Much like asthma, hypertension, and diabetes, substance use disorders are chronic, relapsing conditions (1). As the articles in this special section of Psychiatric Services show, this characteristic is no less true for people with severe mental illnesses than for others. A majority of clients with co-occurring disorders attain full remission of their substance use disorder (at least six months without evidence of abuse or dependence), but they also tend to relapse by returning to problematic substance use. Even those who remain in remission for years continue to be at risk of relapse.
Research on substance use disorders suggests that recovery is a longitudinal process that often involves return to problematic substance use during the course of building a different life by acquiring new attitudes, skills, supports, relationships, habits, and coping strategies (1,2,3). During this process, clients' intentions and efforts interact complexly with social, neighborhood, cultural, and societal factors (4).
One might question whether "remission," "relapse," and "recovery" are in fact the best terms for describing the complex process of overcoming problematic substance use and adopting a completely different lifestyle. Customary usage makes the terms unavoidable, but the rhetoric of medical terminology may oversimplify what the articles in this special section show to be a highly complex mixture of use, risk, context, and consequences, the nature of which depends heavily on the society in which it occurs. When it comes to substance use, we need to keep in mind that people with mental illnesses, like the rest of us, are human agents, not just passive sites of biomedical conditions. That said, we grant that the oversimplified terms are probably here to stay.
Given the chronic, recurrent nature of substance use disorders and the longitudinal nature of recovery, relapse prevention should logically be a major focus of dual disorder treatment programs (5). However, recent reviews of dual disorder interventions (6,7,8) have documented that essentially all existing studies focus on engaging clients in treatment (the engagement stage), motivating them toward remission (the persuasion stage), or initiating remission (the active treatment stage). Meanwhile, no controlled studies have been conducted of preventing relapse during the maintenance or recovery stage. Recognizing this omission in the dual diagnosis literature in fact inspired this multisite, multiperspective view of relapse prevention.
As McGovern and colleagues (9) describe, theory and research on substance abuse relapse and prevention in the general population are extensive. As theoretical models of relapse prevention become more sophisticated (10), evidence-based approaches to relapse prevention support several common strategies: reducing exposure to substances; fostering motivation for abstinence; self-monitoring of situations, settings, and states; recognizing and coping with cravings and negative affects; identifying thought processes that have relapse potential; and using, if necessary, a crisis plan (11). Note that current approaches to relapse prevention emphasize cognitive aspects and minimize the socioenvironmental and biological contexts of relapse. However, as we describe below, these biases may render current theories and interventions less useful for persons with severe mental disorders. The emerging evidence suggests that relapse and relapse prevention issues for this population are somewhat different from those that have been identified for the general population (12). Thus, using current models or developing a new model of relapse prevention for persons with severe mental disorders may be premature.
Several early studies addressed relapses of substance use disorders and the prevention of such relapses among clients with a dual diagnosis. These studies identified risk factors for substance abuse relapse, including symptoms of mental illness (12,13), dysphoria (12,14), lack of protective housing (15,16,17), social pressures within high drug-using networks and neighborhoods (12,14,18,19,20), interpersonal stressors (12,14,18,19), social isolation for some (12,18), lack of meaningful activities (12,15,21), and lack of trusting treatment relationships (15,19,22,23). For example, manic episodes and severe anxiety are particularly likely to precipitate relapse (12,13). Individuals often make the transition to independent housing after they attain abstinence, but some protection from social pressures (17) and of abstinence of several months' duration (6) appear to be critical. Social pressures include having drug-abusing social networks (12,18) and being victimized by drug dealers who pretend to be friends when the individual receives his or her Social Security check (20). These findings come from quantitative studies (14,21), qualitative studies (12), ethnographic studies (15,18,19), focus groups with outreach workers (20), and numerous self-reports (22,23).
