The purpose of this application is to investigate a number of health-related variables in middle-aged and older Mexican Americans. Data will be employed from the southwestern sample of the recently completed Hispanic Health and Nutrition Examination Survey (H-HANES). This is the largest study of the health of Hispanics. Moreover, it oversampled persons 45 to 74 enabling the performance of relatively detailed multivariate analyses for a number of subgroups (by age, gender, etc). Our findings will be generalizable to the Mexican Origin population of the five Southwestern states. Variables of primary interest include self-ratings of health, activity limitation due to health problems or disability, presence of hypertension, physical activity, depression, preventive health care utilization, smoking, and alcohol consumption. We propose to approach the analysis of the data within a conceptual model in which acculturation into the larger society is the key independent variable influencing health and mental health. Acculturation has been receiving increased attention in the literature on Hispanics as a potential determinant of health because of the stress it may involve or because it alters certain health behaviors (e.g., smoking, alcohol consumption, utilization of services). However, we argue that the association between level of acculturation and health is quite complex. We hypothesize that a number of factors """"""""buffer"""""""", or modify the influence of acculturation. These include age, gender, marital status, employment status, socioeconomic status and health insurance coverage. We plan to operationalize and examine this model utilizing a variety of multivariate statistical techniques. First, we plan methodological analyses with the acculturation scale and the two depression scales to examine their applicability to the various age groups in the sample. Our substantive analyses will utilize variations of Ordinary Least Squares multiple regression analysis when the dependent variable is continuous (e.g. total depressive symptoms, number of drinks consumed). We plan to utilize logistic regression analysis when the dependent variable is dichotomous (e.g., presence of hypertension, high depressive symptomatology). Modifier effects of variables will be assessed through subgroup analyses (e.g. by conducting separate regression analyses for persons age 65 to 74 and for younger persons) and/or by introducing multiplicative terms into the equations.
