Goetzel RZ, Anderson DR, Whitmer RW, Ozminkowski RJ, Dunn RL, Wasserman J; Association for Health Services Research. Meeting.

Abstr Book Assoc Health Serv Res Meet. 1998; 15: 129.

The MEDSTAT Group, Ann Arbor, MI 48108, USA.

RESEARCH OBJECTIVE: To estimate the impact of a broad range of self-reported health behaviors and risk factors on health care expenditures, controlling for other measured risk and demographic factors. STUDY DESIGN: Study subjects completed an initial health risk appraisal survey to report information on health risks related to body weight, smoking status, drinking habits, exercise habits, nutrition habits, depression. and stress. Biometric information on blood glucose and cholesterol levels was obtained, and blood pressure was measured. Based on this information, subjects were classified in terms of their risks of having poor health outcomes, using a proprietary classification scheme developed by StayWell Health Management Systems, Inc. Logistic regression models were then used to estimate the probabilities of incurring any medical expenditures during the three year period after subjects completed their initial health risk appraisal (HRA). Ordinary least squares regressions were used to estimate the magnitude of those expenditures (in log terms) for those who incurred any expenditures. After applying a smearing estimate to obtain predicted expenditures in actual (rather than log) dollars, the predicted probabilities of incurring any expenditures were multiplied by predicted expenditures to obtain estimates of the total medical expenditures for the study subjects. A total of 46,026 employees were followed during the three-year study period. These employees came from six large health care purchasers and included those age 18-64 who completed an HRA during the period 1990-1995 and had at least six months of claims experience available post-HRA. These participants contributed 113,963 person-years of experience. PRINCIPAL FINDINGS: Subjects at high risk for poor health outcomes had higher expenditures than subjects at lower risk, for the following risk factors: depression (70% higher expenditures), high stress (46%), high blood glucose (35%), extremely high or low body weight (21%), former (20%) and curren (14%) tobacco users, high blood pressure (12%), and poor exercize habits (10%). These same risk factors were found to be associated with a higher likelihood of having extremely high (outlier) expenditures. Finally, subjects with multiple risk profiles were shown to have higher expenditures than subjects without these profiles for the following diseases: heart disease (228% higher expenditures), psychosocial problems (147%), and stroke (85%). CONCLUSION: Common modifiable health risks are associated with higher medical expenditures. The economic impact of these risk factors had not been measured for such a large, varied sample in earlier research. IMPLICATIONS FOR POLICY, DELIVERY OR PRACTICE: Results suggest substantial health care savings may be acheived by investing in health promotion programs that successfully modify poor health habits.

Publication Types:

• Meeting Abstracts

Keywords:

• Costs and Cost Analysis
• Demography
• Health Behavior
• Health Expenditures
• Health Promotion
• Health Services Research
• Health Status Indicators
• Logistic Models
• Research
• Risk Factors
• economics
• hsrmtgs

Other ID:

• HTX/98619669

UI: 102234233

From Meeting Abstracts