The findings reported in this special section, again based on a variety of methods but all based on naturalistic studies of abstinence and relapse, reinforce and elaborate on these earlier results. In a prospective, quantitative study of clients in full six-month remissions, Xie and colleagues (24) reported that being female, having at least a high school education, not living in independent housing, and continuing in substance abuse treatment were associated with sustained remission. Using a similar design and clinical records rather than research data, Rollins and colleagues (25) found that being older, being employed, and living in congregate housing programs were associated with sustained remission.
On the basis of focus groups with clients who were in remission, Davis and colleagues (26) learned that clients maintained remission by using behavioral strategies (for example, avoiding high-risk situations and focusing on healthy behaviors), using clinical and self-help supports (for example, dual diagnosis and dual recovery groups), relying on spirituality (for example, invoking a higher power as used in Alcoholics Anonymous), and developing meaningful goals and activities (for example, employment).
Using qualitative interviews with female clients who were in different phases of recovery, Harris and colleagues (27) found that having a role model, access to alternative social networks, meaningful activities, and a sense of community enhanced success, whereas severe symptoms, the pull of old habits and social networks, and the stress of developing a radically different lifestyle were significant barriers. Although these findings differ somewhat, probably as a result of study group and setting differences as well as methods, they converge remarkably on findings regarding the importance of meaningful activities and safe living environments.
The initial findings in this special section and elsewhere on preventing relapse of substance use disorders among clients with co-occurring disorders converge on a consistent set of observations, with both social and clinical implications. They highlight several features of relapse to substance use disorders among persons with severe mental illnesses. The temporally proximal risk factors, or precipitants, of relapse for this group appear to be largely similar to those that affect other populations: interpersonal problems, negative emotions, social stresses, long-standing habits, lack of involvement in more satisfying activities, and attempts to escape from painful experiences. One unique proximal risk factor for this group involves the fluctuating nature of long-term mental disorders. People appear to be at increased risk of substance abuse relapse when their symptoms of mental illness recur. These various factors represent risks in the epidemiologic sense, and are not necessarily causal factors.
Temporally more distal, or predisposing, risk factors for relapse of substance use disorder appear to be related to the long-term biopsychosocial sequelae of severe mental illness in this society. These distal risks for relapse would include a complex interaction of many vulnerabilities—psychiatric illness, illness-related deficits, lack of normal developmental experiences, shunting into poverty and dangerous neighborhoods, social and vocational marginalization, victimization and criminalization, and lack of needed treatment resources.
Impinging on this picture is the question of how adequately a culture provides for people who have disabilities in the first place, because these distal risks interact in numerous ways with the general wholesomeness or toxicity of the environment. Consider several examples. First, as the availability of psychiatric hospital beds declines, the rate of criminalization and incarceration of persons with mental illness continues to rise (28). Incarcerations are often related to minor crimes, such as public nuisance and possession of small amounts of marijuana (29), but the personal consequences of incarceration in lieu of hospitalization are profound: physical and emotional victimization; exposure to serious criminal behavior and drug dealers; inability to obtain treatment while incarcerated; loss of housing; and further barriers to work, housing, and treatment on release. Thus it appears to be a bad bargain, for several reasons, to replace mental hospitalization with incarceration.
Second, extreme poverty forces people with psychiatric disabilities into high-crime, drug-infested neighborhoods, where their cognitive and social deficits, combined with their predictable monthly checks, make them easy and sought-after targets for local drug dealers (20).
Third, although the great majority of adults with psychiatric disabilities want to work competitively (30), employment rates remain low because of disincentives in the disability payment system (31), the failure to make supported employment services widely available (32), and perhaps the inadequacy of the Americans With Disabilities legislation to protect individuals with mental health problems (33). Clearly, the return to substance abuse or dependence is more than relapse biomedically defined, because it likely involves numerous contextual factors.
In addition to the psychosocial sources of risk identified in this special section, we also must recognize another kind of distal risk—that people with mental illnesses may be vulnerable to substance use disorders and to relapse of those disorders as a result of biological considerations (34,35,36). Here relapse in the narrower biomedical sense applies. Substance dependence is associated with neurochemical changes in the brain that subserve compulsive use, tolerance, and withdrawal. Individuals with mental illnesses also manifest a heightened sensitivity to the biological effects of psychoactive substances and may react adversely to relatively small amounts of substances (37).
