Trajectories and Influencing Factors of Online Health Information-Seeking Behaviors Among Community-Dwelling Older Adults: Longitudinal Mixed Methods Study.

Background: The internet has emerged as a critical avenue for the public to access health information resources. As online health information is a key resource for older adults, the factors influencing their online health information-seeking behaviors (OHISBs) are crucial. Objective: This study aims to examine the trajectories and influencing factors of OHISB among community-dwelling older adults and to explore their attitudes toward this information in daily contexts. Methods: A longitudinal mixed methods study was conducted involving 346 older adults from three communities in Shandong Province, China. In the quantitative phase, data were collected at three time points: baseline (T1), 6 months postbaseline (T2), and 12 months postbaseline (T3). Latent class growth modeling and logistic regression identified trajectories and predictors of OHISB, while a cross-lagged panel model analyzed longitudinal relationships among digital health literacy (DHL), technology anxiety (TA), and OHISB. Subsequently, we conducted semistructured, in-depth interviews with 16 older adults from different OHISB trajectory subgroups using a descriptive phenomenological approach. Quantitative and qualitative results were integrated via triangulation. Results: Latent class growth modeling identified three OHISB trajectories: low-level declining group (92/346, 26.6%), medium-level stable group (187/346, 54%), and high-level declining group (67/346, 19.4%). Multivariate logistic regression identified household registration, education, income, chronic disease status, internet frequency and duration, online health information attitude and experience, DHL, and TA as significant predictors of OHISB trajectory. TA mediated the relationship between DHL and OHISB, with an effect size of 0.04 (SE 0.014, 95% CI 0.0123-0.068). Qualitative interviews with 16 participants revealed four themes: personal cognition, emotional experience, external environment, and behavioral choices. The quantitative study elucidated the pathway between DHL, TA, and OHISB, providing a data-driven foundation for the qualitative inquiry. In a complementary manner, the qualitative study revealed themes not captured quantitatively-particularly how personal factors influence behavior through attitudes-and elaborated on sources of social support and the role of the medical environment. Together, both methods convergently demonstrate that older adults' information behavior is not static but shaped by individual and environmental factors, resulting in distinct behavioral patterns. Conclusions: Both quantitative and qualitative findings clarified the developmental process of OHISB among older adults in communities and the important effects of DHL, TA, and risk perception on OHISB. Although self-efficacy, health anxiety, self-perceived aging, social support, and health care environment were not addressed in the quantitative study, they emerged as important factors shaping older adults' OHISB in qualitative interviews. Personalized intervention measures should target various OHISB trajectory characteristics and their influencing factors to enhance the health conditions of community-dwelling older adults.