Combined Medication and Psychotherapy: Evidence for Acute-Phase Efficacy
The first question of interest is whether providing combined treatment to patients who are acutely symptomatic does anything to enhance response over that achieved with either modality alone. Only some of the psychosocial interventions have been evaluated in combination with medications. Because claims of differential efficacy have been made for each of these approaches, and because few practitioners are skilled in more than one, we discuss each in turn. Because allegiance to a particular approach is more likely to be found among psychotherapists than among pharmacotherapists, we organize the review with respect to the type of psychotherapy used.

Dynamic Psychotherapy
Early psychodynamic theories of depression focused on the role of retroflected anger and unconscious masochistic drives, whereas more recent work has emphasized the diminution of self-esteem following interpersonal loss in the context of unresolved conflict. Interventions based on these models strive to promote personality change by means of facilitating an understanding of past conflicts and providing insight into unconscious motivations. Despite their widespread use in clinical practice, these interventions have rarely been studied in the treatment of patients with mood disorders, and little is known about their actual clinical utility with such patients.

Only two studies have examined the effects of combining drugs and dynamic psychotherapy, and neither study was particularly supportive. Daneman (1961) found that adding pharmacotherapy (up to 200 mg/day of imipramine) enhanced response relative to dynamic psychotherapy alone in a sample of depressed outpatients, but he was unable to evaluate the efficacy of psychotherapy alone because of the absence of any other control condition. Covi et al. (1974) found no indication that adding dynamic psychotherapy did anything to enhance response over pharmacotherapy alone; imipramine (up to 200 mg/day) was generally effective whether provided alone or in combination with psychotherapy, whereas psychotherapy was generally ineffective unless combined with medications.

These findings suggest that dynamic psychotherapy has little effect on depression and does little in combination to enhance the short-term efficacy of pharmacotherapy alone. However, in both studies, psychotherapy “alone” was instantiated as the combination of dynamic psychotherapy and an inert pill placebo. Although pill placebo conditions have typically been found to be more effective than minimal-contact controls, there is reason to think that adding a pill placebo may undercut the efficacy of psychotherapy, because patients erroneously believe they are receiving an active medication (Hollon and DeRubeis 1981). Moreover, it is unlikely that the dynamic psychotherapy administered by Covi and colleagues was representative of more typical dynamic psychotherapeutic regimens; most proponents of that approach would have little faith in a mere 16 sessions of treatment provided in a group context. Thus, although there is no empirical justification for claiming that adding dynamic psychotherapy does anything to enhance the efficacy of pharmacotherapy alone in the treatment of depression, it is not clear that this particular combination has been adequately tested.

Marital and Family Therapy
Systems theory posits that the interplay of forces within a marital or family system can produce “psychopathology” within the individual (Ackerman 1958). Traditionally, such approaches have placed less emphasis on disorders in individuals and have concentrated on restructuring the larger system to produce lasting change. More recent versions of these approaches have recognized that psychopathology can also be a cause of disruption in the marital or family system and have increasingly emphasized the need to help spouses and families cope with the demands of dealing with a loved one who has a mood disorder (Miklowitz and Goldstein 1997).

Marital or family therapy has been combined with medications in four trials, although two were only minimally informative. Davenport and colleagues (1977) found combined treatment involving marital therapy to be superior to pharmacotherapy alone in a sample of bipolar patients, but in this study the patients were not randomized to treatment. Waring and colleagues (1988) compared combined pharmacotherapy plus marital therapy with marital therapy alone in a sample of dysthymic patients, but the sample was too small to draw any meaningful conclusions.

In the earlier of two fully informative trials in this literature, Friedman (1975) crossed two levels of pharmacotherapy (up to 200 mg/day of amitriptyline versus pill placebo) with two levels of psychotherapy (marital therapy versus brief supportive contact) in a sample of 196 depressed female outpatients. Although there were few differences in terms of the overall magnitude of symptom reduction by the end of 16 weeks of acute treatment, pharmacotherapy produced more rapid symptom change and marital therapy produced greater improvement in the quality of the relationship. Both advantages were retained by the combined condition. This suggests that combined treatment may enhance breadth of response. Moreover, an examination of the reported posttreatment values suggests that the combined treatment may actually have produced greater symptom change than either modality alone. Friedman appears to have assumed that a significant interaction was required to document an advantage for combined treatment with respect to acute response, which is not the case (Hollon et al. 1991b). Thus, this early study provides evidence for the superiority of combined treatment with respect to the breadth, and perhaps the magnitude, of response.

In the other major trial in this literature, Clarkin and colleagues (1990) added psychoeducational family therapy to standard hospital treatment (including medications) in an inpatient sample that included both bipolar and unipolar patients. Clinical ratings at discharge indicated an advantage for combined treatment on measures of symptom change, but only among female patients. Results across an 18-month naturalistic follow-up indicated that these gains were maintained only among women with bipolar disorder (who also showed gains in measures of social role functioning). Male patients (particularly the unipolar depressive patients) actually showed poorer short- and long-term outcomes, with respect to both symptom status and social role functioning, in the combined treatment condition than in standard treatment.

