Digital heart failure management: Insights from the COPILOT study

Abstract

Suboptimal implementation of guideline-directed medical therapy (GDMT) remains a major problem in heart failure (HF) care. The Cooperative Program for Implementation of Optimal Therapy in Heart Failure (COPILOT-HF) trial evaluated an innovative remote strategy to optimize GDMT across the spectrum of HF, irrespective of left ventricular ejection fraction (LVEF). This pragmatic randomized study compared a pharmacist-led, navigator-assisted remote program against traditional provider notification and education. The results, recently presented at the American Heart Association 2025 Scientific Sessions, offer compelling insights into overcoming clinical inertia and improving HF management in the digital age. This mini review examines the trial’s key findings and their potential to reshape HF care delivery.

Introduction

Advances in heart failure (HF) therapy has provided the opportunity for substantial reductions in HF events, prolonged survival and improved quality of life [1]. Guideline-directed medical therapy (GDMT) remains the cornerstone of current HF treatment and includes a foundational four-drug regimen for patients with HF and reduced left ventricular ejection fraction (LVEF; HFrEF) [2, 3]: (1) renin-angiotensin system inhibition with angiotensin-converting enzyme inhibitors (ACEi), angiotensin receptor blockers (ARB), or angiotensin receptor neprilysin inhibitors (ARNi); (2) beta-blockers (BB); (3) mineralocorticoid receptor antagonists (MRAs); and (4) sodium-glucose co-transporter-2 inhibitors (SGLT2i) [3, 4].

Despite compelling evidence, the implementation of GDMT and HF follow-up remains suboptimal [5, 6]. Efforts to improve GDMT adherence have become a major focus of current clinical HF research, and a recent systematic review reported that interdisciplinary HF clinics were associated with substantial improvements in GDMT adherence [7]. Additional interventions, including audits, clinician and patient education, and electronic health record alerts, were also associated with improvements in GDMT utilization. While achieving optimal treatment for all patients in real-world practice may seem daunting, national registry data from Norway offer hope, demonstrating that > 70% of patients with HFrEF can successfully be treated with quadruple therapy [8].

Trial design and methodology

The Cooperative Program for Implementation of Optimal Therapy in Heart Failure (COPILOT-HF) trial is a pragmatic, randomized, open label trial that was designed to prospectively compare two strategies for remote GDMT optimization across the spectrum of HF, regardless of LVEF [9]. The first strategy was an “education-first” approach where the intervention was provider and patient education with the intention of increasing compliance and acceptance to GDMT by improving patients’ understanding of their diagnosis and its management. At 3 months postrandomization, the strategy was followed by a remote clinic that implements a standardized, stepped approach to GDMT optimization. In the second strategy, both interventions were started simultaneously [10]. The primary objective was to examine whether a remote pharmacist-led education program of patients and providers, assisted by non-licensed navigators, could achieve higher rates of GDMT implementation compared with provider notification and education alone. The investigators employed an innovative screening approach, utilizing advanced natural language processing to identify eligible participants [11]. Study staff were randomized to either manual or AI-assisted prescreening methods. The results demonstrated that large language model-based solutions can significantly enhance clinical trial screening performance and reduce costs by automating the process. This novel approach highlights the potential of AI in streamlining clinical research recruitment.

The study included adults aged 18–90 with documented HF, regardless of LVEF, who have undergone a transthoracic echocardiogram and had symptomatic HF within two years before enrollment. Key exclusions were hypotension, type 1 diabetes, known use or intolerance of SGLT2i or ARNI, severe aortic valve disease, kidney failure, amyloid heart disease, and pulmonary arterial hypertension requiring treatment. All eligible participants were randomized 1:1 to either the “education-first” strategy (3-month period of patient and provider education, followed by enrollment in a remote HF management program for the optimization of GDMT), or to the simultaneous initiation of both interventions.

The primary efficacy endpoint was the proportion of participants receiving full GDMT at 3 months. Full GDMT was defined as either (1) all four drug classes (BB, ACEi/ARB/ARNi, MRA or SGLT2i) for participants with LVEF < 50%, or (2) prescription of SGLT2i for participants with LVEF ≥ 50%. The primary implementation outcome at 3 months was adjusted after the interim analysis, with full GDMT defined as: (1) all four drug classes (BB, ACEi/ARB/ARNi, MRA or SGLT2i) for participants with LVEF < 50%, (2) three drug classes (ACEi/ARB/ARNi, MRA or SGLT2i) for participants with LVEF 50–59% or (3) SGLT2i alone for participants with LVEF ≥ 60%.

The secondary endpoint included the initiation or up-titration of any GDMT medication. Safety outcomes included kidney complications (increase in serum creatinine by ≥ 1.5 times the baseline value), hypotension (systolic blood pressure < 90 mmHg with associated symptoms), hypo/hyperkalemia (serum potassium < 3.0 mmol/L or > 6.0 mmol/L), hospitalizations, and death.

A prespecified interim analysis was performed after the first 100 participants completed 4 months in the program. At that time, 4 participants (8%) in the education-first arm achieved the primary endpoint compared with 27 participants (54%) in the simultaneous arm (p < 0.001). Based on these results, the protocol was transitioned to a single-arm intervention with simultaneous initiation of both interventions. The last patient was enrolled in February 2025.

