Large Language Model-Generated Messages to Improve Guideline-Directed Medical Therapy in Heart Failure

NCT07337577 | Not Yet Recruiting FDA Device

• 1 other identifier: 2025P003214
• Study Type: interventional
• Target: 500
• Locations: 1 country
• Sites: 1

Brief Summary

This study is an investigator-initiated, cluster-randomized implementation trial evaluating a large language model (LLM)-based clinical decision support (CDS) tool designed to improve guideline-directed medical therapy (GDMT) for adult patients with heart failure seen in outpatient cardiology clinics at Mass General Brigham. For eligible heart failure encounters, the CDS tool reviews existing electronic health record (EHR) data, including diagnoses, medications, vital signs, laboratory results, and recent notes, and generates brief, clinician-facing messages suggesting opportunities to initiate or optimize GDMT and highlighting relevant safety considerations. Messages are delivered to cardiology providers via Epic InBasket and/or institutional email prior to scheduled visits. The tool is advisory only and cannot place orders or change medications automatically; all treatment decisions remain at the discretion of the treating clinician and patient. Cardiology providers are assigned at the provider/clinic level to early implementation of the CDS tool versus usual care (no messages) during the initial phase. The primary outcome is GDMT optimization within 30 days of an index visit. Secondary outcomes include feasibility of CDS generation and delivery and a 30-day safety composite (e.g., heart failure hospitalization, acute kidney injury, hyperkalemia, hypotension or bradyarrhythmia plausibly related to GDMT).

Conditions

Heart Failure

Keywords

Guideline-directed medical therapy; Artificial Intelligence

Outcome Measures

Primary Outcomes (1)

• Any GDMT optimization within 30 days of index visit

  Among eligible HF encounters, the proportion with initiation of at least one new GDMT class not previously prescribed and/or uptitration of at least one existing GDMT medication during or within 30 days after the index visit, comparing early-implementation vs usual care arms.

  30 days

Secondary Outcomes (2)

• Short-Term Safety Composite (30 days)

  30 days

• Operational Feasibility

  30 days

Study Arms (2)

Early Implementation

EXPERIMENTAL

Providers in this arm receive a large language model-based clinical decision support (LLM-GDMT CDS) intervention. For eligible outpatient heart failure encounters, the CDS tool reviews existing EHR data (diagnoses, medications, vitals, labs, recent notes) and generates a brief, clinician-facing message summarizing HF status, suggesting opportunities to initiate or optimize guideline-directed medical therapy (GDMT), and highlighting safety considerations. Messages are delivered via Epic InBasket and/or institutional email in advance of the visit. The tool is advisory only and cannot place orders or directly change medications; all treatment decisions remain at the discretion of the treating clinician and patient.

Device: LLM-GDMT Clinical Decision Support Tool

Usual Care (Delayed Implementation)

NO INTERVENTION

Providers in this arm continue usual care and do not receive LLM-GDMT CDS messages during the initial evaluation phase. Eligible outpatient heart failure encounters are managed according to routine clinical practice without additional CDS messages. EHR data from these encounters are used to compute GDMT utilization and safety outcomes for comparison with the early-implementation arm. After the initial evaluation phase is complete, the LLM-GDMT CDS tool may be expanded to providers in this arm as part of routine care.

Interventions

LLM-GDMT Clinical Decision Support Tool DEVICE

Software-only, large language model-based clinical decision support tool that reviews structured and unstructured EHR data for adult heart failure patients and generates brief, clinician-facing messages suggesting opportunities to initiate or optimize guideline-directed medical therapy (GDMT) and highlighting relevant safety considerations. Messages are delivered to cardiology providers via Epic InBasket and/or institutional email prior to eligible outpatient visits. The tool is advisory only and cannot place orders or directly change medications; all treatment decisions remain at the discretion of the treating clinician and patient.

