Multiple-Impact Effectiveness of a State Supported Employment Policy Initiative

描述（由申请人提供）：我们提出了一项观察性研究，以检查国家政策倡议所产生的广泛影响，该倡议增加了马里兰州严重精神疾病（SMI）患者支持就业（SE）服务的基于证据的个人安置和支持（IPS）模型的可及性和可用性。我们将研究马里兰州公共精神卫生系统（PMHS）在促进IPS-SE服务在该州的推广方面的经验。我们的研究利用了IPS-SE服务可及性和可用性的跨时间和地理变化：(1)2002-2010年期间的时间变化；(2)马里兰州不同地理区域的地理变化。我们还研究了提供商在IPS-SE模型保真度方面的差异。我们建议研究这一国家政策举措在实际运行和影响中观察到的有效性。本研究将扩展IPS-SE的评价研究，超越目前在有限地点的小规模RCT试验中对疗效/有效性的关注。我们还建议将来自多个公共部门数据库的数据结合起来，以扩大SE影响研究的范围，超越目前对就业结果的研究重点，包括对心理健康状况的影响，以及对心理健康和躯体健康治疗成本的影响。促进IPS-SE服务实施和推广的政策倡议是由马里兰州精神卫生管理局（MHA）在2002年作为一个试点项目开始实施的。卫生部为培训服务提供者和监测提供者对IPS-SE模式的忠诚提供了资金。从2006年开始，向提供IPS-SE服务和达到保真基准的供应商提供财政激励。卫生部的这一举措增加了重度精神障碍患者获得IPS-SE服务的可及性和可获得性，2004年至2010年期间出现了重大增长，而且因地理区域而异。我们计划使用来自几个大型、明确的队列（1)马里兰州医疗补助接受者和(2)其他非医疗补助覆盖的PMHS服务用户）的重度精神障碍患者的数据，随访时间为2002-2010年。利用这些数据，我们将评估MHA政策对以下方面的影响：(1)IPS-SE服务的使用率，(2)IPS-SE服务使用率的种族和民族差异，(3)IPS-SE用户相对于非用户的心理健康状况，(4)IPS-SE服务用户与非用户的心理健康专业服务和其他健康服务的治疗成本，以及(5)IPS-SE服务用户与非用户的就业和收入。我们还将研究不同IPS-SE提供商保真度水平对这些结果的影响。为了估计这些政策影响，提出了各种回归分析方法，包括工具变量-广义矩量法（IV-GMM）程序，简化形式回归模型，基于个人长期数据池的模型，以及基于横截面回归序列的分析。我们提出了IV-GMM方法来解释影响选择使用IPS-SE服务的不可观察因素，以及我们对IPS-SE使用指标内质性假设的检验。

DESCRIPTION (provided by applicant): We propose an observational study to examine a broad range of impacts produced by a state policy initiative that increased accessibility and availability of the evidence-based individual placement and support (IPS) model of supported-employment (SE) services for persons with severe mental illness (SMI) in Maryland. We will study the experience of Maryland's public mental health system (PMHS) in facilitating the diffusion of IPS-SE services in the state. Our research exploits inter-temporal and geographic variations in accessibility and availability of IPS-SE services 1) over time in the period 2002-2010 and 2) across geographic areas of Maryland. We also examine variations across providers in fidelity to the IPS-SE model. We propose to study the observed effectiveness of this state policy initiative in its actual operation and impact. This study will extend evaluative research on IPS-SE beyond the current focus on efficacy/effectiveness in small-scale RCT trials of a limited number of sites. We also propose to combine data from multiple public sector databases to expand the range of SE impacts studied, beyond the current research focus on employment outcomes, to include impacts on mental health status, and on mental health and somatic health treatment costs. The policy initiative that promoted the implementation and diffusion of IPS-SE services, undertaken by the Maryland Mental Hygiene Administration (MHA), began as a pilot project in 2002. MHA provided funding for training service providers and for monitoring providers for fidelity with the IPS-SE model. Beginning in 2006, financial incentives were given to providers for offering IPS-SE services, and for meeting fidelity benchmarks. This MHA initiative increased accessibility and availability of IPS-SE services for persons with SMI, with major increases, occurring in the period from 2004 to 2010, that also varied by geographic area. We plan to use data on individuals with SMI from several large, defined cohorts 1) of Maryland Medicaid recipients and 2) of other non-Medicaid-covered users of PMHS services, followed over the 2002-2010 period. Using these data, we will assess the impacts of the MHA policy initiative on (1) the take-up rate of IPS-SE services, (2) disparities by race and ethnicity in the rate of IPS-SE service use, (3) the mental health status of IPS-SE users relative to non-users, (4) costs of treatment for mental health specialty services and other health services for users vs. non-users of IPS-SE services, and (5) employment and earnings of users vs. non-users of IPS-SE services. We also will examine the impact of varying IPS-SE provider fidelity levels on these outcomes. A variety of regression analysis methods are proposed for estimating these policy impacts, including instrumental variables-generalized method of moments (IV-GMM) procedures, reduced-form regression models, models based on pooling of data on individuals over time, and analyses based on sequences of cross- sectional regressions. We propose IV-GMM methods to account for unobservable factors influencing selection into use of IPS-SE services, as well as tests for our assumption of endogeneity of the indicator of IPS-SE use. PUBLIC HEALTH RELEVANCE: This proposal addresses one the focus areas Healthy People 2010 that is proposed to be retained for Health People 2020: Increase the proportion of persons with serious mental illness (SMI) who are employed. The research will examine the effectiveness of a state-wide initiative to increase the provision of evidence-based supported employment services for those with SMI in the state of Maryland. We assess the initiative's impact on multiple outcomes including supported employment service receipt, race and ethnic disparities in service receipt, employment, earnings, mental health treatment costs, somatic treatment costs, mental health status.