Alerts Improve Prescribing of HF Quadruple Therapy: PROMPT-HF

Marlene Busko

April 13, 2022

Physicians treating patients with heart failure and reduced ejection fraction (HFrEF) were more likely to prescribe an added drug class, out of the four guideline-directed medical therapy (GDMT) classes, when they were prompted by an alert in a patient's electronic health record (EHR).  

Clinicians treating more than 1000 outpatients with HFrEF in a large healthcare system were randomized to EHR prompts or no prompts.

At 30 days, more patients with than without an alert in their EHRs (26% vs 19%) were prescribed an additional GDMT class: a beta blocker; a renin-angiotensin-aldosterone system (RAAS) inhibitor, including an angiotensin-converting-enzyme inhibitor (ACEI), an angiotensin-receptor blocker (ARB), or an angiotensin-receptor-neprilysin inhibitor (ARNI); a mineralocorticoid receptor antagonist (MRA); or a sodium-glucose cotransporter-2 inhibitor (SGLT2i).

Tariq Ahmad, MD, MH, presented these findings from the aptly named PROMPT-HF (Cluster-Randomized Pragmatic Trial Aimed at Improving Use of Guideline-Directed Medical Therapy in Outpatients With Heart Failure), in a late-breaking clinical trials session at the American College of Cardiology 2022 Scientific Session held in Washington, DC, and virtually. The study, with lead author Lama Ghazi, MD, PhD, was also simultaneously published online April 3 in the Journal of the American College of Cardiology.

"There are four medications that have been shown to dramatically improve outcomes in heart failure, reduce hospitalizations, make patients live longer," Ahmad, from Yale University School of Medicine, New Haven, Connecticut, said during a press briefing after the session. The HF "guidelines that just came out yesterday show that they have the highest level of recommendation," he noted.

"However, what we find from practice" is that "less than 5% of patients in the best-case scenario are on the right medications, and if you look at the right medications and the right doses, it's almost less than 1%," he said.

SGLT2is are the most underutilized of the four recommended drug classes, with the greatest potential for increased uptake.  

Findings from PROMPT-HF suggest that "this low-cost tool can be rapidly embedded into the EHR at integrated healthcare systems and lead to widespread improvements in the care of heart failure patients," Ahmad concluded.

"Important, Simple"

"This is simple, beautiful," and "it can work in most settings," Craig Beavers, PharmD, University of Kentucky College of Pharmacy, Lexington, the assigned discussant in the press briefing, commented.

"It's exciting to see that you can do embedding with our EHR, and instead of looking at it as a burden, using it as a tool," he said.

"This study is really important," said Mary Norine Walsh, MD, Ascension St. Vincent Heart Center, Indianapolis, the assigned discussant at the session. Therapy inertia is "something we see every day."

"We don't use EPIC," the EHR system used in the trial, Walsh said in an email to theheart.org | Medscape Cardiology. "But many other such prompts are in the works, and I am sure that they will become more common, regardless of the EHR."

She asked Ahmad what might explain the finding that a not insignificant number of patients "went backward," from a higher to a lower number of these four drug classes. The researchers are looking into this, Ahmad replied, adding that the least prescribed drug class, SGLT2is, are extremely well tolerated.

Walsh further wondered if "alert fatigue" by busy clinicians might explain the lack of uptake. Ahmad noted that 79% of alerted providers agreed that the alert was helpful, and anecdotally, several electrophysiologists said the prompts for suggested medications for the specific patient in front of them "walked me through it."

The main study finding "is that a customized, automated system actually can change provider behavior," coauthor F. Perry Wilson, MD, Yale University School of Medicine, and a columnist at Medscape, told theheart.org | Medscape Cardiology in an email.

The next step, he said, "is to see if it can improve clinical outcomes (like hospitalization or even death rates) in patients with HFrEF."

Trial Limitations

"The PROMPT-HF trial demonstrates the effectiveness of EHR-embedded decision support in optimizing GDMT, Harriette G.C. Van Spall, MD, MPH, from McMaster University, Hamilton, Ontario, Canada, and colleagues agree in an accompanying editorial, but they add that "some limitations must be considered."

