An individual's neighborhood provides important context for cardiovascular disease risk, and neighborhood characteristics as defined by U.S. census-level socioeconomic measures or the social and physical environment in which an individual lives have been associated with cardiovascular disease and prevalent cardiovascular risk factors. Obesity as a cardiovascular risk factor appears to be particularly influenced by an individual's neighborhood environment. The exponential rise in obesity prevalence over only three decades, with more than one-third of the U.S. population now having a body mass index (BMI)greater than or equal to 30 kg/m2, is largely consistent with behavioral and environmental rather than biological causal factors. Prior cross-sectional studies support a relationship between neighborhood socioeconomic status (SES), prevalent obesity and prevalent cardio-metabolic risk factors. However, these results are likely subject to self-selection bias, or a tendency for healthier, more financially secure individuals to live in areas of higher socioeconomic status. Few longitudinal studies have evaluated the relationship between change in neighborhood SES due to moving and obesity as a cardiovascular risk factor. Prior studies have been limited by use of self-reported measures of weight, use of intermediate surrogates for weight gain or cardiovascular health, small sample sizes, or limited numbers of movers. Therefore, we examined the impact of change in neighborhood socioeconomic deprivation with moving on weight change over 7 years in the Dallas Heart Study (DHS), a multi-ethnic sample of Dallas County residents aged 18-65. Body weight (kg) was measured in 2000-02 and at 7-year follow up (N=1835). Geocoded baseline and follow-up home addresses were linked to census block groups in Dallas County. A block group-level neighborhood deprivation index (NDI) was created (higher scores = more socioeconomic deprivation). Repeated measures linear mixed modeling with random effects was used to determine weight change relative to NDI change. Heckmans correction factor (HCF) was used to adjust for the non-random chance of moving to an area of higher NDI based on age, sex, race, education, income, employment, marital status, and home ownership. 49% of the DHS population moved within Dallas County during the 7-year study period. Blacks were more likely to move than whites or Hispanics (p<0.01), but there were no differences in baseline body mass index or waist circumference in movers vs. non-movers (p>0.05 for both). Adjusting for HCF, sex, race, and time-updated covariates (age, smoking, education, income, physical activity, length of residence), those who moved to areas of higher NDI gained more weight compared to those who remained at the same NDI or moved to lower NDI (0.69 +/- 0.29 kg per 1-unit NDI increase, p=0.02). Among those who moved, the impact of change in NDI on weight gain increased with time in the new neighborhood: mean weight gain per 1-unit NDI increase was 0.91 +/- 0.43 kg (p=0.03) for those living in the new neighborhood greater than median of 4 years, but not significant for those living in the new neighborhood less than or equal to median 4 years (0.55 +/- 0.39 kg, p=0.2). Thus, moving to a neighborhood of higher socioeconomic deprivation was associated with weight gain among DHS participants. Until economic and policy factors reduce neighborhood socioeconomic deprivation, more work is needed to identify individual or community interventions that reduce its adverse effects on cardiovascular health. In addition, we examined the relationship between neighborhood-level socioeconomic deprivation and prevalent diabetes in the DHS. Diabetes was defined by self-report, use of anti-hyperglycemic medication, or fasting glucose≥126 mg/dl. Logistic regression modeling was used to determine odds of prevalent diabetes for those in the highest versus lowest NDI tertile. In DHS, diabetes prevalence was 5%, 13%, and 16% across NDI tertiles (p<0.001). In modeling diabetes, we found a significant interaction between race and NDI (p=0.03); therefore, models were race-stratified. White, Hispanic, and black DHS participants in neighborhoods in the highest NDI tertile were up to seven times more likely to have diabetes than those living in the lowest tertile. In Whites and Hispanics, higher deprivation remained associated with a greater likelihood of diabetes after adjustment for age, sex, smoking, and education and was only attenuated after adjusting for income. In contrast, adjustment for confounders attenuated the relationship between NDI and diabetes among blacks. Residing in socioeconomically deprived neighborhoods is associated with prevalent diabetes among whites and Hispanics in DHS. These data suggest racial/ethnic disparities in cardio-metabolic risk within areas of higher socioeconomic deprivation in Dallas County.
Showing the most recent 10 out of 14 publications
