Artificial intelligence-enhanced management system for secondary prevention of coronary heart disease: a randomized clinical trial.

Abstract

Secondary prevention of coronary heart disease (CHD) remains suboptimal due to fragmented care and therapeutic inertia. While digital health interventions offer a promising strategy, few existing tools provide comprehensive, closed-loop management of multidimensional risk factors without increasing clinician workload. This study aimed to evaluate the efficacy of an artificial intelligence (AI)-enhanced management system (AIM-CHD) in improving multidimensional risk factor control among patients with CHD. This single-center, open-label, parallel-group, randomized clinical trial was conducted from November 2024 to June 2025 in China. A total of 1100 adults with confirmed CHD were randomized 1:1 to receive either the AIM-CHD intervention (n = 549) or usual care (n = 551) for 3 months. The AIM-CHD system featured automated multi-source data capture, guideline-directed risk stratification, and closed-loop feedback via smartphone. The primary outcome was the low-density lipoprotein cholesterol (LDL-C) level at 3 months. Secondary outcomes included goal attainment rates for LDL-C (< 70 mg/dL) and blood pressure (BP, < 130/80 mm Hg), glycated hemoglobin, smoking cessation, body mass index (BMI), medication adherence, and a composite cardiovascular endpoint at 3 months. Analysis was performed on an intention-to-treat (ITT) basis. Of 1100 randomized participants (mean age 61 years; 75.3% male), 943 (85.7%) completed the 3-month follow-up. The intervention group achieved a significantly lower mean LDL-C level compared with the control group (60.4 ± 23.4 vs. 63.7 ± 26.0 mg/dL), with an adjusted mean difference of - 3.2 mg/dL (95% CI, - 6.2 to - 0.3; p = 0.03). Furthermore, participants in the intervention group were more likely to achieve the LDL-C target (71% vs. 64%; RR, 1.10; 95% CI, 1.01-1.20; p = 0.03) and the BP target (45% vs. 35%; RR, 1.27; 95% CI, 1.09-1.48; p = 0.002). No significant differences were observed for HbA1c, BMI, or medication adherence. The AIM-CHD system significantly improved short-term lipid and BP control compared with usual care. These findings support interoperable, low-burden digital management systems as a scalable strategy in routine secondary prevention. ClinicalTrials.gov Identifier: NCT06686056. Registered on 11 November 2024.