Helping (Some) SSDI Beneficiaries With Severe Mental Illness Return to Work

The significance of mental disorders as a source of disability has grown over recent decades. This has meant that people with severe and persistent mental illnesses have been making increasing claims on public disability programs like Supplemental Security Income (SSI) and Social Security Disability Insurance (SSDI). For example, about 28% of SSDI recipients have a mental disorder (1). People with severe mental illness represent one of the fastest growing segments of the SSDI population. Policy makers greet these facts with both alarm and suspicion. Alarm because the numbers are large and the financial claims made on SSDI amount to tens of billions of dollars annually. Suspicion because identifying disability in people suffering from mental disorders is viewed as more subjective and hence more prone to manipulation. These concerns arise at a time when the SSDI Trust Fund is projected to be exhausted in 2016.

Proposed policy responses to the growth in SSDI beneficiaries with severe and persistent mental illnesses take three general forms: 1) expand the use of programs that help already disabled individuals return to work through treatment and vocational rehabilitation; 2) implement measures that would tighten criteria for program entry and continued participation; and 3) develop early intervention/prevention efforts to reduce the demand for SSI and SSDI. Based on a variety of small-scale clinical trials, the mental health field has been bullish on the first approach. Advocates see the combination of evidence-based clinical care and supported employment as the platform for returning people to work. The recommended approach to clinical care derives from the well-established chronic disease model, and the best-studied supported employment model is individual placement and support (2). Both of these approaches were used in the Social Security Administration’s Mental Health Treatment Study (MHTS). The article in this issue by Drake and colleagues reporting on the MHTS (3) is critically important as a rigorous evaluation of the favored approach. It was conducted by a team made up of some of the strongest proponents of individual placement and support (Drake and Bond) and a group of highly skilled evaluators. The results obtained by the MHTS are enormously instructive and offer a cautionary tale.

The MHTS compared the clinical and vocational outcomes of two randomly assigned conditions. Evaluators compared usual care with free access (no cost sharing and no care management other than clinical case management) to evidence-based clinical care in combination with individual placement and support. The intervention was implemented with a reasonable degree of fidelity and minimal patient dropout. The study produced the following key results. First, only 14% of SSDI beneficiaries contacted about joining the study agreed to enroll. Compared with those who did not participate, those who did were younger, had been on SSDI for less time (8 years compared with 12 years), and were more likely to have attempted to get work in recent periods. Thus, the MHTS was conducted with a relatively highly motivated segment of those with severe mental illness on the SSDI rolls; in many respects this represents a best-case scenario. Second, the intervention group was significantly more likely than control participants to engage in both paid and competitive employment (60.5% compared with 40.3%). Third, the level and duration of employment was also significantly greater for the intervention group. The intervention group experienced 6.2 months of employment in various jobs over the 2 study years, compared with 3.7 months for the control group. Weekly hours of work for those employed were about 12 hours for the intervention group and 7.6 hours for the control group. Fourth, the intervention group incurred significantly fewer overnight hospital days and mental health-related emergency department visits.

The implications of these key findings are several. First, although the objective of increased workforce participation was achieved, few participants realized sufficient levels of work activity to exit SSDI. In fact, nearly no one worked enough to come close to program work ceilings, so the well-known SSDI work disincentives were clearly not a binding constraint on the work activities of MHTS participants. The intervention produced an expected level of earnings: $1366 over 2 years, compared with $227 for the control group, or a difference of $1139 (my back of the envelope calculation, based on the probability of paid competitive employment times the expected months worked times the expected earnings per month). This is a meaningful amount, but not one large enough to make a real difference in the lives of these vulnerable people. Finally, there is evidence that the intervention produced some cost offsets in hospitalization and emergency department spending.

What, then, are the take-home messages for policy and practice? The results replicate earlier studies and remind us that individual placement and support is the best we have, and when combined with other evidence-based clinical care, it produces meaningful levels of work activity in a highly disabled population. People with severe mental illness often want to work and to feel socially included, and the MHTS demonstrated an ability to improve those outcomes. The MHTS also establishes strongly that the best we have will have only a small effect on employment and earnings and essentially no effect on exit from SSDI into mainstream economic life—and this result comes from a study focusing on the most motivated 14% of SSDI recipients with a severe mental illness.

Where does that leave us? While there are no doubt mismatches between SSI/SSDI program participants and disability levels, the mismatches occur in both directions; some less needy find their way into the program, and many disabled people are not receiving appropriate income support. Evidence suggests that simply tightening the rules is unlikely to result in improved matching (4). Recent small-scale studies suggest some potential for early interventions to attenuate the disabling effects of severe mental illness and thus possibly blunt the growth in SSDI participation (5). These studies also rely on a mix of evidence-based clinical care and individual placement and support. Thus, perhaps the most profound implication of the MHTS is that it may be time to give greater priority to early intervention and to shift our focus to “front door” efforts to keep people at work while in the early stages of their illness. Doing so presents a new set of challenges that requires addressing ethical questions regarding privacy, new approaches to financing such interventions (as they do not fit comfortably with existing insurance arrangements), and infrastructure development related to health care information technology that does not currently exist or is not structured to facilitate early identification and tracking.

The MHTS represents the best in rigorous policy-relevant evaluation research. It is persuasive; it teaches us much; and it humbles and redirects us in the process.