Self-medication of mental health problems: new evidence from a national survey

Health Services Research, Feb, 2005 by Katherine M. Harris, Mark J. Edlund

The high rate at which mental and substance-use disorders occur together has been well documented in epidemiological and clinical studies (Regier and Farmer 1990; Kessler, Nelson et al. 1996). For example, in the National Comorbidity Study, 51 percent of those who met criteria for a substance disorder at some time in their life also met criteria for a mental disorder at some point, and in the large majority of cases individuals reported that the mental disorder preceded the substance disorder (Kessler et al. 1996). Researchers and clinicians have advanced a number of theories to explain the high rates of co-occurrence. One prominent explanation for the high rates of co-occurrence is that individuals use psychoactive substances to "self-medicate" painful or disturbing psychiatric symptoms (Khantzian 1997; Chilcoat and Breslau 1998; Strakowski and DelBello 2000). Other theories suggest that substance-use disorders cause mental health problems or that substance use and mental health problems have common underlying genetic and environmental causes (Chilcoat and Breslau 1998).

Understanding the underlying causes of co-occurrence is important for improving the treatment and prevention of mental health and substance-use problems. If self-medication is common, then timely screening and treatment of mental health problems may prove the key in preventing the onset of substance-use disorders among the population with mental disorders. Although predictions about the substitutability of psychoactive substances and mental health care are implicit in the self-medication hypothesis, they have gone unexplored in the health services literature.

At the same time, empirical tests of the self-medication hypothesis in the clinical and epidemiological literatures have produced largely equivocal results (Khantzian 1997; Chilcoat and Breslau 1998; Raimo and Schuckit 1998; Dixit and Crum 2000; Strakowski and DelBello 2000). Here, mental health symptoms that pre-date the onset of substance use disorders are considered evidence of self-medication. With the exception of several studies based on long-term follow-up of longitudinal cohorts (Chilcoat and Breslau 1998; Vaillant 1998), study designs have been less than ideal, relying on the long-term recall of highly selected samples of patients with advanced substance-use disorders. In clinical interviews, patients may not accurately recall the temporal sequence of the onset of mental health symptoms and substance-use disorders. Further, among patients who did not confuse the temporal sequencing, advanced substance abuse may have exacerbated mental health symptoms.

This study tests for the presence of behavior consistent with self-medication by examining the relationships between drug and alcohol use and perceptions of unmet need for mental health care and use of mental health care. Specifically, we hypothesize that (1) unmet need for mental health care is associated with higher rates of substance use and (2) the mental health care is associated with lower rates of substance use, controlling for clinical and demographic characteristics.

To test these hypotheses, we pool data from the 2001 and 2002 waves of the National Household Survey on Drug Abuse (NHSDA) (renamed the National Survey on Drug Use and Health [NSDUH] in 2002). These data are well structured to identify patterns of substance use that are consistent with self-medication in several important respects. First, we have a reasonable degree of confidence that measured mental health care use and need precede our substance-use measures in time, because mental health care use and need are measured over the past year and substance use is measured in the past month. Second, our data contain a sufficient sample size and clinical symptom measures to eliminate individuals with substance dependence from our analyses. This exclusion makes it possible to isolate a subpopulation where self-medicating behavior, to the extent it occurs, is least likely to be confounded by mental disorders brought about by established substance-use disorders. Third, the large sample size also allows us to examine separately alcohol, marijuana, and other illicit drugs. This is important because these substances vary not only in terms of their psychoactive properties, but also in terms of the risks and costs associated with their use.

CONCEPTUAL FRAMEWORK

Central to the idea of self-medication is the notion that individuals with mental disorders perceive their symptoms as treatable and act purposefully to mitigate them through the use of substances and/or professionally sanctioned mental health care. Figure 1 illustrates the key relationships among the factors that influence the relative use of mental health care and substances to treatment mental health symptoms. In this context, the relative desirability will be influenced by preferences and beliefs about the relative effectiveness, health, and legal risks associated with substance use, resulting side-effects, and social acceptability of mental heath care and substance use. It is important to emphasize the notion of relative because both substance use and mental health care use are perceived as stigmatizing behaviors--but not necessarily equally so. Notions about relative desirability are formed through one's own experiences (indicated by the dashed arrow from mental health outcomes to desirability in Figure 1), learning through the experiences and opinions of others, and from formal sources of information, such as substance abuse prevention campaigns or programs promoting mental health screening. At the same time, the relative desirability of substance and mental health care use is moderated by access. In the case of mental health care, access is driven by the out-of-pocket cost of care determined by the existence and generosity of health insurance coverage, and the availability of willing providers. While alcohol is widely available to individuals over 21 years of age, safe access to illicit drugs requires a network of social relationships.