Beijing survey examines older adults’ perceptions of LLM-enabled healthcare quality

A cross-sectional study published in BMC Public Health examined perceived quality of large language model (LLM)-enabled healthcare among urban older adults in Beijing, China. Participants had used hospital-provided LLM-based medical devices. Researchers developed and pilot-tested an adapted, perception-based SERVQUAL questionnaire and analysed 298 valid responses, with a valid response rate of 91.9%. Structural equation modelling assessed associations between five service dimensions and perceived healthcare service quality.

Tangibility, reliability, responsiveness, assurance and empathy were each significantly and positively associated with perceived quality. Assurance and empathy had the largest standardised path coefficients, but formal comparison tests found no statistically significant differences between coefficients. The structural model explained 57.0% of the variance in perceived service quality.

The authors suggest the findings could inform age-friendly technical optimisation, hybrid online–offline service models and AI governance. For clinicians and service planners, the results offer preliminary evidence about older users’ experiences rather than evidence of clinical effectiveness or safety. The cross-sectional, perception-based design cannot establish causal relationships, and the surveyed population was limited to urban older adults who had already used these hospital-provided services.