Finally, the same neurobiological dysfunctions may underlie both severe mental illnesses and substance use disorders (38,39), perhaps related to neurobiological deficits in brain reward systems (40). Note that this is different from speculation about "self-medication," which denotes the self-selection of specific pharmacologic antidotes for particular symptoms or illnesses (41). Although intuitively appealing, the self-medication model has little current scientific support (37).
Most of the factors identified in the preceding articles as protective against substance use disorder relapse for clients with dual disorders also resemble the usual targets of relapse prevention interventions: cognitive strategies, continued involvement in treatment or self-help, supportive relationships, and meaningful activities (42,43,44). However, people with severe mental illnesses differ from others with substance use disorders in several important ways: they tend to have pervasive cognitive and social dysfunctions, they usually need long-term mental health treatment, and they often need assistance to find and maintain housing and employment. Cognitive deficits can be expected to reduce the effectiveness of learning new skills, and social deficits can lead to isolation and victimization.
Medications and long-term, trusting treatment relationships are clearly effective in controlling symptoms of mental illness and may also induce hope, relatedness, and strategies for recovery. The need for housing and vocational supports is salient in many studies. Both the qualitative and quantitative data reported in this special section indicate that protective housing arrangements are experienced as a needed bulwark against the perceived dangers of neighborhoods and local social forces, which are rife with substance purveyors and users who are eager to victimize vulnerable individuals. The studies also identify the importance of work and other meaningful activities in the recovery process.
Consider these factors—cognitive and social deficits, difficulties with housing and work, and long-term mental health treatment—in a broader perspective. The protective functions that mental hospitals previously served have, in the era following deinstitutionalization, never been adequately compensated for by comparably safe and secure community living arrangements. At the same time, adequate outpatient mental health treatment and rehabilitation services have never grown sufficiently to fill the gap left by the closing of mental hospitals (45,46,47).
This is not to glorify hospital care, but to state the necessary conditions for its termination. "Community" loses meaning if low-cost safe housing and sufficient support, treatment, and rehabilitation are unavailable. This society's problem may be the ambivalence with which it maintains the ideal of community in practice—an ideal that conflicts, after all, with America's focus on autonomy and independence as prime virtues. For example, independent housing may be a social ideal, based on overvalorization of independence and autonomy, with undesirable consequences for people with multiple disorders. Clients themselves often recognize the need for a different type of housing, which provides safety, security, and support for recovery, but find it unavailable.
What are the implications for next steps? First, there is an urgent need to develop healthy environments that are experienced as nurturing of recovery—living situations that provide meaningful opportunities rather than pervasive dangers and that facilitate the acquisition of supports and skills needed to maintain stable and complete remission from substance abuse and to build more satisfying lives—not because we view abstinence and work moralistically as virtues but because clients with severe mental illnesses have identified these goals themselves.
We unfortunately know relatively little about what these arrangements should look like. Certainly, they should not be large housing projects with intense crowding and segregation of risky people and risky situations. Independent apartments in neighborhoods that prove dangerous for vulnerable individuals should also be considered to be inappropriate. Healthy environments in urban areas might include arrangements such as the residential programs at Thresholds in Chicago (48) and Community Connections in Washington, D.C. (49), which honor people's preferences for choice, privacy, peer and professional supports, and protection. The society's tendencies toward bureaucratic solutions should not be reborn outside mental hospitals after their failures within them. As in the Thresholds and Community Connections programs, modest size and a variety of housing arrangements seem important desiderata. Many other Western societies provide better incomes and low-income housing within economically diverse housing communities for people with disabilities (50)—a more humane approach that would require serious political attention to disability policies in the United States (51,52).
