Keywords
How to Cite
Abstract
Background: Artificial intelligence (AI)-enabled telemedicine platforms are increasingly being developed to support chronic disease management through remote consultation, automated reminders, patient monitoring, personalized communication, and clinical decision support. However, the magnitude of their potential effects remains uncertain because outcomes depend on patient engagement, intervention effectiveness, and disease characteristics. Objective: This study aimed to estimate the potential impact of an AI-enabled telemedicine platform on treatment adherence, patient satisfaction, glycemic control, and blood-pressure management using an explicitly performed Monte Carlo simulation.
Methods: A probabilistic model was constructed to represent 10,000 simulated chronic-disease patient trajectories. Baseline adherence, patient satisfaction, HbA1c, systolic blood pressure (SBP), intervention effects, and patient engagement were sampled from predefined probability distributions. A reproducible random seed (20260812) was used. Patient engagement was modeled using a beta distribution and incorporated as a modifier of intervention effectiveness. Ten thousand Monte Carlo iterations were executed, and means, medians, standard deviations, 2.5th and 97.5th percentiles, and probabilities of achieving predefined improvement thresholds were calculated.
Results: The simulation estimated a mean improvement in treatment adherence of 7.51 percentage points (95% simulation interval: 1.98–13.22), an increase in patient satisfaction of 0.359 points on a 5-point scale (0.062–0.653), an HbA1c reduction of 0.492% (0.093–0.896), and an SBP reduction of 4.25 mmHg (0.58–7.97). The simulated probability of achieving ≥5 percentage-point improvement in adherence was 81.3%; ≥0.30-point improvement in satisfaction, 64.4%; ≥0.30% reduction in HbA1c, 82.5%; and ≥3 mmHg reduction in SBP, 74.1%.
Conclusion: The executed Monte Carlo model predicts potentially favorable effects of AI-enabled telemedicine on adherence, patient satisfaction, and selected chronic disease indicators. These findings are model-generated predictions rather than clinical observations and should be validated through prospective clinical research.