Arline T. Geronimus photo

Measurement Error in Population Health Inequity Research using Novel Biomeasures

a MiCDA Research Project Description

Investigators: Arline T. Geronimus, John Bound, Colter Mitchell

Funding: National Institute on Aging, 2014-2017 (1 R01 AG 047167 01)

Researchers studying health disparities are increasingly interested in using telomere length (TL), a biomeasure of aging, as a marker of cumulative wear and tear on the body. Current research on racial/ethnic or socioeconomic differences in TL tends to be rudimentary, often based on highly select, racially homogeneous, or convenience samples, and lacking controls for socioeconomic variables. Leukocyte-derived DNA from fresh venous blood samples, the state-of-the-art for TL measurement, is a labor-intensive and very expensive collection process. Likewise, it is expensive to mount studies using large, representative or community-based samples to collect data on the broad range of socioeconomic, psychosocial, and environmental factors needed to advance interdisciplinary health disparities research.

However, several ongoing national data collections that include a broad range of social measures also isolate blood DNA and bank it in repositories after the cells have been transformed and immortalized using Epstein-Barr virus (EBV). Others collect, store and extract DNA from saliva. While neither of these sample types is ideal for gauging individual TL measurement, a critical question is whether DNA from these blood or saliva samples is valid for use in population research of TL. Validity hinges on whether the error introduced in these methods is random or systematic with respect to original TL or to populations of interest.

To address this question, we draw and analyze blood and saliva from 150 adult black, white, or Mexican-origin women and measure each woman's TL multiple times: using DNA directly isolated from fresh blood cells, DNA isolated from cells that are EBV-immortalized, and DNA extracted from saliva. By stratifying the sample along key axes of comparison – race/ethnicity, socioeconomic status, and neighborhood – and within-woman results by immortalized and fresh blood samples, and by fresh blood samples and saliva, we will directly gauge the effect that immortalization or tissue type has on the validity and reliability of using these biomeasures. Given that biosamples of these materials are currently available in large, representative data sets, our findings have implications for the pace of health disparities research.

Research Signature Themes:

Aging, Genetics, and Social Science
Health and well-being in later life: Disparities