Second, we should emphasize helping people to make fundamental changes in their lives, such as finding meaningful jobs and abstinent friends, which would enable them to manage chronic illnesses and to pursue satisfying life experiences. Researchers have demonstrated repeatedly that people with substance use disorders need help in multiple areas, such as with legal, family, and vocational problems, and that the addition of psychosocial services improves outcomes (53,54). Similarly, mature dual disorders treatment programs incorporate many of the same elements, such as medical care, legal assistance, vocational services, psychological counseling, and assistance with family and other relationships (55).
Third, recovery clearly involves long-term supports and relationships. Clients themselves identify the need for new friendships with people who do not abuse substances as a critical step in recovery (15). Self-help programs for people with dual disorders, such as Double Trouble in Recovery (56), are developing in many areas, and other self-help programs and consumer-run programs are beginning to incorporate attention to co-occurring disorders (57). Traditional self-help groups such as Alcoholics Anonymous have been naturally expanding to accommodate new types of clients, including persons with severe mental illnesses. Similarly, family psychoeducation programs are beginning to attend to dual disorders (58).
Fourth, because most people with co-occurring disorders actually have multiple disorders, not just dual disorders, they may need specific and individualized treatments for those particular problems in order to maintain remission. For example, people with posttraumatic stress disorder appear to need trauma-focused treatments to support treatment adherence and to sustain remission (59,60). Similarly, many clients with co-occurring disorders need specialized medical attention for comorbid medical conditions, such as HIV and hepatitis C (61).
Fifth, client-centered approaches demand that we pay more attention to meeting their spiritual needs—to a sense of meaning that lifts people beyond themselves (62). Many surveys show that clients with mental illness identify spirituality as an unmet challenge (63). Alcoholics Anonymous and other substance abuse self-help organizations have long recognized its importance (57). How spirituality becomes incorporated into self-help, treatment interventions, and residential and other programs remains unclear, but emerging ideas regarding "recovery communities" incorporate notions about finding a sense of purpose by joining a community and helping others (27). We also need to recognize that such attention to spiritual needs is often overlooked and can be a source of discomfort to many mental health professionals.
Finally, these data have clear and significant implications for the future of research. More research attention needs to be directed toward understanding long-term abstinence and preventing relapse rather than continuing to focus on initiating remissions. Because relapse prevention and recovery from dual disorders are long-term issues, we also need much more emphasis on longitudinal research. Furthermore, clients are clearly heterogeneous in many respects, and understanding their heterogeneity should lead to better combinations of psychological, social, pharmacologic, and structural interventions that match their needs. For example, physiological addiction, antisocial personality disorder, and severe cognitive deficits may represent client characteristics that will require different types of treatment and support. It seems unlikely that the current movement toward criminalizing and incarcerating these individuals will be helpful, but the development of alternative approaches is only beginning.
Perhaps most important, intervention research needs to emphasize the environmental context (64) and not just new medications. People with dual disorders are particularly vulnerable in the impoverished, drug-infested neighborhoods in which they often reside, yet they may have a much greater chance for recovery in more benign, more supportive, and relatively protected environments. Perhaps we should be developing and studying healing communities rather than more sophisticated technical treatments. Treatment addresses disease, but healing addresses the person's life and context as well (65). This goes against the individualistic and technological grain of our society but acknowledges that in some degree the problem at hand emerges from the conditions for living that the society makes available to its more troubled members.
Mental health and housing policies are inevitably intertwined (51). Addressing economic and political considerations will be necessary in order to create appropriate and decent housing and a sense of community. We might still gain inspiration from the centuries-long history of places like Geel, Belgium, that took individuals with severe mental illness into its households as boarders and offered local farming and other jobs as employment (66).
Dr. Drake and Dr. McGovern are affiliated with the department of psychiatry of Dartmouth Medical School, 2 Whipple Place, Lebanon, New Hampshire 03766 (e-mail, firstname.lastname@example.org). Dr. Wallach is with the department of psychological and brain sciences of Duke University in Durham, North Carolina. This article is part of a special section on relapse prevention among patients with co-occurring substance abuse and other mental disorders. Robert E. Drake, M.D., Ph.D., served as guest editor of the section.