Our new paper led by the lab’s Postdoc Researcher Fengrui Jing, titled “From neighborhood contexts to human behaviors: Cellphone-based place visitation data contribute to estimating neighborhood-level depression prevalence in the United States”, is published in Cities, a top interdisciplinary journal focusing on urban planning and policy.
Abstract: The elucidation of neighborhood-level mental illness is pivotal to effective community need assessment and public health interventions. However, aggregating individual behaviors linked to mental health at the neighborhood level has proven to be a challenge. In this study, we collected place visitation data from extensive mobile phone records as a proxy measure of health behaviors to investigate whether and how the place visitation data can contribute to improving the estimation of nationwide neighborhood-level depression prevalence. Using nationwide place visitation data from 2019, we measured eight types of health behaviors at the neighborhood level in the United States, including positive and negative health behaviors (PNHB) and health service utilization behaviors (HSUB). The study revealed that visitations to different types of places of interest (POI) (i.e., fitness visitation, drinking place (alcoholic beverages) visitation, pharmacy visitation, general hospital visitation, and specialty hospital visitation) were significantly associated with neighborhood-level depression. Incorporating cellphone-based place visitation data (i.e., the proxy of health behaviors) into the models enhanced modeling fitness, with the model that included neighborhood context variables exhibiting the strongest fitness, followed by PNHB, POI features, and HSUB variables. These improvements are greater in models for self-reported mental health status compared to depression. The model fitness exhibits spatial differences, with smaller differences between actual and predicted values in urban areas (2.834) compared to rural areas (2.956) and the Midwest (2.212) compared to other regions. Overall, this study represents the first nationwide investigation of the role of cellphone-based place visitation data in estimating neighborhood-level depression prevalence. It expands a novel conceptual framework for explicating neighborhood-level depression prevalence by incorporating neighborhood-level health behaviors.
Keywords: depression; mobile phone data; health behavior; neighborhood; United States
Jing F., Li Z., Ning H., Lessani N., Qiao S., Li X., (2024). From neighborhood contexts to human behaviors: Cellphone-based place visitation data contribute to estimating neighborhood-level depression prevalence in the United States, Cities. https://doi.org/10.1016/j.cities.2024.104905
