Neighborhood Environment and Physical Activity Among Persons with Dementia

Background and Purpose: Physical activity is a well-established, non-pharmacological strategy for delaying further cognitive and functional decline in persons with dementia (PWD). Yet, many community-dwelling PWD do not engage in sufficient physical activity, highlighting a critical gap in public health efforts to support this vulnerable population. Guided by the Social-Ecological Model, which emphasizes the multifaceted and layered influences on health behaviors, neighborhood environments may serve either as key enablers or critical barriers to promoting physical activity among PWD. Despite this theoretical foundation, the specific neighborhood-level factors that influence physical activity in this group remain largely unexplored. A handful of existing studies have been primarily done in qualitative research design with region-specific, small samples. To address this gap, our study merges a large, nationally representative cohort of older adults that include PWD with multiple contextual datasets to examine how diverse aspects of the neighborhood (social, economic, physical, and built) are associated with physical activity levels among PWD living in communities across the U.S.

Methods: We used data on PWD from the 2006-2018 waves of the Health and Retirement Study, focusing on the first wave in which they were identified as having dementia (N=3,216). Dementia classification was based on the validated Langa-Weir algorithm, which incorporates cognitive test performance and proxy-reported information. Physical activity was operationalized as a weighted score (range: 0–31) reflecting self-reported frequency of light, moderate, and vigorous activity. Neighborhood social conditions were assessed using self-reported measures of neighborhood cohesion and safety, as well as interviewer-observed neighborhood disorder. Economic, physical, and built environmental characteristics were drawn from census-tract-level data for the corresponding year, including neighborhood affluence and deprivation (American Community Survey), annual average temperature (GridMet), fine particulate matter (PM₂.₅) concentration (U.S. Environmental Protection Agency), and park area (NanDA). We estimated linear regression models adjusting for sociodemographic covariates (i.e., age, sex, race/ethnicity, marital status, education, wealth, and urban/rural status) to predict physical activity levels by each neighborhood variable.

Results: PWD residing in neighborhoods with greater socioeconomic affluence (B=0.45; 95% CI: 0.07, 0.84) and stronger social cohesion (B=0.70; 95% CI: 0.32, 1.09) had significantly higher levels of physical activity. In contrast, the presence of any neighborhood disorder was associated with reduced activity (B=–1.23; 95% CI: –2.09, –0.38). Other neighborhood environmental characteristics, including air quality, temperature, safety, and park area, were not significantly associated with physical activity.

Conclusions and Implications: Our findings underscore the importance of the neighborhood socioeconomic environment in shaping physical activity behaviors among PWD. Features such as affluence and social cohesion may exert downstream effects by facilitating more supportive and socially engaging environments that encourage physical activity. These results offer robust empirical support for the Social-Ecological Model, affirming the role of broader contextual determinants in influencing health behavior among cognitively impaired older adults. Future interventions aimed at increasing physical activity among PWD should consider neighborhood-level strategies that promote social cohesion and reduce visible disorder, alongside traditional individual-focused approaches.