Association for Public Policy Analysis & Management

Background: Health screening provides information on disease risk and diagnosis, but whether this promotes health is unclear. We estimate the behavioral responses to and impact of the National Health Screening Program in Korea for diabetes, obesity, and hyperlipidemia. In this setting, information on disease risk and prompting for a secondary examination vary at different biomarker thresholds.

Method: We apply a regression discontinuity design to estimate differences in outcomes at these thresholds, using administrative data that includes medical claims, biomarkers, and surveys of health behaviors for more than 350,000 baseline screening participants over four years after screening.

Results: We find evidence for increased diabetes medication and weight loss around the high risk threshold for diabetes, where information on disease risk is combined with prompting for a secondary examination and subsequent medical treatment. However, we do not find differences around other risk classification thresholds, where information on disease risk is not combined with further intervention. Specifically, we find that those who are just above the blood sugar 126 threshold who are classified as high risk for diabetes and are prompted for a secondary examination take more medications for diabetes and exhibit reduced BMI and waist circumference in the subsequent screening. On the other hand, we find little to no differences around the medium risk threshold for diabetes, as well as the risk classification thresholds for obesity and hyperlipidemia where intervention is not provided beyond provision of the screening report.

Implications: Our results suggest that information itself might not be sufficient to lead to behavioral changes, and that further intervention in addition to the information may increase the marginal benefits of screening.