Neighborhood, physical activity and women's health: A multilevel analysis in Hermosillo, Mexico

Objective: To estimate the effect of neighborhood's physical and social characteristic in the levels of physical activity participation of women and the importance of physical activity in predicting objective and subjective health outcomes.

Methods: Data for this analysis come from a survey to women (25-54 years) residents in the city of Hermosillo and from publicly available information on the neighborhoods where they reported to have their regular residence. The survey is a cross-sectional study based on a probabilistic sample (n=1,306) using a systematic sampling procedure. Data were gathered through face-to-face interviews at the domicile of the respondents and through anthropometric measurements performed by medical personnel working for the Health Department of the State of Sonora. Multilevel logistic regression was used to estimate variations in low impact physical activity (LIPA) and overall physical activity (OPA), and the importance of physical activity levels in predicting individual health outcomes.

Results: There were significant variations between individuals in low impact and in overall physical activity in educations adjusted models. The z-value associated with the random coefficient variance estimate is significant (p=0.04) for OPA, indicating that participation in physical activity varied significantly across neighborhoods. In contrast, the z-value associated with LIPA was not significant, suggesting homogeneity. Living in the most socioeconomically deprived neighborhood was associated with a decreased likelihood of participating in overall physical activity.

Conclusion: These research findings support the need to focus on improving local environments to increase physical activity participation, and ultimately, health status of women living in urbanized areas.