Assessment of a national health interview survey-based method of measuring community socioeconomic status

Purpose Increasingly researchers are interested in assessing the role of community socioeconomic status (SES) in poor health outcomes, above and beyond the influence of low individual SES. However, the feasibility of conducting these multi-level studies is often limited by restrictions on release of confidential identifiers for linkage to census data, resources for the linkage, and the availability of data sources with individual SES measures. This study assessed a new method of measuring community socioeconomic status (SES) that can be used with the publicly available National Health Interview Survey (NHIS) and preserves confidentiality and can be used with individual SES measures from the NHIS. Methods The associations between community SES and mortality from all causes and breast cancer in women were assessed in two samples: 1) deaths in 1987–1993 NHIS respondents linked to community SES measures developed with the new method; and 2) deaths in 1991 from the National Multiple Cause of Death Files linked to 1990 county-level census SES measures. The magnitude of crude mortality rates, direction of trend, and age-adjusted relative risk of mortality for low vs. high SES were compared in the two samples. Results Crude all-cause mortality and breast cancer mortality rates were similar in both samples in terms of magnitude and direction of trend. In both samples, as SES decreased, rates of all-cause mortality increased, whereas breast cancer mortality rates tended to decrease. Age-adjusted relative risks of mortality from all causes and breast cancer for low vs. high SES were similar in the two samples. Conclusions Similarity of associations between community SES and mortality from all causes and breast cancer in the two samples provides support for the validity of a new NHIS-based method of measuring community SES. Since the NHIS is a large, nationally representative survey with high response rates and low loss to mortality follow-up, this method represents an important resource for multi-level studies.