Integrating AI-powered Chatbots Into Patient Education Under the Digital Intelligent Precision Nursing Framework: A Qualitative Study (Preprint)

Abstract

<sec> <title>BACKGROUND</title> Current patient education is constrained by high nurse-to-patient ratios and heavy clinical workloads, leading to inefficient communication and inadequate content, urgent reform is needed. Generative artificial intelligence, particularly chatbots, presents significant opportunities to meet the demands for immediacy and continuity in patient education, leveraging natural language interaction, repeatable access, and round-the-clock availability. </sec> <sec> <title>OBJECTIVE</title> This study aimed to examine the current use, challenges, and recommendations for integrating AI-powered chatbots into clinical patient education from the perspectives of patients and nurses, and to delineate the requirements for nursing role transformation and capacity building, while, guided by the Digital Intelligent Precision Nursing framework, specifying pathways for responsible integration. </sec> <sec> <title>METHODS</title> This qualitative study was conducted from April to July 2025. Patients and caregivers were recruited via purposive sampling at a tertiary general hospital, and nurses were recruited through snowball sampling from six hospitals of varying tiers. Data were collected using a sociodemographic questionnaire and semi-structured, in-depth interviews. Interview recordings were transcribed verbatim and analyzed using inductive thematic analysis, with NVivo used for coding and theme construction. Sociodemographic data were analyzed descriptively. </sec> <sec> <title>RESULTS</title> This study enrolled 60 participants: 46 patients and caregivers along with 14 nurses. The mean (SD) age was 38.8 (11.4) years for patients and caregivers, and 34.64 (5.85) years for nurses. Thirty-four patients and caregivers (73.9%) had used AI-powered chatbots to obtain disease-related information, most commonly treatment-related information. All nurses reported using AI-powered chatbots to assist their work, however, only 4 (28.6%) had attempted to use them for patient education. Transparency and accuracy of chatbot outputs were the primary concerns shard by both parties. Four themes with 12 subthemes emerged: (1) acceptance and limitations of AI-powered chatbots for information retrieval; (2) conflicts and reconciliation mechanisms between AI-powered chatbots and traditional patient education; (3) improvement needs for applying AI-powered chatbot in patient education; and (4) challenges to the nursing profession and coping strategies. </sec> <sec> <title>CONCLUSIONS</title> While AI-powered chatbots offer transformative opportunities for patient education, their clinical integration still faces fundamental challenges. Drawing on the Digital Intelligent Precision Nursing (DIPN) framework, this study proposes targeted optimization pathway to enhance chatbot’s output accuracy, transparency, and personalization, and develops a human-AI-powered chatbot collaborative model to guide their use in patient education. Moreover, we outline nurse capacity-building and organizational support strategies to empower nurses to reshape traditional education models and to advance a coordinated, efficient, patient-centered educational paradigm. </sec>

Ähnliche Arbeiten