Citation
Gimm, G., Ireys, H., Gillman, B., & Croake, S. (2011). Impact of early intervention programs for working adults with potentially disabling conditions: evidence from the national DMIE evaluation. Journal of Vocational Rehabilitation, 34, 71–81.
Highlights
- This study’s objective was to examine short-term (that is, 6- to 12-month) impacts of the Demonstration to Maintain Independence and Employment (DMIE). DMIE was implemented in Hawaii, Kansas, Minnesota, and Texas with the aim of preventing or delaying people with disabilities from leaving the workforce and applying for Social Security Disability Insurance (SSDI) and Supplemental Security Income (SSI) benefits.
- The study was a randomized controlled trial examining DMIE’s impact on employment, hours worked, and benefits applications. Study data sources included the Ticket Research File (TRF), 831 File, and Uniform Data Set (UDS). Because DMIE recruitment methods, participant populations, and services offered varied substantially across states, the authors reported impacts separately for each state and for Minnesota and Texas pooled together.
- With few exceptions, DMIE did not achieve statistically significant impacts on employment, hours worked, or disability benefit applications.
- The quality of causal evidence is high for most outcomes, which means that we are confident that the estimated impacts are attributable solely to the DMIE. However, for some outcomes including employment in Kansas and Hawaii, the quality of the causal evidence is moderate, which means that we have confidence that the estimated effects are attributable at least in part to the DMIE, though other factors might also have contributed.
Intervention Examined
Demonstration to Maintain Independence and Employment (DMIE)
Features of the Intervention
The DMIE, which was authorized under the 1999 Ticket to Work and Work Incentives Improvement Act and funded by the Centers for Medicare & Medicaid Services, aimed to delay or prevent reliance on Social Security disability benefits through medical assistance and other supports. Hawaii, Kansas, Minnesota, and Texas received federal funding to design and test innovative programs to obtain this objective. Each state provided health care services beyond existing health insurance coverage, as well as discounted deductibles, premiums, and copayments. DMIE participants also received employment services such as vocational rehabilitation and service coordination and a personal case manager. DMIE enrollment occurred from 2006 through 2008 and varied by state. All DMIE services expired on September 30, 2009.
The age range for eligible participants varied slightly across states but was roughly 18 to 62. Eligible participants also had to work at least 40 hours per month, not have pending disability applications, and neither be receiving Social Security disability benefits nor have pending applications for benefits. Minnesota (888 treatment and 267 control group members) and Texas (888 treatment and 697 control group members) recruited low-income residents with severe mental illness from certain parts of the states. In Texas, participants also had to be uninsured and have a physical disability. Kansas (225 treatment and 275 control group members) recruited statewide participants with both physical and mental disabilities from its high-risk insurance pool and Hawaii (124 treatment and 60 control group members) recruited people with diabetes who lived in the city and county of Honolulu. Given the different recruitment strategies, participants varied across states in terms of age, marital status, ethnicity, college degree attainment rates, physical and mental capabilities, and employment characteristics.
Features of the Study
The study was a randomized controlled trial examining DMIE’s impact on employment, earnings, benefit applications and receipt, and health status. In Hawaii, 124 eligible participants were randomly assigned to a treatment group eligible to receive DMIE services and 60 were assigned to a control group that could not access DMIE services. The treatment and control group sample size in Kansas was 225 and 275, respectively; in Minnesota, it was 888 and 267, respectively; and in Texas, it was 888 and 697, respectively. Because the recruitment methods, participant populations, and services offered varied substantially across sites, the authors reported impacts separately for each of the four sites. The authors found that participants in Minnesota and Texas were similar enough to pool, so they also estimated impacts for the two states pooled together.
This analysis estimated regression-adjusted impact estimates based on outcomes measured 6 to 12 months after enrollment. The UDS data included demographic information (age, gender, and race); self-reported health status; baseline employment; percentage working in past month; and monthly hours worked. In addition, the analysis relied on Social Security Administration administrative data files: the TRF provided data on disabling conditions and the 831 file supplied disability benefit application information.
Study Sites
- Honolulu city and county, Hawaii
- State of Kansas
- Eight counties in Minnesota
- Harris County, Texas
Findings
- With few exceptions, DMIE did not achieve statistically significant impacts on employment, hours worked, or disability benefit applications in the four states examined.
- In Kansas, treatment group members had a 4.6 percentage point higher employment rate than control group members. In Hawaii, treatment group members worked 31.4 fewer hours, on average, than control group members. And participants in the pooled Minnesota and Texas sample were 2 percentage points less likely to submit an application for Social Security disability benefits than control group members.
Considerations for Interpreting the Findings
The authors cautioned that the analysis considered a short time frame—6 to 12 months after enrollment—and therefore might not represent eventual program effects. They also cautioned that the self-reported employment and hours worked measures in the UDS might be imprecise. Finally, the DMIE’s varying, somewhat narrow target populations limit the generalizability of study findings to the working-age adult population.
Causal Evidence Rating
The quality of causal evidence presented in this report is high for most outcomes and study sites examined. This means we are confident that the estimated effects are attributable to the DMIE, and not to other factors. However, DMIE attrition rates—and therefore the quality of the evidence produced by them—varied by outcome measure and state. There was high attrition for the percentage not working and average hours worked in Kansas and Hawaii. For these outcomes in these sites, the quality of causal evidence is moderate, which means we are confident that the estimated effects are attributable at least in part to the DMIE, but cannot rule out that other factors might have contributed.