Better But Not Well: Mental Health Policy in the United States since 1950

7 Assessing the Well-being of People with Mental Illness

c h a p t e r 7 Assessing the Well-being of People with Mental Illness Over the past five decades, Americans have witnessed vast improvements in living standards. Incomes have more than tripled (adjusting for inflation). The share of U.S. families living in poverty has fallen by half. The life expectancy of adults has increased seven years since 1950. Most Americans have benefited from a rich array of new treatments for life-threatening diseases, and they are better protected against the financial costs of those treatments. The quality of most of the other goods and services we purchase has also improved markedly. This chapter examines whether mental health delivery has kept pace with these developments and whether people with mental illnesses have shared in this generally rising tide of fortunes over the past fifty years. As we have shown, the United States has invested heavily in treatment for mental disorders over the past five decades. There has been considerable innovation in the treatment of mental illnesses. By themselves, these investments and innovations would be expected to translate into improved well-being. At the same time, however, there have been profound changes in the nature of delivery of mental health services. Public mental hospitals have shrunk drastically. New private institutions, such as managed behavioral health organizations, have taken over many of the responsibilities of public mental health authorities. They have also replaced much of the private fee-for-service insurance system. Many observers have raised alarms that the development of new pharmacologic treatments has come at the cost of displacing other forms of therapy. Much of the popular discussion of the state of mental health suggests that the negative consequences of these institutional and therapeutic changes more than outweigh any gains from increased spending on services and innovation. In this chapter we will evaluate these conflicting claims through an assessment of the well-being of people with mental illness over time. Assessing the Well-being of People with Mental Illness 105 The ideal way to make such an assessment would be to study a comprehensive, longitudinal database. Using such data, we could examine whether, for example, Americans are better protected against the financial risks of mental illness, are receiving increasingly effective therapy, and are better supported in the event of a severe mental disorder. Unfortunately, no such single database exists. Instead, our strategy in making assessments over time is to combine information from multiple sources. Administrative data provide reasonably consistent information on the number of people served by the public mental health system in various settings. Epidemiologic surveys conducted at various times provide information on the number of people with specific types of mental illnesses. Numerous general health and medical surveys provide information on patterns of treatment and financing over time. Research studies offer assessments of the effectiveness of specific therapies. We assembled these various sources of information to assess the availability and quality of treatment over the last five decades in a series of steps. First, we estimated the size of the population with mental illness and divided that population into categories that, in a particular context, are likely to be meaningfully different in terms of treatment and living arrangements. In assessing changes in the quality of treatment received, we divided the population of people with mental illness by major disorder category. Second, we used information from surveys conducted at different times to calculate the proportion of these populations that received any mental health treatment at all. Third, we further subdivided the population according to the type of treatment received. For example, we used information from health care surveys to measure the share of people with depression who were receiving an antidepressant at each of several points. Fourth, we used the most recently available information from research studies to project the likely effect of a particular type of treatment on a person with a particular disorder. These data are rarely rich enough to permit a precise measurement of effectiveness. Rather, we divided treatments received into broad quality categories —likely to be highly effective, likely to be somewhat effective, no evidence of effectiveness, and likely to be harmful. We then extrapolated the probable effectiveness of observed patterns of treatment in the past on the basis of our understanding of treatment today. We used a similar process to assess changes in living conditions. We divided the population by the severity of their functional disabilities. We then focused mainly on those with severe mental illness. We identified and classified the...