Cook 2008 — The Employment Intervention Demonstration Program Major Findings and Policy Implications

Summary

Synthesizes the published results of the Employment Intervention Demonstration Program (EIDP), a federally-funded multi-site randomized controlled trial (RCT) (7 sites, $n = 1{,}273$, 2-year follow-up) testing supported employment (SE) for unemployed adults with severe mental illness. Experimental participants were significantly more likely to achieve competitive employment ($55\%$ vs. $34\%$), work 40+ hours/month ($51\%$ vs. $39\%$), and earn more ($\$122$ vs. $\$99$/month), with effects increasing over time. Results were robust across diverse SE model variants, geographic sites, and participant populations.

Key Claims

• Main RCT result: $55\%$ experimental vs. $34\%$ control achieved competitive employment; advantage grew over 24 months, suggesting model effects are sustainable
• Job development is the lynchpin service: participants receiving job development were $\approx 5\times$ more likely to obtain competitive employment; those with no prior work history had virtually zero chance of competitive employment without it
• Clinical-vocational integration: models integrating clinical and vocational services (same agency, location, case file; $\geq 3$ staff meetings/week) produced $2.5\times$ greater competitive employment odds and $1.75\times$ greater odds of working 40+ hours/month
• Local unemployment rates matter but do not nullify SE: unemployment rate (UR) at site significantly affects outcomes; SE participants in high-UR areas outperformed controls in low-UR areas — SE is especially critical in weak local economies
• SSDI self-sufficiency barrier: only $4\%$ of Social Security Disability Insurance (SSDI) beneficiaries ($n = 450$) earned enough during 24 months to complete the trial work period and leave rolls, despite $69\%$ working during the study; SE helps people work but not reach financial self-sufficiency
• Costs: total program costs $\$2{,}000$–$\$6{,}000$/client/year; direct vocational service costs $\$500$–$\$2{,}000$/client/year; costs driven more by site nature than by SE model
• Demographic predictors: younger, female, Hispanic/Latino, better work history, high school (HS)/college education all predict competitive employment; African Americans more likely to hold any paid job but less likely to hold competitive job vs. Caucasians
• Clinical predictors: diagnoses other than schizophrenia, fewer prior psychiatric hospitalizations, lower psychiatric symptoms, and higher self-rated functioning predict better vocational outcomes; SE advantage persists regardless of clinical characteristics

Concepts Introduced or Extended

• Supported Employment
• DI Return-to-Work Milestones

Entities Mentioned

• Judith A. Cook
• Employment Intervention Demonstration Program

Quotes

"While employment services are successful in helping SSDI beneficiaries return to work, their labor force participation is not at a level that would enable them to leave the public disability income rolls. This may be due, in part, to work disincentives in the policies of public disability income support programs, particularly those that eliminate cash payments for those who earn enough to complete their trial work periods."

"Individuals re-entering the labor force in geographic regions with high unemployment are particularly vulnerable to continued unemployment without best-practice services."

"For most EIDP participants, work did not serve as a springboard to financial self-sufficiency, but instead as a route to membership into a social stratum of society referred to as the 'working poor'."

My Take

Cook 2008 is a well-organized summary of a major federal RCT rather than a primary data paper. The core result — supported employment substantially increases competitive employment — is well-established. The most policy-relevant finding for a disability economics audience is the SSDI self-sufficiency result: $4\%$ of beneficiaries left rolls despite most working. This is primarily a consequence of Social Security Administration's (SSA) substantial gainful activity (SGA) threshold and trial work period earnings cliff, not a failure of supported employment per se. The paper's discussion of work disincentives in SSDI benefit design is unusually direct for a rehabilitation journal article. The underlying EIDP sub-studies (Cook et al. 2005 in Archives of General Psychiatry; Cook et al. 2005 in American Journal of Psychiatry; Leff et al. 2005 in Psychiatric Services) are the primary quantitative sources; this article is most useful as an index and interpretive overview.