For example, the study did not report number of clinical encounters, they note. Among the 79% of clinicians who reported the alerts as useful, only 25% implemented the recommendations, 14% reported that patients weren't suitable candidates, and the rest either postponed therapeutic changes (49%) or ignored the alert (12%), representing missed opportunities for improved care.

The proportion of patients with an increase in GDMT classes at 30 days was modest, and less than 8% of patients were prescribed all four classes of agents, they write.

Only 9% of patients in the alert group were newly initiated on SGLT2i, despite a very modest 11% baseline prescription rate, and a vast majority of patients in the two groups did not receive any changes in the number of GDMT classes prescribed, the editorialists point out.

Outpatients With HFreEF

For the study, the researchers randomized the top 100 providers caring for HFrEF patients in outpatient internal medicine and cardiology practices and clinics affiliated with the Yale New Haven Health System in 2020.

Fifty providers with 685 patients and 50 providers with 625 similar patients were randomized to the alert and no-alert groups.

When a clinician opened the order entry of a patient in the alert group, he or she saw the patient's latest left ventricular ejection fraction (LVEF), blood pressure, heart rate, and serum potassium and creatinine levels, which are used when considering medications for patients with HF.  

If a patient was currently prescribed GDMT in the four classes, the specific drugs were displayed. Missing drug classes were highlighted in bold and red, and the EHR provided direct links to available medications in those classes, along with US Food and Drug Administration indications.

The provider could respond a) "I will adjust medications," b) "Med changes not clinically indicated," or c) "Defer for other reason."

The patients had a mean age of 72 years, 31% were female, and 18% were Black, similar to the nationally representative CHAMP-HF population.

At baseline, most patients were on a beta blocker (84%), or an ACEi/ARB/ARNI (71%), but fewer than a third (29%) were on an MRA, and only 11% were on an SGLT2i.

The primary outcome, an increase in the number of GDMT classes prescribed within 30 days from baseline, occurred in more patients with than without alerts in their EHRs (adjusted relative risk [RR], 1.41; 95% CI, 1.03 - 1.93; P = .03).

The number of patients who needed to have an alert in their EHR to result in an added drug class was 14.

Alert Dismissal, Uptake Barriers

In his Trials and Fibrillations with Dr John Mandrola blog, Mandrola writes about PROMPT-HF, saying "there are just too many [EHR] alerts, and we have become too accustomed to ignoring them. Clicking through alerts has figured prominently in an ongoing criminal medical error case at Vanderbilt."  

This is "a great point," Wilson agreed. "These systems need to be rigorously evaluated, ideally in the context of a randomized trial, lest a proliferation of alerts render all alerts obsolete." If alerts have no effect, they should be abandoned.

Alert "dismissal" is common, even in PROMPT, Wilson added. "That said, we have spoken to providers who dismiss the alert, only to come back later and order the meds, perhaps because prescribing was not in their workflow at the moment."

According to Walsh, clinicians might override the prompts for several reasons, including previous medication intolerance. "The EHR can't know everything about the patient that the clinician does."

Wilson identified uptake barriers as cost, particularly for SGLT-2is, which don't have generic equivalents, lack of knowledge, inexperience with some of these agents, and clinical inertia.

"We are often loathe to change anything when a patient is doing well, and so gentle nudges like that from PROMPT might help to improve care even among those who are relatively stable," he said. "The best we can do here is look at the empiric data. Does the alert increase the use of GDMT? It appears the answer is yes."

The study was funded by AstraZeneca. Ahmad and Wilson have no relevant financial disclosures. The disclosures of the other authors are listed with the publication. Walsh has no relevant financial disclosure. Van Spall is funded by the Canadian Institutes of Health Research and Heart and Stroke Foundation of Canada. The disclosures of the other editorialists are listed with the publication.

American College of Cardiology (ACC) 2022 Annual Scientific Session: Abstract 406-12. Presented April 3, 2022.

J Am Coll Cardiol. Published online April 3, 2022. Abstract, Editorial

For more from theheart.org | Medscape Cardiology, join us on Twitter and Facebook

Comments

3090D553-9492-4563-8681-AD288FA52ACE
Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.

processing....

Recommendations