The Relationship Between Modifiable Health Risks and Future Medical Care Expenditures: The Korea Medical Insurance Corporation (KMIC) Study

Abstract

CONTEXT: The relationship between lifestyle risk factors, morbidity, and mortality is well established, but the relationship between lifestyle risk factors and medical care costs is not as well defined. OBJECTIVES: To determine the ability of modifiable biometric and lifestyle risk factors to predict future medical care costs. DESIGN: Prospective cohort study. SETTING AND PARTICIPANTS: Data on modifiable risk factors collected in 1992 and medical care costs collected in 1998 by the Korea Medical Insurance Corporation in South Korea. Data were examined for a final cohort of 78,728 men and 50,414 women enrolled in the health insurance plan from 1990 through 1998. MAIN OUTCOME MEASURES: Outcome measures included likelihood of any inpatient, outpatient, and total medical care costs and outlier costs; amount of inpatient, outpatient, and total medical care costs; and portion of total medical costs attributable to each risk factor through unadjusted and adjusted multivariate analyses. RESULTS: Baseline modifiable risk factors measured in 1992 (including lifestyle factors such as smoking, high body mass index, exercise, and biometric measures such as cholesterol, blood sugar, blood pressure, and urinary sugar) were important predictors of the amount of medical care costs incurred 6 years later in 1998, even after controlling for age, perceived health status, and each of the other modifiable variables. These risk factors were generally better predictors than nonmodifiable demographic risk factors, including income level and type of job. For men, lifestyle risk factors were associated with total costs that were 2.4% (for high blood pressure) to 16.1% (for former smokers) higher than among men without those risk factors. Biometric risk factors were associated with costs ranging from 9.2% (for cholesterol) to 38.2% (for positive urinary glucose) higher. For women, lifestyle risk factors were associated with total costs that were 2.5% (for exercise) to 6.4% (for current smokers) higher than among those without the risk factors. Biometric risk factors were associated with costs ranging from 10.2% (for cholesterol) to 60.4% (for positive urinary glucose) higher. For men, a cluster of six heart disease risk factors were associated with total costs 54.7% higher, and a cluster of three stroke risk factors were associated with total costs 22.2% higher than in men who had none of these risk factors. Modifiable risk factors accounted for 23.1% of medical costs for men and 8.7% for women. CONCLUSIONS: These results suggest that modifiable biometric and lifestyle risk factors can predict a moderate portion of future medical care costs. If these risk factors can be reduced, future medical care costs may be reduced.