Electronic Alerts to Improve Heart Failure Therapy in Outpatient Practice: A Cluster Randomised Trial
- J Am Coll Cardiol 2022; S0735-1097(22):04489-8
What was known?
Guideline-directed medical therapy (GDMT) for heart failure with reduced ejection fraction (HFrEF) includes four medication classes: β-blockers (BB), Renin-Angiotensin-Aldosterone System inhibitors [angiotensin converting enzyme inhibitors (ACEI), angiotensin receptor blockers (ARB), or angiotensin receptor-neprilysin inhibitors (ARNI)], mineralocorticoid receptor antagonists (MRA), and sodium-glucose cotransporter-2 inhibitors (SGLT2i). However, real-world evidence shows these medications are under-prescribed in this patient group.
The PRagmatic Trial Of Messaging to Providers about outpatient Treatment of Heart Failure (PROMPT-HF) aimed to investigate whether alerting on recommendations about medical treatment of HFrEF customised for each patient would lead to higher rates of prescription of these therapies when compared with usual care/no alerts.
What this study adds:
Electronic health record (EHR)-based alerts providing tailored recommendations for outpatients with HFrEF resulted in accelerated use of underutilised GDMT.
These data suggest that personalised clinical decision support which could be rooted into EHR at integrated healthcare systems could result in major advancements in the management of HF.
This was a pragmatic, EHR-based, cluster-randomised comparative effectiveness trial (NCT04514458). Healthcare providers involved in HFrEF patient care were randomised to either intervention (alert) or usual care (no alert). An EHR-embedded best practice alert was used that would trigger for patients who were eligible, notifying providers of individualised GDMT recommendations along with patient characteristics. The provider could acknowledge or dismiss the alert.
- Number of patients with HFrEF with an increase in the number of prescribed GDMT classes at 30 days post-randomisation
- Any increase in dose of currently prescribed GDMT
- Filling of prescriptions
- Total healthcare costs
- Hospitalisations and emergency department visits
- Age ≥18
- left ventricular ejection fraction (LVEF) ≤ 40%
- Not already on all four classes of GDMT for HFrEF
Notable exclusion criteria:
- Patients opted out of EHR-based research
- Patients in hospice care
- 1,310 patients were enrolled across 100 providers according to eligibility criteria
- No significant differences in the baseline characteristics between the two groups; 84% of patients were on BB, 71% on an ACEI/ARB/ARNI, 29% on an MRA, and 11% on an SGLT2i
- The primary outcome was noted in 25.7% of the alert arm and 18.7% of the no alert arm [adjusted relative risk (RR), 1.41 (1.03–1.93); p=0.03].
- There was a consistent rise in GDMT utilisation of the alert across sex, race, LVEF, provider type, insurance coverage, and baseline GDMT
- Addition of a class of GDMT or increased dosage was noted in 36.2% of the alert arm vs. 26.2% of the no alert arm [adjusted RR,1.39; 95% confidence interval (CI), 1.08–1.79; p=0.01].
- At 30 days, no significant differences were found in the rate of emergency department visits or hospitalisation between the two groups
- Providers feedback was that the alert was effective for empowering improved prescriptions for HF
The personalised alerts rolled out during this trial, led to a significantly higher number of patients receiving the appropriate GDMT and the clinicians taking part in this trial found the alerts helpful in optimising their patient care. These findings suggest that this low-cost and broadly scalable intervention could rapidly improve the quality of care in HF.
- This study was performed at a single centre
- The study did not look at up-titration of dosing, it concentrated on examining the increase in medication initiation
- The alert was created and tested with the Epic EHR ecosphere only
- During the study period there were rigorous efforts across the Yale New Haven Health System to improve medical therapy for HF; therefore, this may have resulted in a bias in the results