By Roger D. Weiss, Shelly F. Greenfield, and Lisa M. Najavits


In the past decade, there has been substantially increased interest in patients with coexisting psychiatric illness and substance use disorders (Meyer 1986a; Minkoff and Drake 1991; Mirin 1984). Two main factors have contributed to the interest in this subject.

First, studies of both primary substance abusers (Mirin et al. 1991; Ross et al. 1988; Rounsaville et al. 1991) and patients with primary psychiatric disorders (Caton et al. 1989; Drake and Wallach 1989; McLellan and Druley 1977) in clinical settings have revealed substantial rates of co-morbidity with the other disorder. Moreover, results of the National Institute of Mental Health Epidemiological Catchment Area study (Regier et al. 1990) confirmed that the frequent association between substance use disorders and psychiatric illness is not due to the bias inherent in studying clinical populations, but occurs in the general population at a significantly higher rate than would be expected by chance alone.

A second reason for the interest in these so-called ??odually diagnosed??? patients is the fact that research conducted in the early 1980s (McLellan et al. 1983) revealed that substance abusers with high levels of psychiatric severity (regardless of the exact nature of the specific coexisting psychiatric disorder) had poor treatment outcomes.

In a series of studies by McLellan (1986, pp. 97-139), the level of psychiatric severity was the most robust predictor of treatment outcome in their population of alcohol and drug-dependent patients. Moreover, certain forms of traditional drug abuse treatment, such as the confrontational approach utilized in residential therapeutic communities (TCs), were found to be particularly ill-suited for patients with coexisting psychiatric illness, as demonstrated by the finding that such patients who were treated in TCs had worse outcomes with longer treatment.

The combined findings of high rates of co-morbidity and a growing recognition of the ineffectiveness of traditional forms of substance abuse  treatment for dually diagnosed patients has led clinicians and researchers in recent years to (a) characterize more clearly the relationship between substance abuse and psychiatric disorders, and (b) search for effective approaches to the treatment of this patient population.


In treating a patient with a substance use disorder and coexisting psychiatric illness, it is important to understand the potential relationship between the patient??Ts two disorders. Meyer (1986b) has described six different ways in which substance abuse and psychopathology may interrelate: (a) Axis I or Axis II disorders may act as risk factors for substance use disorders, (b) psychopathology may affect the course of a substance use disorder, (c) psychiatric symptoms may develop in the course of chronic intoxication, (d) chronic substance use may lead to the development of psychiatric disorders that do not remit despite cessation of substance use, (e) substance use and psychiatric symptoms may become meaningfully linked over time, and (f) the two disorders may coexist without being related to each other.

In addition to the multiple potential relationships between substance use disorders and psychiatric illness, it is important to recognize the multiplicity of clinical presentations that can be subsumed by the term ??odually diagnosed??? patient (Weiss et al. 1992a). For example, the nature, length, and severity of psychopathology and of the substance use disorder may vary widely.

Dually diagnosed patients thus include a broad range of individuals, including patients with chronic severe mental illness and relatively mild substance use disorders, as well as patients with severe substance dependence and mild psychopathology, e.g., a simple phobia that is unrelated to the substance use. Moreover, even patients with the same two disorders, e.g., alcohol dependence and depression, may have different severity patterns based on a number of factors, including which disorder occurred initially (i.e., which disorder was ??oprimary???).

For example, Weissman and colleagues (1977) noted that patients with secondary depression and primary substance use disorder had less severe depressive symptoms than patients with primary mood disorder and secondary substance use disorder. This heterogeneity in patients who are dually diagnosed underscores the need to develop a variety of treatment approaches when working with this population. 


A number of studies (Kosten et al. 1987; McLellan et al. 1981) have shown that although patients with substance use disorders may have a wide range of problems associated with their addiction, these problems (including psychiatric problems) are not necessarily caused by their addictive disorder, and therefore do not necessarily improve merely as a result of achieving abstinence.

Therefore, there has been increasing recognition that patients with coexisting substance use disorders and psychiatric illness need to receive treatment for both disorders, as well as for associated problems such as vocational, legal, medical, and interpersonal difficulties (McLellan et al. 1992, pp. 231-252).

Although there are differences of opinion regarding which specific techniques to utilize in dual diagnosis treatment, several stages have commonly been described in the treatment of these patients: crisis intervention, medical and psychiatric stabilization, engagement, motivation or ??opersuasion??? of the patient to seek substance abuse treatment, asking the patient to make a commitment to pursue active treatment, and relapse prevention (Fariello and Scheidt 1989; Kofoed and Keys 1988; Minkoff 1989; Ridgely 1991, pp. 29-42).

Since both psychiatric illnesses (particularly those of greater severity) and addictive disorders are frequently accompanied by minimization or denial of symptoms, overcoming this resistance to treatment (which is frequently related to feelings of shame, stigma, and hopelessness) is an important early step in the treatment process.

