Predicting physical activity adoption among minority women using structural equation modeling and participatory research

The purpose of this study was to explore individual, social and environmental correlates of physical activity adoption among African American and Hispanic or Latina women using a mixed-method approach. First, a conceptual model of individual, social and environmental factors that influence physical activity was developed by conducting a systematic review of the literature on psychosocial factors related to physical activity. Next, 18 African American and Hispanic or Latina women participated in an in-depth interview to gain a deeper understanding of the underlying factors within the Ecologic Model of Physical Activity framework that contribute to physical activity adoption among minority women. The transcribed interviews were coded and analyzed using thematic content analysis and a constant comparison approach. Women discussed individual level beliefs and attitudes about physical activity and the need for confidence, motivation and time. Women also discussed micro-environmental factors, including their physical neighborhood environments, and safety of and accessibility to physical activity resources, and emphasized caretaking for others and social support, factors in the meso- and exo-environments that demonstrate the linkages between physical environments and physical activity. Themes from in-depth interviews were used to refine the conceptual model and latent constructs included in the model. The model was then analyzed using existing objectively assessed physical activity and neighborhood environment data and questionnaires on body image, self-efficacy, motivational readiness, social support, home environment for physical activity, and perceived environment from the Health Is Power project (PI: Lee, 1R01CA109403). The final model included six latent constructs measured by 61 indicator variables: Measured neighborhood environment, perceived neighborhood environment, home environment, body composition and image, motivational readiness and social support. The proposed structural model fit was acceptable (χ 2=2924.7, df=2080, RMSEA=.035, CFI=.878). Body composition and image and motivational readiness (β=0.155, p=.035) and exercise self-efficacy and home environment ( β=-0.220, p=.001) were significantly associated, and there were significant direct effects on physical activity from exercise self-efficacy (β=-0.164, p=.043) and neighborhood environment (β=0.194, p\u3c.001). Findings from this study contribute knowledge about ecologic, multilevel mechanisms underlying adoption of physical activity among minority women. The resulting model may be used to guide future intervention development and physical activity research.