Application of Narrative and AI-Assisted Follow-Up After Voluntary Medical Male Circumcision: Multicenter, Double-Blind, Prospective, Randomized Controlled Trial.

Background: Postoperative anxiety following voluntary medical male circumcision (VMMC) poses a significant health challenge, with limited telemedicine access and inadequate communication compromising recovery and adherence. Narrative-based interventions have shown promise in reducing psychological distress in other contexts, and large language models may enable automated follow-up, but their role in VMMC care remains underexplored. Objective: We evaluated the effect of a narrative-enhanced tool (NET) on anxiety, sleep quality, quality of life, and pain management and identified risk factors for postoperative anxiety. We also assessed the feasibility of an artificial intelligence-assisted consultation (AAC) system in improving follow-up efficiency. Methods: From October 1, 2023, to April 29, 2024, patients aged ≥15 years undergoing VMMC were recruited and randomized 1:1 to a standardized risk tool (SRT) or NET group. In addition to the routine postoperative communication, the NET group received a narrative video highlighting positive recovery experiences. Both groups accessed an AAC chatbot for automated follow-ups. Primary outcomes were anxiety levels measured by the 7-item Generalized Anxiety Disorder scale (GAD-7), sleep quality measured by Pittsburgh Sleep Quality Index, quality of life measured by 3-level EuroQoL 5D questionnaire, and pain levels measured by Numerical Rating Scale. Secondary outcomes included analgesic use, satisfaction, and health care worker efficiency. Repeated measures ANOVA assessed trends and regression identified risk factors for anxiety. Results: Between October 1, 2023, and April 29, 2024, 671 eligible participants were enrolled, with 388 completing the 30-day follow-up (SRT group: n=189, mean age 26.21, SD 3.69 years; NET group: n=199, mean age 26.41, SD 3.56 years; P=.60). Both groups exhibited increased anxiety levels, diminished quality of life, and poorer sleep quality during the 30-day postoperative period. However, compared to SRT, the NET group demonstrated lower GAD-7 scores (7.06, SD 2.73 vs. 9.95, SD 3.50; P<.001), improved sleep quality (12.29, SD 3.57 vs 13.20, SD 3.54; P=.01), higher quality of life scores (0.87, SD 0.07 vs 0.84, SD 0.09; P<.001), more regular analgesic use (154/173, 89.02% vs 100/169, 59.17%; P<.001), reduced opioid consumption (5/173, 2.89% vs 25/169, 14.79%; P<.001), and higher pain medication satisfaction (4.21, SD 0.69 vs 3.76, SD 0.97; P<.001). Multivariate analysis identified SRT assignment, inability to recall opioid risk levels, hematoma, swelling, and pain as independent risk factors for elevated GAD-7 scores. Implementation of the AAC substantially reduced health care worker follow-up time (2.34, SD 1.95 min vs 7.85, SD 2.65 min; P<.001). Conclusions: The study demonstrates that narrative is effective in reducing anxiety, improving quality of life, and improving pain management post-VMMC. The integration of artificial intelligence into clinical follow-up protocols has the potential to enhance health care worker efficiency without compromising patient satisfaction.