There has been some controversy over whether the treatment of dually diagnosed patients should occur in an ??ointegrated??? or a ??osequential??? program (Minkoff 1989). Integrated treatment programs, which provide simultaneous substance abuse and psychiatric treatment, have recently gained favor (Minkoff 1989; Ries and Ellingson 1990), although the authors are aware of no studies that have clearly demonstrated the superiority of this approach over sequential treatment, in which dually diagnosed patients receive episodes of substance abuse treatment and psychiatric treatment in sequence (in either order, depending on the patient, program, or response to treatment).

Despite some fundamental differences in the integrated versus sequential models, one similarity between the approaches is the general use of pharmacotherapy to primarily treat the patient??Ts psychiatric disorder (Siris 1990), with the implicit hope that improvement in psychiatric symptoms  will (a) help make a patient more accessible to psychosocial treatment for substance abuse, and (b) reduce the patient??Ts vulnerability to relapse to substance use by diminishing symptoms such as psychosis, depression, or anxiety.

Indeed, one of the problems with some of the early studies of antidepressant treatment of alcoholic patients was related to this implicit assumption. Indeed, some such studies failed to measure changes in both depression and drinking behavior as outcome measures (Ciraulo and Jaffe 1981). One of the advances in more recent clinical and research approaches to this topic has been the clear understanding that the treatment of dually diagnosed patients requires specific attention to both disorders, and measurement of outcome in both domains.


The integration of psychotherapeutic and pharmacologic approaches to psychiatric illnesses other than substance abuse has been the subject of a great deal of research (Beitman and Klerman 1991; Karasu 1982; Sarwer-Foner 1983, pp. 165-180). Klerman (1991, pp. 3-19) has outlined a number of potential interactional effects between pharmacologic treatment and psychotherapy.

He has divided these into both positive and negative effects; the potential effects of pharmacotherapy on psychotherapy are listed in table 1. Klerman (1991, pp. 3-19) also has described the potential beneficial and detrimental effects of psychotherapy on psychopharmacologic treatment.

First, some individuals may hold the belief that since psychotropic drug treatment is designed to correct an underlying metabolic or biochemical imbalance or dysfunction, then adding psychotherapy (while not necessarily harmful) would represent an unnecessary investment in time, energy, and expense. Moreover, it is possible that exploratory psychotherapy, particularly when undertaken early in the treatment process, may disrupt early defenses and undo some of the healing and ??osealing over??? that is facilitated by the use of medications.

It is important to note that these potential objections to psychotherapy are theoretical and not based on empirical studies that demonstrate the worsening of patients when psychotherapy is added to their pharmacotherapeutic regimen. Theoretical benefits to the addition of psychotherapy to medication treatment include (a) the facilitation of medication compliance by helping the patient to further understand the nature of his or her illness and   enhancing motivation for positive change; and (b) the correction of associated difficulties such as interpersonal problems and poor self-esteem, which may occur as a result of having a psychiatric illness.

TABLE 1. Potential positive and negative effects of pharmacologic treatment on psychotherapy. [See source article.]

Positive Effects 1. Medications facilitate accessibility to psychotherapy. 2. Medications influence the ego-psychological functions (cognitive functioning, attention, verbal skills, concentration) required for participation in psychotherapy. 3. Medications may promote abreaction.

Negative Effects 1. Reduction of symptoms may lead patients to stop psychotherapy. 2. Medications may undercut defenses. 3. For patients who value psychotherapy, the use of medications may be seen as a failure on their part.

Patients with substance use disorders, even in the absence of associated psychopathology, are frequently noncompliant with medication regimens and suffer from poor self-esteem, shame, interpersonal difficulties, and a variety of other associated problems.

It therefore could be posited that psychotherapeutic interventions with dually diagnosed patients, who experience these difficulties in a more profound way as the result of having more than one illness, would serve to both help improve compliance and to assist in the rehabilitative process.

Since ensuring medication compliance is one of the primary treatment goals in working with psychiatric patients, and since dually diagnosed patients tend to have poorer medication compliance than either patients with substance use disorders alone or psychiatric illness alone (Drake et al. 1989), addressing this issue is critical.


Despite evidence from studies of both substance abusers and other psychiatric patients that a combination of psychotherapy and  pharmacotherapy is more effective than either alone, there have been virtually no studies of this subject in dually disordered patients.

Rather, most treatment studies of patients with substance use disorders and coexisting mood or anxiety disorders have thus far involved trials of medications that are primarily designed to treat the coexisting psychiatric illness, with the hope that by carefully identifying and treating coexisting psychiatric disorders in substance abusers, the outcome of their substance use disorders can be improved as well.

