Literacy, health, and health services use in a nationally representative sample

The relationship of adult literacy and health in the United States is a topic of growing interest for researchers, health care providers, and policy makers. Standard health status items have not been included in existing literacy surveys, nor have literacy measures been included in existing health surveys, limiting our understanding of the health/literacy relationship in a nationally representative sample. This study used the non-standard health items in the 1992 National Adult Literacy Survey (NALS), a nationally representative literacy data set, and the standard health items in the 1996 Medical Expenditures Panel Survey (MEPS), a nationally representative health data set, to investigate the health/literacy relationship, particularly considering mutual associations between socioeconomic status and race/ethnicity. As the MEPS does not contain a direct literacy measure, a proxy literacy variable was created from the overlapping demographic variables available in both data sets.;Though the strength and nature of the health/literacy relationship varied by health status measure, literacy was positively associated with some measures of health status even among adults who were equivalent in socioeconomic status, race/ethnicity, and other relevant characteristics. Thus, literacy is more than just another problem in a constellation of negative factors (poverty, minority status, low education) that adversely affect health. Better literacy is associated with better health, and poorer literacy is associated with poorer health, even when other positive or negative factors are controlled. This relationship was most clear for the work-impairing condition variable, the NALS health status index score, and the MEPS self-reported health status item. Literacy was associated with some measures of health status across a gradient, but the relationship was stronger beneath the "inadequate" threshold than it was above that threshold. Including literacy in some predictive models of health decreased the explanatory power of SES and race/ethnicity variables and, thus, provides new insights into health disparities. Some literacy-related health benefits were stronger for Whites compared to African Americans and for non-Hispanics compared to Hispanics. Unexpectedly, literacy was not significantly positively associated with preventive services use in this sample.