Results

The investigators ultimately enrolled a total of 503 HF patients. Overall, the mean age was 72 years, 52% were male, and 89% were white. The majority (65%) had LVEF ≥ 60% and only 14% had LVEF < 40%. The most common comorbidities were hypertension (87%), obesity (mean BMI 33 kg/m²) and type 2 diabetes mellitus (30%), and 29% were hospitalized for HF within 1 year of enrollment. The mean baseline labs were NT-proBNP 1213 pg/mL, potassium 4.3 mEq/L, and eGFR 68 mL/min per 1.73 m². Baseline medications were: 89% BB, 33% MRA, 32% SGLT2i and 30% ARNI for EF < 50% and 77% BB, 18% MRA, 2% SGLT2i and 4% ARNI for EF ≥ 50%.

In the cohort of patients included in the original design component (n = 300), more patients in the simultaneous arm achieved the primary endpoint at 3 months compared to education first (32.0% vs. 0%, p < 0.001). In the single-arm component (n = 203), 35% of patients achieved the primary endpoint at 3 months. Patients with EF ≥ 50% reached the primary endpoint faster and more frequently than patients with EF < 50%. The investigators conclude that a strategy of remote medication implementation, combined with education, appeared to result in a higher proportion of patients with HF receiving optimal medical therapy compared with education first.

Discussion

The healthcare landscape is increasingly focusing on remote follow-up of patients with HF as a promising strategy to enhance GDMT implementation. The COPILOT-HF trial provides compelling evidence that a remote HF clinic capable of initiating and titrating GDMT can serve as an effective alternative to traditional curated patient and provider education methods. Findings from the trial and the “education-first” arm may suggest the persistence of clinical inertia towards GDMT prescribing, even when patients and providers are offered state-of-the-art educational materials, including summary sheets on indications, outcomes, prescribing, and monitoring information [9]. One possible explanation for clinicians’ reluctance to escalate GDMT is the misinterpretation of stable symptoms as reflecting low risk, as well as the interpretation of certain biological findings as contraindications to up-titration (e.g., low eGFR or asymptomatic low blood pressure), coupled with limited awareness of the potential harms associated with omitting medications shown to reduce risk of mortality and/or hospitalization [12]. While the data do not provide information about reasons for underutilization of GDMT, such as side effects or contraindications to initiation or up-titration, the results are nonetheless notable. After just three months, 32% of participants in the simultaneous group achieved full GDMT, compared to 0% in the control group. Although it is concerning that two-thirds of patients in the intervention group did not achieve optimal GDMT, the results are encouraging and may support this strategy for improving GDMT implementation. This significant improvement in a relatively short timeframe may highlight the potential of targeted interventions in HF management.

The use of multidisciplinary HF clinics is known to reduce HF hospitalizations and all-cause mortality, which is why they are recommended in current guidelines [2, 3]. Safety is of great importance when implementing new treatment strategies. In COPILOT-HF the chosen safety endpoints – worsening renal function, potassium disturbances and hypotension – appear reasonable, as these are common adverse effects of GDMT initiation and titration and may limit achievement of full GDMT. Further long-term data on HF hospitalizations and mortality in the trial will also be of great importance when comparing this strategy with other multidisciplinary team-based approaches. Detailed data on safety outcomes remain to be published and will contribute with valuable information whether this strategy should be incorporated in broader clinical practice.

Although there are both pros and cons to digital follow-up (Fig. 1), the COPILOT-HF results suggest that similar benefits may be achievable in a remote setting. This finding is particularly important, as it opens up possibilities for expanding access to high-quality HF care, especially for patients in rural or underserved areas who may face barriers to accessing traditional in-person clinics.

Fig. 1.

Balancing act: Pros and cons of digital follow-up for heart failure patients

In COPILOT-HF, specially trained personnel without formal medical training were primarily responsible for direct communication with patients. The uptitration of GDMT was performed by a team consisting of these navigators collaborating with pharmacists, under the supervision of an advanced practice practitioner and a cardiologist. Medication adjustments were determined by a clinical pharmacist based on a predefined algorithm [10]. In COPILOT-HF, this model appeared to be effective in improving GDMT implementation, but its applicability in broader settings remains uncertain. Although GDMT pharmacist clinics are available at selected centers, where data suggest they are effective in improving GDMT use [13, 14], the widespread uptake of these clinics across the United States remains to be seen. In European countries, pharmacists are less often directly involved in medication management for individual patients and more commonly serve as consultants. Such differences may limit the generalizability of the COPILOT-HF model. However, modern healthcare systems worldwide face similar challenges related to high patient volumes, aging populations, and physician shortages and models in clinical follow-up that allow for task redistribution may be beneficial, enabling cardiologist to supervise larger patient populations.

Author contributions

J. Joyson Paul: Conceptualization, Data curation; Resources, Formal analysis, Writing – original draft. R. Durairaja: Investigation, Methodology. N. Jayakumar: Writing – review & editing. M. Sharumathi: Methodology, Software. K. Karuppasamy: Validation, Writing – review &amp; editing. Percy Paulin J: Software, Validation, Visualization. S. Parkavi and S. Pearlin Arul Granap: Formal analysis

Funding

Open access funding provided by University of Oslo (incl Oslo University Hospital)

Data availability

No datasets were generated or analysed during the current study.

Declarations

Competing interests

The authors declare no competing interests.

Disclosures

JMR reports no disclosures.