.1%.

“What we would say here is the number needed to remind is 10,” Virani said. “That is, 10 reminders need to be sent to clinicians for one patient to be titrated to high-intensity statin therapy.”

There was no change between synchronous and asynchronous reminders (11.6% vs 9.6%; P = 0.58), but there was a slight attenuation of the effect when looking at patients with statin-associated side effects compared with those without (9.1% vs 10.9%; P = 0.02).

The secondary outcome of any statin use went down in both study arms, but significantly more in the usual-care arm compared with the intervention (-5.2% vs -2.4%; OR 1.12; 95% CI 1.06-1.18). However, this decrease was not seen when looking solely at those who did receive an alert.

Lastly, an exploratory analysis looking statin adherence showed that this endpoint went up in both the cohorts, but by 2.8% more in the intervention arm (3.6% vs 6.4%; OR 1.38; 95% CI 1.32-1.45).

‘Repeatable, Scalable Solutions’ Needed

Commenting on the study in a press briefing, David Cho, MD (UCLA Health), said the benefits of the study were its size as well as the intervention’s ability to “centrally process and use different algorithms to not only figure out whether [patients are] on a statin but also [check] for statin-associated side effects being highlighted in the notification.”

However, “the reminder is only one piece of the puzzle,” Cho said, noting that clinician work flows will vary especially outside of the VA system. “I think the study also highlights a very important component about alert fatigue and how many alerts are too much. We could have an alert for everything that's potentially present. So how do you optimize and balance a physician's time with the secondary outcome that you will get for the patient?”

Ultimately, he said, “what we need to find are the repeatable, scalable solutions that can identify the patients with that care gap, create that touch point with the healthcare system, but then actually can trigger an action to address that care gap and close that loop back to the patient.”

Watson pointed out that while the intervention may increase the use of high-intensity statin therapy, there might be some “opportunity cost” to factor in. First, it’s possible that the alerts subsequently decreased use of antiplatelet agents or blood pressure-lowering medications. Also, the “high number” of clinicians opting out likely “relates to alert fatigue,” Watson said.

On the other hand, she continued, “Many clinicians who likely could most benefit from an alert never received it. . . . So while I enjoy any kind of intervention that can help us get more patients on these lifesaving medications, I think there are a number of questions to ask.”

Virani agreed that its entirely valid and worth investigating whether an intervention that increases statin prescribing might have a negative impact on other therapies. Further work will also look into statin adherence issues, especially for unintentional reasons like forgetfulness and not necessarily due to side effects, he added. “It will be important to see what is the long-term impact of this if we follow these patients for another year or so.”

  • The study was supported by the Department of Veterans Affairs.