Abstract
Loneliness is increasingly recognized as a critical public health concern, yet evidence on the associations between long-term residential conditions and loneliness trajectories remains limited, particularly in developing countries. Using nationally representative longitudinal data from the China Health and Retirement Longitudinal Study (CHARLS), this study identified loneliness trajectories among middle-aged and older adults and examined the associations between housing and neighborhood characteristics and trajectory membership. Growth mixture modeling (GMM) was applied to identify latent trajectory classes, and multinomial logistic regression was used to estimate these associations. We identified three distinct trajectories: stable-low (70.6%), increasing (17.6%), and decreasing (11.9%). Statistically significant associations with trajectory membership were more common for individual-level characteristics than for housing and neighborhood characteristics. Among the latter, certain housing structure types and higher housing quality were associated with a greater likelihood of belonging to the stable-low trajectory, whereas declines in housing quality were associated with membership in the increasing trajectory. Larger residential land area was associated with the decreasing trajectory among rural residents, while associations with floor area varied across demographic and regional groups. In addition, greater availability of neighborhood amenities and better transportation conditions were associated with stable-low trajectory membership among rural residents. Overall, these findings extend existing research by linking built environment characteristics to longitudinal patterns of loneliness. By identifying vulnerable subpopulations and highlighting the potential relevance of housing and neighborhood conditions, this study provides evidence that may inform future urban planning, community development, and mental health prevention strategies.