A Proposed Linkage between the General Social Survey and the National Death Index

• Peter Muennig, Health Policy and Management
• Bruce Link, H&SS co-director, Sociomedical Sciences
• Awarded:  Spring 2008

Social factors—social support networks, access to social capital, job control, tempered emotional expressivity, and low levels of psychological stress, to name a few—appear to be stronger determinants of health than medical care.  Socio-economic disparities, for instance, are the leading preventable cause of death, accounting for 17.4 million quality-adjusted life years lost annually in the US.  Over the past four decades, researchers have learned a great deal about these disparities.  This research, however, has largely been conducted using large-scale datasets that were primarily designed to improve our understanding of traditional risk factors for disease, such as dietary factors or smoking.  Social determinants, though, mostly arise from socially-constructed phenomena, and researchers require a dataset with a rich source of both sociological and health variables to study them.

One way of obtaining such data is to link older sociological datasets to information on which subjects have died and which remain alive. Doing so not only provides a rich source of socio-medical data, but also converts cross-sectional datasets into prospective datasets.  Peter Muennig (Health Policy and Management) and Bruce Link (H&SS co-director) are working to link a long-standing and complete social science dataset—the General Social Survey (GSS)—to death certificate data in the National Death Index (NDI).  When linked to the NDI, the GSS will allow public health researchers to tackle many vexing questions surrounding the social determinants of health that have been out of their reach due to data limitations.  For example:  Do measures of pessimism, distrust, or civic disengagement predict future mortality?  Do measures of positive subjective well-being (e.g., happiness) predict mortality?  Are racial disparities in health explained by positive attitudes towards civil rights (i.e., support for the Equal Rights Amendment)?  Does an individual’s choice of low paying altruistic professions (e.g., social work), place persons at lower risk of disease than less desirable professions of comparable pay?

This linkage will be highly cost-effective, bringing 28 years of mortality follow-up data into the dataset in a matter of days.  The end product will be an unprecedented medical sociology dataset. With the support of an H&SS seed grant, Muennig and Link have shown that the linkage will work; of 52 records tested to date, 51 can be successfully linked to mortality data. Thus H&SS seed money has opened the door to obtaining the hundreds of thousands of dollars that will be required to ultimately merge these data. Drs. Muennig and Link are now taking the next step and applying for funding from the National Institutes of Health to pave the data highway for the next generation of social determinants of health researchers.