When Digital Fits and When Human Matters: Attachment, Trust, and User Profiles in Communication Formats in Primary Care from Face-to-Face to AI.

Abstract

Objective: This study examined how preferences for face-to-face, video, written, AI-chatbot, and AI-avatar consultations in the context of primary care vary between routine and emotionally sensitive contexts, how attachment orientations and trust in the diagnosis contribute to these preferences, and whether distinct user preference profiles can be identified.

Methods: In an online survey of 934 German adults, participants rated the appropriateness of five formats across routine and emotionally sensitive contexts. Predictors included trust in the diagnosis, attachment anxiety and attachment avoidance, perceived efficiency, gender, and technical resources. Multilevel models and cluster analysis were applied.

Results: Face-to-face was preferred for sensitive consultations, while video and written formats were acceptable for routine tasks. AI formats were less accepted overall but judged more appropriate when trust in the diagnosis was high. Trust in the diagnosis strongly supported acceptance of AI formats, while perceived efficiency was positively associated with higher acceptance of video and AI formats . Other predictors had minor effects. Three user preference profiles emerged: face-to-face-oriented, digital without AI-oriented, and broadly digital & AI-oriented.

Discussion: Format-dependent trust in the diagnosis outweighed efficiency in predicting acceptance (especially for AI), and attachment orientations offer an interpersonal perspective on these patterns.

Conclusion: Digital formats can support routine care, but sensitive consultations require face-to-face contact. Trust in the diagnosis is important for AI acceptance, while attachment orientation helps to interpret interpersonal differences. Patient needs are heterogeneous; services should provide transparent, efficient, and flexible options without replacing face-to-face care.