GUARD-PH: Guided Use of AI-ECG for Risk Detection of PH in Surgery (GUARD-PH Trial)

Overview

This study is a prospective, open-label, randomized controlled trial designed to evaluate a new artificial intelligence (AI) tool for heart monitoring. Researchers will use an AI-enabled electrocardiography (ECG) system to screen patients before they undergo surgery. The main goal is to determine if this AI system can accurately detect pulmonary hypertension and related heart diseases in the preoperative setting. The study is being conducted at Taipei Veterans General Hospital.

This is an open-label, prospective, randomized controlled trial at Taipei Veterans General Hospital. Approximately 1,380 adult patients who already have a preoperative electrocardiogram (ECG) scheduled as part of routine care and have an established plan for elective surgery will be enrolled, along with approximately 30-60 participating physicians responsible for preoperative assessment and perioperative clinical care. The study evaluates whether integrating the Taiwan Medical Imaging Pulmonary Hypertension Detection System (TAIMedImg PHDS), an ECG-based clinical decision-support tool, during the preoperative preparation period can enable earlier identification of patients at risk of pulmonary hypertension and cardiopulmonary complications, and supports assessment of the potential cost-effectiveness of earlier risk recognition. Patients are randomized to determine whether their responsible participating physician will receive TAIMedImg PHDS output derived from the patient's routine ECG. In the intervention arm, research staff upload the ECG to TAIMedImg PHDS and provide the analysis result to the participating physician; all other processes proceed as usual care. In the control arm, usual care proceeds without provision of TAIMedImg PHDS output. Randomization affects only access to this supplementary ECG-based output; it does not change patients' scheduled tests, clinical workflow, or rights to care. Clinical evaluation and management remain at the physician's discretion according to standard practice. The study measures physicians' responses to ECG interpretation with versus without TAIMedImg PHDS output and documents the clinical decision-making trajectory, using routinely available medical record data for analysis. Results may inform improved preoperative pulmonary hypertension risk assessment and future perioperative care.