Mental Disorders, Private Insurance, and SSDI Benefits

该项目将详细说明并估计因精神障碍而领取有限责任保险的员工从私人长期残疾（LTD）福利向社会保障残疾保险（SSDI）福利过渡的计量经济模型。使用来自个人有限责任公司索赔的数据，过渡到SSDI福利的概率将建模为以下函数：1)员工个人、雇主和残疾特征；2)有限公司政策规定；3)雇主提供的精神健康保险；4)精神卫生治疗资源的可得性；5)雇主的残疾管理工作；6)州际间SSDI和其他补偿计划的可用性差异。此外，还将根据个人索赔，将SSDI费用数据汇总到雇主年水平，以估计每年每位员工（有限责任公司保险）的SSDI费用模型。该项目的数据将涉及约1400名雇主及其50万名受有限公司保险覆盖的员工。行政数据、索赔数据、公共资源数据以及对这1400名雇主中278名雇主的调查数据将进行分析。计量经济学方法将包括风险（持续时间）回归以及多重风险模型（其中重返工作岗位和向SSDI福利过渡是相互竞争的风险）。雇主层面成本的估算将使用曼宁和穆拉希最近提出的医疗支出建模方法。对缺失数据进行估计的几种方法也将作为将我们的雇主调查数据与研究中所有1400名雇主的信息相结合的策略进行测试。研究中获得的经验估计将提供关于精神健康福利变化的可能影响的信息，以及精神障碍受益人的SSDI成本持续快速上升的福利和补偿计划的变化。将对SSDI接收和成本的影响，或增加管理护理/切割覆盖范围，增加地理上可获得的精神卫生专业护理，以及雇主的残疾管理工作的影响进行估计。

This project will specify and estimate econometric models of transitions from private, long-term disability (LTD) benefits to Social Security Disability insurance (SSDI) benefits by employees who are on LTD as a result of a mental disorder. Using data from individual LTD claims, the probability of transitioning to SSDI benefits will be modeled as function of: 1) individual employee, employer, and disability characteristics; 2) LTD policy provisions; 3) mental health coverage provided by the employer; 4) availability of mental health treatment resources; 5) employers' disability management efforts; and 6) interstate differences in availability of SSDI and other compensation programs. Data on SSDI costs will also be aggregated over individual claims to the employer-year level for estimating models of SSDI costs per employee (covered by LTD insurance) per year. Data for the project will pertain to approximately 1,400 employers and their 500,00 employees covered by LTD insurance. Administrative data, claims data, data from public sources, and survey data on 278 of these 1,400 employers will be analyzed. Econometric methods will include hazard (duration) regressions as well as multiple hazard models (where return to work and transition to SSDI benefits are the competing risks). Estimation of employer-level costs will use the methods for modeling health expenditures recently proposed by Manning and Mullahy. Several approaches for estimation with missing data will also be tested as strategies for combining our employer survey data with information for all 1,400 employers in the study. The empirical estimates obtained in the study will provide information on the possible implications of changes in mental health benefits and changes in benefit and compensation programs for the continuing rapid rise in SSDI costs for beneficiaries with mental disorders. Estimates will be derived of the impacts, on SSDI receipt and costs, or increased managed care/carve-out coverage, of increased geographic availability of mental health specialty care, and of employers' disability management efforts.