Early Implementation

Eligibility Criteria

• Age: 18 Years - 85 Years
• Sex: all
• Healthy Volunteers: No
• Age Groups: Adult (18-64), Older Adult (65+)

You may qualify if:

• Age ≥18 years
• Scheduled outpatient visit with a participating cardiology provider in an MGB outpatient cardiology clinic
• At least one prior cardiology clinic visit in the MGB system within the past 2 years
• Diagnosis of heart failure by ICD code within the past 2 years
• Heart failure diagnosis supported by at least one of the following:
• Current or recent use of a loop diuretic
• Left ventricular ejection fraction ≤40% on the most recent echocardiogram
• Explicit documentation of heart failure diagnosis or heart failure signs/symptoms in a prior cardiology note

You may not qualify if:

• Systolic blood pressure \<90 mmHg on the most recent recorded measurement
• Heart rate \<50 beats per minute on the most recent recorded measurement
• eGFR \<20 mL/min/1.73 m² on the most recent laboratory assessment
• Known cardiac amyloidosis or hypertrophic cardiomyopathy
• History of heart transplant or presence of a left ventricular assist device
• Severe aortic stenosis, severe aortic insufficiency, or severe mitral stenosis on the most recent echocardiogram
• Encounter occurs in an adult congenital heart disease clinic

Sponsors & Collaborators

• Brigham and Women's Hospital: lead

Study Sites (1)

Mass General Brigham

Boston, Massachusetts, 02115, United States

Location

Related Publications (3)

• McDuff D, Schaekermann M, Tu T, Palepu A, Wang A, Garrison J, Singhal K, Sharma Y, Azizi S, Kulkarni K, Hou L, Cheng Y, Liu Y, Mahdavi SS, Prakash S, Pathak A, Semturs C, Patel S, Webster DR, Dominowska E, Gottweis J, Barral J, Chou K, Corrado GS, Matias Y, Sunshine J, Karthikesalingam A, Natarajan V. Towards accurate differential diagnosis with large language models. Nature. 2025 Jun;642(8067):451-457. doi: 10.1038/s41586-025-08869-4. Epub 2025 Apr 9.

  PMID: 40205049 BACKGROUND

• Greene SJ, Butler J, Albert NM, DeVore AD, Sharma PP, Duffy CI, Hill CL, McCague K, Mi X, Patterson JH, Spertus JA, Thomas L, Williams FB, Hernandez AF, Fonarow GC. Medical Therapy for Heart Failure With Reduced Ejection Fraction: The CHAMP-HF Registry. J Am Coll Cardiol. 2018 Jul 24;72(4):351-366. doi: 10.1016/j.jacc.2018.04.070.

  PMID: 30025570 BACKGROUND

• Heidenreich PA, Bozkurt B, Aguilar D, Allen LA, Byun JJ, Colvin MM, Deswal A, Drazner MH, Dunlay SM, Evers LR, Fang JC, Fedson SE, Fonarow GC, Hayek SS, Hernandez AF, Khazanie P, Kittleson MM, Lee CS, Link MS, Milano CA, Nnacheta LC, Sandhu AT, Stevenson LW, Vardeny O, Vest AR, Yancy CW. 2022 AHA/ACC/HFSA Guideline for the Management of Heart Failure: A Report of the American College of Cardiology/American Heart Association Joint Committee on Clinical Practice Guidelines. J Am Coll Cardiol. 2022 May 3;79(17):e263-e421. doi: 10.1016/j.jacc.2021.12.012. Epub 2022 Apr 1.

  PMID: 35379503 BACKGROUND

MeSH Terms

Conditions

Heart Failure

Condition Hierarchy (Ancestors)

Heart Diseases; Cardiovascular Diseases

Central Study Contacts

Jonathan W Cunningham, MD, MPH

CONTACT

617-732-8534; jcunningham3@bwh.harvard.edu

Study Design

• Study Type: interventional
• Phase: not applicable
• Allocation: RANDOMIZED
• Masking: NONE
• Purpose: HEALTH SERVICES RESEARCH
• Intervention Model: PARALLEL
• Sponsor Type: OTHER
• Responsible Party: PRINCIPAL INVESTIGATOR
• PI Title: Assistant Professor of Medicine

Study Record Dates

• First Submitted: January 12, 2026
• First Posted: January 13, 2026
• Study Start (Estimated): June 1, 2026
• Primary Completion (Estimated): July 1, 2027
• Study Completion (Estimated): December 1, 2027
• Last Updated: April 21, 2026

Record last verified: 2026-04

Locations

US (1)