A randomized controlled trial of a novel AI-based prognostic decision support tool integrated into the electronic health record for ischemic heart disease: Rationale and design of the PM Heart trial.

BACKGROUND

Ischemic heart disease (IHD) is the leading global cause of years of life lost. Although the clinical presentation and progression of IHD differ widely, clinical management is relatively uniform. To improve risk stratification and clinical management, we developed a decision support tool - the PMHeartIHD Algorithm (also named PMHNet) - based on extensive clinical, registry and biochemical data. This algorithm automatically calculates one-year survival following an IHD event. We hypothesize that presenting the prognoses calculated by the PMHeartIHD Algorithm to the treating physicians will improve both the clinical management and outcome of the individual patient.

OBJECTIVE

The objectives of the PM Heart trial include clinical and technical aspects. First, the primary aim is to investigate if the clinical use of the prognosis calculated by the algorithm improves patient prognosis and minimizes the risk of re-hospitalization, compared to standard of care. A secondary objective is to generate knowledge on how decision support tools in cardiology may be implemented, received, and used directly in clinical practice.

METHODS

The PMHeartIHD Aalgorithm has been integrated into the electronic health record (EHR) system (Sundhedsplatformen (SP), EPIC) for clinical use. The PM Heart trial is designed as a multicenter, prospective, randomized, open-label, clinical trial, automatically enrolling and randomizing hospitalized patients with coronary artery disease at the cardiology departments in The Capital Region and Region Zealand in Denmark. Patients are randomized (1:1 ratio) immediately after coronary angiography, to either show or hide the calculated PMHeartIHD-prognosis or not in the electronic health record.

RESULTS

The primary outcome is a composite outcome consisting of re-admissions within 30 days and one-year survival rates after randomization. Inclusion began September 2023, and follow-up is expected to finish in 2026.

CONCLUSIONS

We present the design of and rationale for the PM Heart trial. Our aim is to assess the clinical value of implementing an artificial intelligence-based decision support tool in clinical practice.