System-level integration of electronic health record-based tools to improve heart failure prescribing: learnings from the I-I-CAPTAIN-HF Trial implementation
L A Allen, K Trinkley, J Thompson, M Viny, C K McIlvennan, F Ahmad, J Stehlik, F P Wilson, S Yan, S Carter, K Kawamoto, L Gleason, B Wallace, K Noonan, D D MatlockAbstract
Background
Guideline-directed medical therapy (GDMT) for heart failure with reduced ejection fraction is highly effective but underused, in part due to clinical inertia. To improve GDMT, the EPIC-HF trial demonstrated effectiveness of a patient-directed video and checklist and the PROMPT-HF trial demonstrated effectiveness of tailored electronic health record (EHR) alerts delivered to clinicians. Neither EPIC-HF nor PROMPT-HF has been implemented widely.
Purpose
I-I-CAPTAIN-HF is a type-2 hybrid trial designed to test the effectiveness and implementation of the EPIC-HF and PROMPT-HF interventions across 5 diverse health systems in the United States. The interventions’ effects on GDMT intensification are being assessed in 2,000 patients among 256 clinicians. We present results of integration of the tools within the EHR.
Methods
The lead site collected field notes throughout the implementation period, recording details from regular project meetings, email communications, and informal interactions with site teams on implementation status. Sites documented implementation activities, emerging challenges, and contextual factors. Thematic analysis was used to determine emergent themes.
Results
All 5 sites used their own institution’s instance of their EHR (Epic® Systems) for tool delivery. Sites leveraged existing information technology (IT) teams who conducted the builds for (1) identifying eligible patients, (2) EPIC-HF tool delivery to patients, and (3) PROMPT-HF alerts to providers. Based on this information, the lead site decided on EPIC-HF tool delivery through Epic’s Care Companion feature using the MyChart interactive patient portal. Build guides and pre-built, importable packages (Epic Turbocharger) were provided to sites for implementation. All sites faced a complex multilayer process of institutional approval of the IT build. Each institution experienced at least one issue that prevented the interventions from working accurately, and the resolution process identified that IT teams were often siloed and did not communicate well across teams or roles (Figure).
Conclusion
EHR-integrated interventions are being developed at a rapid pace, enhanced by advances in decision-support tools. Our study identifies key processes that need to be improved in the implementation stage, including the need for streamlined institutional approvals, cross-team relationship building, expanded IT support, and use of a structured piloting period prior to release of new EHR-based tools.