Effectiveness of a personalized rehabilitation programme using an artificial intelligence–based telediagnostic platform in patients with myocardial infarction and atrial fibrillation

Background

Myocardial infarction (MI) complicated by atrial fibrillation (AF) is associated with an increased risk of mortality, a more severe clinical course, and delayed recovery. Patients in this category represent a very high cardiovascular risk group and therefore require safe cardiac rehabilitation programmes adapted to their individual needs. Previous studies have demonstrated a higher effectiveness of personalized rehabilitation programmes in patients with myocardial infarction and comorbid conditions.

Purpose

To develop and evaluate the effectiveness of a personalized rehabilitation programme using an artificial intelligence (AI) based telediagnostic platform in patients with myocardial infarction and atrial fibrillation over a 30-day period.

Methods

The study included 32 patients with myocardial infarction and atrial fibrillation (mean age 69.23 ± 9.56 years; 68.75% male). Patients were divided into two groups: a control group (n = 17) receiving cardiac rehabilitation according to the standard protocol, and an intervention group (n = 15) receiving, in addition to the standard protocol, a personalized rehabilitation programme using an AI-based telediagnostic platform.

Exercise tolerance was assessed using the 6-minute walk test (6MWT), daily step count, and stair-climbing performance on days 10 and 30 of the study. Left ventricular systolic function was evaluated by echocardiography on days 10 and 30. Hemodynamic parameters and electrocardiographic data were monitored using the AI-based telediagnostic platform throughout the study period.

Results

No significant differences in hemodynamic parameters or exercise tolerance were observed between the two groups on the tenth day of rehabilitation.

On the 30th day, patients in the intervention group demonstrated significantly higher exercise tolerance compared with the control group: 6MWT distance 360.9 ± 12.02 vs 315.1 ± 13.09 m (p = 0.015); daily step count 5239.6 ± 224.21 vs 4478.3 ± 299.82 steps (p = 0.049); stair-climbing performance 38.3 ± 1.82 vs 35.7 ± 1.50 steps (p < 0.001).

Left ventricular ejection fraction on the 30th day showed a trend toward improvement in the intervention group compared with the control group (44.5 ± 2.20% vs 42.2 ± 4.34%, p = 0.06).

The number of AF recurrences during the study period was lower in the personalized rehabilitation group (6.7% vs 35.3%, p = 0.043).

Conclusion

A personalized cardiac rehabilitation programme using an AI-based telediagnostic platform in patients with myocardial infarction and atrial fibrillation improves exercise tolerance, shows a trend toward improved left ventricular systolic function, and reduces the number of atrial fibrillation recurrences. This approach may be used to optimize rehabilitation strategies in this patient population.