Social
Assessing the Measurement Properties of Neighborhood Environment Scales in Rural Alabama Mahasin S. Mujahid* Khin Oo Juan Cabrera Stephanie T. Broyles Tené T. Lewis Viola Vaccarino Apurv Soni Vasan Ramachandran Mahasin S. Mujahid
Despite robust evidence that neighborhood environments influence disease risk, most studies have employed neighborhood measures in urban settings. To assess the validity of neighborhood scales in rural settings, we examined the psychometric and ecometric properties of survey-based measures of physical and social environments in rural neighborhoods (i.e., census tracts) in Alabama. Data were obtained from the Risk Underlying Rural Areas Longitudinal (RURAL) Study, a cohort study designed to investigate the heart and lung health of rural dwelling adults aged 25-64 years of age in four southeastern states (Alabama, Mississippi, Louisiana, and Kentucky). A preliminary Alabama subsample collected in 2021-2022 includes 657 residents across 22 census tracts, with a mean of 30 participants per tract (SD = 19). The mean age was 48.7 (SD=10.9), 73.1% were assigned female sex at birth, 74.6% self-identified as Black, and 31.4% held a high school diploma or GED. Using a mHealth app, we collected six dimensions of neighborhood environments: aesthetic quality, walkability, access to healthy foods, safety, social cohesion, and activities with neighbors. Cronbach’s alpha was calculated to assess the internal consistency across scale items within these six dimensions. Intraclass correlation coefficients (ICCs) and neighborhood reliability were calculated to assess the ecometric properties of these scales. Cronbach’s alpha coefficients indicated good internal consistency ranging from 0.74 to 0.88. Neighborhood reliabilities indicated good reliability, ranging from 0.75 (activities with neighbors) to 0.89 (safety). However, ICCs were lower, ranging from 0.08 (activities with neighbors) to 0.20 (safety), possibly due to the larger square mileage of rural areas compared to urban areas and the lower population densities. Overall, these measures have good psychometric and ecometric properties and will be investigated in relation to rural health outcomes in future studies.