NCT06593314

Using Ultromics EchoGo HFpEF Algorithm to Identify and Treat High Heart Failure Risk in Patients With Type 2 Diabetes

Identifying Undiagnosed HFpEF Among Patients With Type 2 Diabetes Using Ultromics AI HFpEF Algorithm

• Status: Recruiting
• Phase: N/A
• Study type: Interventional
• Enrollment: 800 (estimated)
• Sponsor: University of Texas Southwestern Medical Center · Academic / Other
• Sex: All
• Age: —
• Healthy volunteers: Not accepted

Summary

A pragmatic electronic health record (EHR) based randomized controlled trial to evaluate the utility of providing Ultromics EchoGo analysis results and recommendations for HF risk prevention therapies using an EHR embedded clinical decision support tool.

Detailed description

Historic echocardiograms will be analyzed using the Ultromics EchoGo algorithm. For patients that have a positive EchoGO result i.e. HFpEF detected, the provider will get an clinical decision support alert flagging high risk of HFpEF based on randomized assignment. Experimental: Alert Group Provider will receive a computer-based provider-to-provider message notifying the provider that the patient has subclinical HFpEF as determined by the Ultromics EchoGo algorithm and associated guideline recommendations for the management of these patients. The alert will include guideline-based and standard-of-care recommendations for the use of SGLT-2 inhibitors, non-steroidal MRA, or GLP-1 RA (if obesity is present). The purpose of the alert is to inform the providers about the risk of heart failure and provide them guidance regarding the guideline-recommended standard of care. The providers can choose to provide care as deemed fit based on the information provided. The investigators will assess the practice patterns of providers in response to the EHR-based alert over the study period (3, and 6-month follow-up). The investigators will also assess the downstream hospitalization events for HF within 12 months of the initial alert. Control arm: Standard Message Providers in the control group will receive a standard message that will recommend either SGLT2i, GLP-1RA, and/or ns-MRA for treatment of diabetes and for prevention of heart failure. This group will not receive any information about the presence of subclinical heart failure detected by the EchoGo algorithm. The investigators will monitor the practice pattern in this group as well over the study period. Follow Up. Adherence to SGLT-2i and GLP-1 RA medications will be assessed by evaluating the electronic health record and documenting if the patient had a follow-up with a healthcare provider at 3 and 6 months and medication listed in the active prescription medication list. Sample Size: The investigators plan to enroll 800 anticipated patients using a parallel design with 1:1 allocation and a binary primary endpoint (SGLT2i use). Using a two-sample test for difference in proportions with the normal (Fleiss) approximation, pooled variance without continuity correction, and assuming a control proportion of 30%, α=0.05 (two-sided), and 80% power, an N=800 (400/arm) provides a minimum detectable absolute increase of \~9.4 percentage points (30.0% to 39.4% in the intervention arm). This corresponds to RR = 1.31(95% CI 1.08, 1.59) and Cohen's h = 0.20.

Conditions

• Heart Failure
• Diabetes

Interventions

Type	Name	Description
BEHAVIORAL	Message with EchoGo	This alert will inform the provider that the patient has subclinical HFpEF
BEHAVIORAL	Standard Message	This alert will inform the provider of guideline directed treatment options for patients with diabetes to prevent heart failure.

Timeline

• Start date: 2025-08-06
• Primary completion: 2026-03-15
• Completion: 2027-06-15
• First posted: 2024-09-19
• Last updated: 2025-08-12

Locations

1 site across 1 country: United States