Studies of patients with psychotic and substance use disorders have, on the other hand, primarily focused on psychosocial strategies that integrate the treatment of the two disorders; the medications used are generally held constant and are typically those medications ordinarily prescribed for the treatment of psychosis.


A number of studies have examined the treatment of depressed substance abusers with antidepressants (Weiss and Mirin 1989). Despite the aforementioned methodological flaws of early antidepressant studies, more recent research has suggested the potential benefit of this treatment approach, at least for improving mood.

Nunes and colleagues (1993) studied the efficacy of imipramine in patients with primary depression and alcoholism. They treated 60 such patients in a 12-week open-label trial; the 35 patients (58 percent) who were judged to be responders during this initial period (i.e., they had substantial improvement in both mood and drinking behavior), were then offered the opportunity to enter a double-blind, placebo-controlled, 6-month discontinuation trial.

Twenty-six patients entered this phase of the study, 23 of whom completed the trial. Four of 13 patients (31 percent) relapsed on imipramine, as compared with 7 of 10 (70 percent) who relapsed on placebo (p = 0.09). The authors noted that in a subgroup of patients, imipramine had a more powerful effect on mood than on drinking.

Moreover, patients with coexisting panic disorder appeared to have a more robust response to imipramine than did patients with depression alone. A small study of desipramine for patients with depression secondary to alcoholism also suggested its potential utility.

Mason and colleagues (1992) compared 11 patients on desipramine with 10 patients on placebo in a 6-month random assignment trial and found that patients treated with desipramine had significantly more sober days and significantly fewer depressive symptoms than patients who were given placebo. 

Thus, recent studies of antidepressant treatment of coexisting depression and alcoholism suggest the possibility of a positive response, although the major benefit of this treatment approach may be the reduction of depressive symptoms. Although this is intrinsically helpful, mood improvement is not necessarily associated with a corresponding reduction in drinking.

These studies have generally been hampered by small sample sizes and a number of confounding variables (e.g., the mixture of patients with major depression and dysthymia, primary alcoholism and primary depression, and patients with and without coexisting panic disorder), all of which render clear interpretation of these data difficult.

Studies of antidepressants in depressed opioid addicts receiving methadone maintenance treatment have been plagued by analogous methodological problems, and have thus yielded similarly modest results. In most such studies, depression was diagnosed on the basis of a current assessment of depressive symptoms rather than a lifetime clinical historical assessment.

Moreover, virtually all of these studies have had small sample sizes, thus increasing the possibility of a type II error (i.e., accepting a false-negative result as true) in the interpretation of results. As with the studies of depressed alcoholics, the effect of antidepressants on mood has been more robust than the effect on drug use (Weiss and Mirin 1989).

In a recent study of imipramine in 17 methadone maintenance patients with either primary or chronic depression, 9 (53 percent) improved on measures of both mood and drug use after being treated with imipramine for a period of time ranging from 6 weeks to 11 months (Nunes et al. 1991).

However, patients with dysthymia and major depression were both included, and the potential confounding effect of coexisting panic disorder in some patients may have affected these results. Moreover, this was an open-label study, and previous work with this population has shown the potential importance of a response to either a placebo or the extra attention and psychosocial treatment given to research subjects (Kleber et al. 1983). Ziedonis and Kosten (1992, p. 365) conducted a comparative study of amantadine, desipramine, and placebo in 20 depressed and 74 nondepressed cocaine-abusing methadone maintenance patients; all patients also received relapse prevention treatment.

The depressed patients who were treated with placebo had a significantly worse treatment outcome than the nondepressed group. However, the depressed patients who were treated with medication reported significantly less cocaine use than the depressed patients who were given placebo.

Thus, these data suggest that relapse prevention treatment alone is not  particularly effective for depressed, cocaine-abusing methadone maintenance patients.

However, a combination of relapse prevention treatment and medication may be beneficial for this population. In sum, while there are some encouraging findings regarding the potential efficacy of antidepressants in substance abusers with coexisting depression, methodological difficulties involved in performing these studies have limited the generalizability of their results.

Moreover, the most powerful effect of antidepressants in these patients appears (not surprisingly) to be a reduction in depressive symptoms. Unfortunately, while this may be associated with a corresponding reduction in substance use in some patients, this is not universally true.

This appears to be an area in which the interaction between psychotherapeutic interventions and pharmacotherapies could be very important and should be studied.

For example, it would be important to know which patients exhibit improvement in their mood symptoms and are thus able to reduce or stop their substance use, and which patients are not. It is possible, for instance, that factors that influence the likelihood of improvement in substance use are independent of the nature and/or severity of the patient??Ts coexisting mood disorder.

Conversely, it would be important to study patients who do not respond to an antidepressant with mood improvement, but who are able to stop their drug use anyway; such patients may be responding more powerfully to a psychosocial intervention.