Objectives: Using data from the nationally representative Medical Expenditures Panel Survey (MEPS), we explored the extent to which health care utilization and health risk-taking, together with previously examined mediators, can explain the education–health gradient above and beyond what can be explained by previously examined mediators such as age, race, and poverty status. Method: Health was measured using the Physical Component Score (PCS) from the Medical Outcomes Study 12-Item Short Form (SF-12). Educational attainment was self-reported and categorized as 1 (less than high school), 2 (high school graduate or GED), 3 (some college), 4 (bachelor’s degree), and 5 (graduate degree). Results: In bivariate analysis, we found systematic graded relationships between educational attainment and health including, SF-12 PCS scores, self-rated health, and activity limitations. In addition, education was associated with having more office visits and outpatient visits and less risk tolerance. Those with less education were also more likely to be uninsured throughout the year. Multivariate regression analysis suggested that adjustment for age, race, poverty status and marital status explained part, but not nearly all, of the relationship between education and health. Adding a variety of variables on health care and attitudes to the models provided no additional explanatory power. This pattern of results persisted even after stratifying on the number of self-reported chronic conditions. Conclusions: Our findings provide no evidence that access to and use of health care explains the education–health gradient. However, more research is necessary to conclusively rule out medical care as a mediator between education and health.