Enhancing Outpatient Follow-Up Efficacy through AI-Assisted Medical Technology

Abstract Background The responsibility of managing chronic diseases and executing postoperative follow-ups imposes a significant workload on medical professionals. The current absence of unified and consolidated plans to improve follow-up efficiency led to our exploration of the impact of varying patient follow-up procedures on compliance levels, with an objective to develop a more effective follow-up strategy. Methods A cohort comprising a total of 14,012 patients from seven hospitals located across five provinces in China was subjected to a follow-up questionnaire delivered via one of four methods: a phone call, a text message, WeChat, or an AI-assisted procedure. Each method's efficacy was assessed based on the completion rate of the follow-up questionnaires. Further, comparisons were drawn among these methods considering diverse patient characteristics such as age and gender and considering the date and time of the follow-up. Results The average completion rate across all follow-up questionnaires was calculated at 86.6% (standard deviation: 19.56%). Among the methods, phone calls boasted the highest completion rate of 99.97%, followed by WeChat at 89.48%, and finally, AI at 85.55%. Regardless of the method employed, women displayed significantly higher completion rates than men (OR: 2.39, 95% CI: 2.202.60, P < 0.001). In the case of patients over 50 years of age, the AI method demonstrated a significantly higher completion rate than both text messages and WeChat, even though phone calls still dominated with the highest completion rate. As the number of questions in the questionnaire increased, the completion rates saw a consistent drop across all methods. Notably, participant compliance was higher during working weekday mornings than non-working mornings (OR: 1.92, 95% CI: 1.622.28, P < 0.001). Despite variations in age and gender, AI emerged as the most potent method, with completion rates hitting 71.47% in men aged between 40–49 years during non-working hours on a workday, and 94.79% in women aged under 40 years on non-working days. Conclusions While the AI follow-up method did not outperform traditional methods in terms of completion rate, it did represent the best approach for obtaining patient data across a wide range of demographics and times.