Guideline-informed AI chat access and perceived mental effort in a simulated home-visit nursing task: a randomized controlled trial

Abstract

Abstract Background Home-visit nurses often make autonomous clinical judgments in patients’ homes, where access to colleagues and evidence-based resources is limited. Guideline-informed artificial intelligence (AI) systems may facilitate access to clinical information, but evidence of their effects on nurses’ perceived mental effort remains limited. This study examined whether access to a guideline-informed AI chat system was associated with perceived mental effort, task difficulty, and confidence in answer correctness during a simulated home-visit nursing task. Methods This web-based randomized controlled trial was conducted in Japan from March 9 to March 27, 2026. Home-visit nurses were randomly allocated 1:1 to AI or non-AI groups. Both groups completed a simulated task involving an older adult with chronic heart failure living alone. The AI group could use a NotebookLM system preloaded with publicly available clinical guidelines, whereas the comparison group was instructed not to use external resources. The primary outcome was perceived mental effort, assessed using a 9-point single-item rating based on Paas’ scale. The exploratory secondary outcomes were perceived task difficulty and confidence in answer correctness. The primary analyses used a complete-case approach and included participants who completed the task and provided post-task outcome data. Outcomes were analyzed using analysis of covariance adjusted for home-visit nursing experience, non-home-visit nursing experience, and educational attainment. Results Of 140 randomized participants, 105 completed the task and were included in the complete-case analysis: 53 in the non-AI group and 52 in the AI group. Compared with the non-AI group, the AI group reported lower perceived mental effort (adjusted mean difference − 1.13, 95% CI -1.81 to -0.45), lower perceived task difficulty (-0.74, 95% CI -1.37 to -0.12), and higher confidence in answer correctness (0.71, 95% CI 0.14 to 1.27). Conclusions Among participants who completed the task, guideline-informed AI chat access was associated with lower perceived mental effort and task difficulty and higher confidence in answer correctness than no access to external resources. These are complete-case, not intention-to-treat estimates; the study neither isolated the independent effect of AI nor evaluated clinical reasoning quality, accuracy, or safety. Trial registration UMIN Clinical Trials Registry, UMIN000061415. Retrospectively registered on 30 April 2026.