Education, Knowledge, and the Production of Health: Evidence from Cancer Screenings

Several studies have documented a positive association between educational attainment and health, though the mechanisms through which education improves individual health are less clear. For example, the more educated could be healthier because they are better human capital with higher stock of knowledge (productive efficiency) or because they are better at selecting inputs used in health production (allocative efficiency). In practice, however, direct tests of the allocative efficiency hypothesis are rare as they require estimating the effect of education on both health knowledge and health behaviors; a relationship that is difficult to disentangle.

In this paper, I provide new evidence of the importance of allocative efficiency in the production of health by examining individuals’ responses to changes in cancer screening guidelines and how these responses vary by differences in education and health knowledge. Cancer screening increases early disease detection, which improves survival probability following a cancer diagnosis. In the U.S., standard cancer screening guidelines are published by two different organizations: the American Cancer Society (ACS) and the United States Preventive Services Task Force (USPSTF). Notably, the screening guidelines promulgated by these two organizations differ in their scope and are periodically updated. For example, the USPSTF recommends biennial mammograms for women over the age of 40, while the ACS suggests that women over 40 receive a mammogram each year. I take advantage of the variation introduced by these guideline changes to examine the role of allocative efficiency in health production.

Using data from the National Health Interview Survey (NHIS) and the Behavioral Risk Factor Surveillance System (BRFSS), I estimate difference-in-differences models to quantify the change in screening rates associated with a guideline change and how that change differs for those with more education and greater health knowledge. I find evidence to support the allocative hypothesis: 1) education is measure by years of schooling, which alone isn’t statistically sufficient enough to explain the discrepancy in the cancer screening behaviors. Higher educated peers are 0.1 percent (insignificant at p-value<10%) less likely to take pap smear and mammogram when both test were recommended to take less frequently; 2) given the guideline recommends against the PSA test, higher educated individuals are 0.9 percent (significant at p-value<5%) more likely to screen for PSA if physician recommends the test regardless the recommendation; however, given particular health knowledge about the advantage and disadvantage in taking the test, higher educated peers are 1.7 percent (significant at p-value <1%) less likely to be screened for PSA.

These findings indicate that both education and health knowledge are important contributors to the production of health. Besides the traditional view that more highly educated have better access to health care resources, it is increasingly the case that modern medical care requires patients’ knowledge and self-management ability to make efficient health care decisions. If education and health knowledge equip one with skills that are necessary to participate in the gains from medical advancements, then policies that promote improvements in these areas would be one solution to the increasing disparities in health outcomes.