Dr. Meara Replies

To the Editor: Drs. Scott and Borenstein argue that Social Security disability caseloads reflect far more than Social Security policy. Indeed, this is something we argued in our commentary when we posited that one potential explanation for rising disability as a result of mental illness could be the shift from manufacturing toward service-sector employment, which may place those with mental illness at a particular disadvantage in the labor market. There are many reasons caseloads change. However, we think it is important to recognize—more explicitly than prior authors—the role of policies that do affect caseloads. For example, the stringent Continuing Disability Review policies of the early 1980s led to the disproportionate termination of disability cases among individuals with mental illness (1) . Furthermore, the evidence we presented of increased homelessness and eviction among former welfare recipients who were unsuccessful Social Security Income applicants (compared with successful applicants and nonapplicants) suggests that award policy is highly salient, since it may be applied in ways that exclude individuals for whom disability benefits are appropriate.

Drs. Scott and Borenstein argue that it would be more appropriate to identify health care funding mechanisms that would allow the extended treatment necessary to maintain work function in people with mental disorders. We, along with others, welcome the recent federal parity legislation that brings this goal closer. However, even with high-quality treatment, individuals with mental illness can and often do suffer disabling relapses or changes in their condition. Current eligibility policy and program rules do not recognize the episodic nature of mental illness. Drs. Scott and Borenstein also call for clarification of the appropriate workplace and treatment supports that can help contribute to more stable work among individuals with mental illness. There is strong evidence for the efficacy of supported work for individuals with psychiatric illness (2) .

Our commentary raised four questions regarding policy research and practice. First, for mental and substance use disorders, are the criteria for eligibility into income support programs we reviewed too stringent, too lax, or both? Second, are current mechanisms to prevent moral hazard in each program appropriate? Third, what are the mechanisms driving the rising caseload for mental disorders in each program over time? Finally, how might supported work or partial work waivers be used to help address disruptions to work that are unique to mental illness? In this final case, evidence from a randomized experiment would be invaluable for future program design.