Knowing Is Not Half the Battle: Impacts of Information from the National Health Screening Program in Korea
Health behavior is an important determinant for health, especially in industrialized countries where morbidity and mortality are primarily related to chronic or lifestyle diseases (Cawley and Ruhm, 2012). For instance, the World Health Organization (2009) identifies that the leading causes of mortality and morbidity in high income countries are modifiable risk factors, including overweight and obesity, physical inactivity, high blood pressure, high blood sugar, high cholesterol, and tobacco and alcohol use. However, people often resist engaging in healthy behaviors that have positive future health outcomes. One explanation is that individuals have imperfect information about the benefits of healthy behaviors or about their own health status (Kenkel, 1991; Sloan et al., 2003). To address the lack of information, many developed countries provide public health screening. These policies assume that the information provided from screening will promote desirable health behaviors and early treatment that would prevent disease or reduce complications.
Our study setting is one of the world’s largest health screening programs, the National Health Screening Program (NHSP) in Korea. The NHSP provides free general health screening to the entire population aged 40 and over. The NHSP includes a variety of tests for health screening including diabetes, obesity, and hyperlipidemia, and is combined with a survey that collects information on health behaviors. We use data on a 2% random sample of the population from administrative data provided by the National Health Insurance Service (NHIS) in Korea, which includes more than 350,000 screening participants observed from 2009 to 2013.
Health screening provides information on disease risk and diagnosis, but whether this promotes health is unclear. We estimate the impacts of information provided by Korea’s National Health Screening Program by applying a regression discontinuity design around different biomarker thresholds of diabetes, obesity, and hyperlipidemia risk using administrative data that includes medical claims, biomarkers, and behavioral surveys over four years after screening. Generally, we find limited responses to disease risk information alone. However, we find evidence for weight loss around the high risk threshold for diabetes, where information is combined with active prompting for a secondary examination for diagnosis and treatment.