Purpose: To address whether policy makers should adopt Supported Employment (SE) over Vocational Rehabilitation (VR) as an employment intervention for SSDI beneficiaries diagnosed with Serious Mental Illness (SMI). While Randomized Controlled Trials (RCT) have shown that SE improves employment outcomes for SMI clients compared to VR, the economic value of SE has not be assessed.
Method: Using decision analysis, the cost-effectiveness of SE compared with VR was assessed over a 10-year time horizon based on cost per work hour gained. For our cost-effectiveness analysis, we considered annual costs of VR ($10,141), SE services ($13,600) and SSDI unemployment stipends ($6,912) and for effects we considered total number of work hours for SE (405 hours) and VR (137 hours) per year. Both costs and effects were discounted at 3% per year. Initial probabilities of employment and enrollment were based on RCTs comparing SE and VR. To predict future employment outcomes, we used the work trajectories of patients enrolled in the New Hampshire Dual-Diagnosis study to calculate the probabilities of staying employed for 10 years among those who had jobs and the probabilities of becoming employed among those who did not secure jobs initially. We made the following assumptions: 20% of clients immediately become completely dependent on SSDI, 20% of clients who are enrolled in services and unemployed will drop out every year, and the probability of steady employment and late employment is the same for SE and VR.
Results: The estimated number of work hours for SE was 3067 hours compared to 1013 hours for VR services with an additional cost of $44,000 for SE services over 10-year time horizon. Sensitivity analyses reported that the decision was dependent on annual number of work hours. If the VR work hours exceeded 414.9 hours per year, then SE would no longer be the cost effective option. This would require more than doubling the effect of VR services, which seems unattainable with existing resources
Conclusions: Compared to VR services, SE appears to be cost-effective; these results are sensitive to the effect of total work hours. Although SE programs are more costly than VR over a 10-year period, the value of SE appears favorable. This suggests that SE should be considered as an alternative to widely disseminated VR programs.
Candidate for the Lee B. Lusted Student Prize Competition