Preliminary Data
The impetus for the theme of this Center came from two sets of preliminary findings, which will be briefly described here and in more detail in the Preliminary Data Sections of Projects 1 and 2. We found that neighborhood composition was a major influence on health of Hispanics. These findings were generated by merging census data with two different data sets. The first is the Hispanic Established Populations for Epidemiologic Studies of the Elderly (EPESE) which is a population-based longitudinal study of 3,050 older Mexican-Americans living in the five southwestern states. The second data set is the Surveillance, Epidemiology and End Results (SEER) cancer registry, which is a population-based tumor registry currently covering approximately 14% of the U.S. population.
In both data sets, the prevalence (in EPESE) or incidence (in SEER) of cancers in Hispanics decreased as the percentage of Hispanics composing the census tract increased. For example, in SEER, compared to Hispanics living in census tracts with fewer than 20% Hispanic population, Hispanics living in census tracts with more than 60% Hispanic population had a 29% lower incidence of breast cancer, 25% lower incidence of colorectal cancer, and a 20% lower incidence of prostate cancer. In the Hispanic EPESE there was a more than three-fold difference in prevalence of cancer among the subjects as a function of the percentage of Mexican-Americans in their census tracts. The above analyses controlled for census tract rates of poverty and other socioeconomic characteristics.
There was a similar pattern seen with mortality during seven years of follow-up in the Hispanic EPESE, with Mexican Americans living in neighborhoods with very high percentages of Mexican-Americans experiencing a 36% lower hazard of death than did subjects living in neighborhoods with very low percentages of other Mexican-Americans. Surprisingly, those findings from the Hispanic EPESE were essentially unchanged after controlling for a measure of acculturation (language of interview) and whether the subject was an immigrant or born in the U.S. In other words, homogeneous Hispanic neighborhoods are associated with substantially better health, and this association is not explained by immigrant status or a measure of acculturation. It should also be noted that neighborhood poverty rate correlates with percent Hispanic; so the lower rates of cancer and lower mortality exist in spite of increased neighborhood poverty.
At the same time, homogeneous Hispanic neighborhoods are also associated with some adverse health indicators. For example, SEER data show that Hispanics living in neighborhoods with high percentages of other Hispanics are more likely to be diagnosed with cancer at a non-local stage, and survival after a diagnosis of cancer is poorer.
In summary, we have found relatively strong, significant effects of neighborhood on health in two studies, one among older Mexican-Americans in the Southwest and another among Hispanics covered by the SEER registries. These neighborhood effects provide us with a perspective with which to examine mechanisms responsible for the relative protection that Hispanic populations seem to have from some of the adverse health consequences of low SES.
