ECONOMETRIC ASSESSMENT OF MENTAL HEALTH INDICATORS

点击翻译按钮获取中文摘要

This First Award proposal requests 4 years of funding to econometrically assess the reliability of both self-reported and clinically-based measures of mental health as they are typically used in empirical models of market behavior. The term 'reliability' encompasses two aspects of widely-used measures of mental health; the degree to which they are correlated with actual health and the extent to which they contain systematic reporting biases. The project will develop econometric models of earnings, utilization of services, and choice of provider which explicitly account for measurement error in mental health indicators. The long-term objectives of the research are twofold: first, to provide consistent estimates of the impact of mental disorders on earnings and on the use of mental health care services. Such estimates are essential for the proper evaluation of policies which affect the provision of services in the mental health care sector. The second major objective is to determine the extent to which measurement errors in mental health variables bias the effects of nonhealth variables on earnings and utilization patterns. These variables include, gender, race, income, and education. The roles played by these factors are also of interest to policymakers bearing on issues related to, accounting for the sources of growth in medical expenditures, the role of consumer information in the choice of provider, access to medical care for the disadvantaged, and the extent of race and gender discrimination in the labor market. The primary source of data for this project will come from the first wave of a longitudinal survey conducted by the National Institute of Mental Health's Epidemiological Catchment Area Program at five sites in the U.S. during the early 1980s. MIMIC (Multiple Indicator Multiple Cause) models provide the basic network for the analysis.