An increasing number of guideline-directed medical therapies (GDMT) have been developed for patients with chronic heart failure with reduced ejection fraction (HFrEF). When used in combination at recommended doses, patients often experience significant improvements in cardiac function, quality of life, and survival.1,2 However, GDMT underuse occurs for the vast majority of patients with HFrEF. Two recent trials demonstrated improved GDMT prescribing during a clinic visit, each using automated delivery of a patient-centered decision support tool to promote a proactive and holistic approach to prescribing: EPIC-HF (NCT03334188) tested a brief video and checklist document sent to patients just prior to a clinic visit encouraging them to work with their clinicians to make at least 1 positive change to their GDMT; PROMPT-HF (NCT05433220) tested tailored electronic health record (EHR) alerts for GDMT intensification delivered to clinicians during clinic visits. The current I-I-CAPTAIN-HF study aims to broadly implement and test the EPIC-HF patient-facing and PROMPT-HF clinician-facing tools for HFrEF medication intensification at 5 health systems around the country through a pragmatic cluster-randomized implementation-effectiveness trial. This will occur through an initial phase of adaptation of the 2 tools at each health system. Once ready, the 2 tools will be tested using a 2x2 randomization at the clinician-level. In parallel, formal assessment of the implementation of EPIC-HF and PROMPT-HF will work to understand the most effective means of intervention design and delivery, as well as adaptations due to contextual factors to optimize use.

I-I-CAPTAIN-HF is a multi-site, type-2 hybrid implementation-effectiveness trial conducted at 5 sites: the University of Colorado (central site), Yale University, Northwestern University, University of Utah, and Sutter Health.

Aim 1. Implement the EPIC-HF and PROMPT-HF interventions at 5 health systems through participatory work with a multi-level team using the PRISM framework to guide the delivery of a host of implementation strategies that respect mandatory aspects of the interventions (function) and explore adaptations (form) to maximize equitable reach, broad adoption, fidelity of implementation, and long-term maintenance.

• Hypothesis 1A: Standard EHR logic (in Epic) can be created for HFrEF identification and automated delivery of both interventions, followed by tailoring to individual system EHR infrastructure.
• Hypothesis 1B: Challenges around automated intervention delivery can be identified, solutions developed, and processes shared for optimal integration into the local clinic environment.
• Hypothesis 1C: Issues of equity-digital access, language, out-of-pocket cost, social determinants of health-can be addressed by intervention and delivery adaptations, respecting uniqueness of recipients.

The initial phase of I-I-CAPTAIN-HF explores a host of implementation strategies that will take the existing EPIC-HF and PROMPT-HF interventions and optimize their ability to be implemented widely while retaining tool fidelity. Current versions of EPIC-HF and PROMPT-HF were tested in single sites under research conditions that have led to barriers to wider uptake. In Aim 1, the investigators and study team will 1) form an External Advisory Committee, 2) apply PRISM domains to define mandatory functions of the interventions, 3) engage site partners to define form adaptations using the iPRISM web tool to assess context, 4) vet adaptable forms to consider fidelity to function, 5) decide final forms for each site, and 6) pilot local versions of each interventions using EHR technical testing to ensure feasibility, usability, and acceptability. The tools will be implemented and automated within each EHR and used as standard of care for all eligible patients.

Prior to conclusion of Aim 1, all sites will pilot test their site-based implementation approach to assess feasibility, usability, and user acceptability before widespread implementation in Aim 2. This process will conclude with interviews with team members to assess the fidelity to the intervention, as well as gather data on adaptations to the intervention by each site. These interviews will be conducted by the central site (University of Colorado) study team and occur with those involved at each site with implementing the interventions and could include site-principal investigators and/or co-investigators, IT leadership or analysts.

Aim 2. Test the effectiveness, as well as the interaction, of the EPIC-HF and PROMPT-HF interventions to improve GDMT prescribing from before to after the clinic visit among 2000 patients at 5 health systems using a 2x2 factorial design randomized at the clinician-level.

• Hypothesis 2A - Effectiveness: Both EPIC-HF and PROMPT-HF will improve GDMT intensification.
• Hypothesis 2B - Interaction: Patient- and clinician-facing tools may work better when used in combination.

Cluster-randomization will occur at the clinician level, where the clinician is randomized to 1 of 4 study arms (usual care, EPIC-HF only, PROMPT-HF only, both). Randomization will be stratified at the site level. Patient assignment follows clinician randomization.

Following processes developed in Aim 1, each site will deliver the assigned intervention(s) to patients and clinicians as per randomization. Those clinicians randomized to intervention(s) will have the intervention delivered to the clinician themselves (PROMPT-HF) and/or their patient (EPIC-HF) prior to the patient's scheduled routine clinic appointment. The interventions will be implemented as standard of care across those in the intervention arms and occur for all eligible patients.

Patient-level data will be collected through the patients' EHRs.

Aim 3. Evaluate ongoing implementation of EPIC-HF and PROMPT-HF using Practical, Robust Implementation and Sustainability Model (PRISM) to understand the most effective means of intervention design and delivery, as well as adaptations due to contextual factors.

• Hypothesis 3: Understanding key mandatory elements and tailored adaptations to the interventions and their delivery will maximize equitable impact across implementation outcomes (reach), increase uptake by the clinics and their clinicians (adoption), ease appropriate deployment (implementation and adaptations), and lead to sustained use after the clinical trial (maintenance).

Implementation of EPIC-HF and PROMPT-HF will be explored using the Reach, Effectiveness, Adoption, Implementation, and Maintenance (RE-AIM) outcomes from the PRISM framework to understand the most effective means of sustainable delivery.

Reach: Investigators will attempt to characterize the proportion and representativeness of eligible patients who receive the intervention. Representativeness will be assessed by comparing participants who opt in versus those who opt out based on demographic/clinical indicators.

Effectiveness: Assessed in Aim 2, above.

Adoption: The absolute number, proportion, and representativeness of settings and intervention agents (clinicians) who are willing to initiate a program will be assessed. Effectively, adoption should be 100% at the site level since all programs who were approached have agreed to participate. Data will be collected on adoption at the clinician level, where each clinician can enroll or decline at study start, and then adopt or ignore PROMPT-HF CDS during study conduct. Clinicians' attitudes towards the new approach will be explored using clinician surveys measured before and after the trial.

Implementation: The investigators will assess the extent to which the intervention is implemented as intended (fidelity) and adaptations made. With a mixed methods evaluation, the experiences of EPIC-HF and PROMPT-HF and their combination will be explored.

Maintenance: The study team will systematically collect data on local decisions regarding maintenance and will also assess staff recommendations for sustainment.

Inclusion Criteria:

Clinician:

• Clinician (MD, PA, NP) who practices in cardiology outpatient clinics
• Regularly sees patients with left ventricular ejection fraction (EF) </=40%, where their panel of patients over the last year included at least 25 patients with heart failure with reduced ejection fraction (HFrEF)

Patient:

• Age > 18 years
• LVEF \

Only aggregate participant data will be reported. A de-identified study dataset will be available on a data repository website after study collection and analysis is completed.