Since the May 2nd publication of the NEJM paper on the Oregon Health Insurance Experiment (OHIE), there has been an uptick in commentary about the effects of poor mental health on physical health. Recall that the OHIE exploited a fortuitous randomized design to show, among other things, that Medicaid enrollment caused a substantial and statistically significant reduction in subjects screening positive for depression. What might this mean for their future health and, more broadly, how is current mental health related to future physical health and other outcomes? On Project Millennial, Adrianna McIntyre reviewed the OHIE depression results and literature that documents associations between reduction in depression and future improvement in functional ability, reduction in mortality, and increase in productivity. On The Incidental Economist, Bill Gardner drilled down into a study that relates mental illness to mortality. The mortality differences found in that study are huge. Men with mental illness die about 16 years earlier than their mentally healthy counterparts. The difference among women is about 12 years.
The mortality difference associated with having a mental illness is comparable to that associated with being a life-long smoker. [...] Mental illness kills. Sometimes by suicide, of which mental illness is a principal cause. But most of the excess deaths among the mentally ill are caused by diseases such as cardiovascular disease or cancer. In a sense, mental illness amplifies the risk or lethality of physical health problems. This occurs for many reasons. Mentally ill people are more likely to develop tobacco, alcohol, and substance abuse addictions. Mentally ill people also experience high levels of stress from the loss of jobs, marriages, and families. Chronic diseases such as diabetes require intensive daily self-care routines and mental illness undermines a patient’s ability to carry these out. The causal story relating mental and physical illness is complex. Each may promote the other and both kinds of illness are promoted by stress and other social determinants of health.For a grant proposal,* I recently summarized some of the outcomes for patients with serious mental illness (SMI), including major depression, schizophrenia, schizoaffective, bipolar, and post-traumatic stress disorders. That literature includes the study by Gale and colleagues that found that men with an SMI suffer substantial morbidity and die 25 years earlier than the U.S. average. A study by Li et al. found that mentally ill Medicare beneficiaries were less likely to be admitted to facilities capable of performing invasive cardiac treatments and were less likely to receive invasive procedures for heart attacks even when admitted to facilities that could perform them. For heart attack patients of the Veterans Health Administration, those with an SMI had higher adjusted 30- and 365-day mortality. SMI patients are also more likely to be readmitted to a hospital after being discharged, an indicator of poor follow-up or self care. A large study found that diagnosis of mental illness increased the risk of readmission for a heart attack to that equivalent the risk of those with diabetes, obesity, cerebral vascular disease and hypertension. Jiang et al. found increased risk of readmission after a hospitalization for congestive heart failure at three months and one year. Higher readmission rates for SMI patients is a consistent finding across many other studies. (See, for example, Saravay et al., Rathore et al., Burke et al., and Brennan et al.) Suffice it to say, there is an extremely large and consistent body of work demonstrating lower quality of care, poorer health outcomes, and higher mortality associated with mental illness. Consequently, an intervention, such as that studied by the OHIE, that substantially improves mental health has the promise of also improving future physical health. Of course, that's a causal connection one would want to test rigorously. It would be interesting indeed to see the OHIE investigators do so. * I am indebted to Mark Bauer, Jessica Gören, and Amy Rosen for assistance with this literature review. –Austin Frakt, Ph.D.