These findings, combined with those from the Friedman study, suggest the possible operation of complex interactions among gender, polarity, and the type of marital or family systems therapy provided. It may be that the type of psychoeducational approach used by Clarkin and colleagues, which emphasized the chronic and recurrent nature of the disorder, may have overstated the implications of pathology among the unipolar depressive males and adversely affected their capacity to meet gender-specific role expectations. At the least, these studies do suggest that adding marital or family therapy to medications affects response, typically, but not invariably, for the better.

Interpersonal Psychotherapy
Klerman, Weissman, and colleagues (1984) developed an approach to psychotherapy predicated on the notion that disturbances in important interpersonal relationships can play a role in causing or maintaining clinical depression. This approach, called interpersonal psychotherapy (IPT), is a neo-Sullivanian intervention that focuses on current life situations and interpersonal relationships for the purpose of resolving problems in those areas and reducing levels of distress. Originally developed for individual work with adult outpatients, it has been modified for different age groups, for psychiatric disorders other than mood disorders, and for conjoint marital and family interventions (Klerman and Weissman 1993). Recent trials have shown that its effects extend to primary care (Schulberg et al. 1996) and HIV-positive patients (Markowitz et al. 1998).

Weissman and colleagues (1979) provided an acute treatment trial in which they compared combined treatment involving IPT with each single modality and a “treatment-on-demand” control in a sample of 96 depressed outpatients (predominantly women). Results at treatment termination (16 weeks) indicated an advantage for combined treatment over each of the two single modalities (which did not differ in efficacy); all three active treatment conditions were superior to the minimal-contact control. (IPT was also found to be as effective as pharmacotherapy alone in the recent National Institute of Mental Health Treatment of Depression Collaborative Research Program [Elkin et al. 1989].) The apparent advantage for combined treatment over either single modality may have been a consequence of the different loci of action associated with each, with pharmacotherapy having had a more rapid effect on vegetative symptoms and psychotherapy having had a greater, but somewhat delayed, effect on mood and interest (DiMascio et al. 1979). There was also some evidence of differential response as a function of subtype of depression, with endogenously depressed patients doing less well in psychotherapy alone than with medications (Prusoff et al. 1980). There was no indication of differential relapse across a subsequent 1-year posttreatment follow-up, although patients previously treated with IPT (either alone or in combination with medications) did evidence greater gains in social functioning by the end of that period (Weissman et al. 1981). Thus, there were indications that combined treatment enhanced not only the magnitude of response but also its probability and breadth.

These findings are in at least partial accord with findings from related studies. Once again, pharmacotherapy appears to produce somewhat more rapid (but not necessarily greater) change in depression, whereas IPT appears to do more to help patients deal with problems in interpersonal relationships. As we shall see, this same pattern is evident even when IPT is introduced during the continuation phase (Klerman et al. 1974; Weissman et al. 1974). As with marital therapy, combined treatment appears to enhance the breadth of response and may also incrementally increase the magnitude of its effect for the average patient.

Behavior Therapy
From a behavioral perspective, depression is viewed as a consequence of insufficient reinforcement or inappropriate conditioning (Lewinsohn 1974). Interventions based on social skills training (Becker et al. 1987) or contingency management (Lewinsohn 1974) have predominated. Closely related is a self-control approach that views depression as a consequence of a deficit in self-reinforcement (Rehm 1977). Although self-control therapy contains some cognitive elements, it is typically classified as a behavioral intervention.

Combined treatment involving behavior therapy has been evaluated in three trials of varying quality. In the first, Hersen and colleagues (1984) randomly assigned 125 depressed female outpatients (including community volunteers recruited via advertisement) to one of four conditions: 1) combined treatment involving behavioral social skills training, 2) social skills training plus an inert pill placebo, 3) pharmacotherapy alone (up to 300 mg/day of amitriptyline), or 4) dynamic therapy plus pill placebo. No significant differences were evident among the conditions after 12 weeks of active treatment, even among patients with endogenous symptom patterns, although a greater proportion of patients assigned to pharmacotherapy alone were withdrawn from treatment due to nonresponse, and combined treatment performed somewhat less well (albeit nonsignificantly) than social skills training alone. Social skills training, either alone or in combination with other modalities, did produce greater improvement in social functioning than did pharmacotherapy alone (Bellack et al. 1983). Treatment gains were maintained across a subsequent 6-month maintenance interval, with no indication of differential relapse.

Wilson (1982) crossed two levels of pharmacotherapy (up to 150 mg/day of amitriptyline versus pill placebo) with three levels of behavior therapy (task assignment versus relaxation training versus minimal contact) in a sample of 97 community volunteers (predominantly women) screened for the presence of depression. No significant differences in efficacy were evident among the treatment conditions at the end of 8 weeks of active treatment, although such differences as were apparent indicated that the combined condition was associated with a more favorable response. Pharmacotherapy was associated with more rapid response, an advantage retained by the combination. Treatment gains were essentially maintained across a 6-month follow-up, with no indication of differential relapse.

Finally, Roth and colleagues (1982) compared combined treatment (involving up to 200 mg/day of desipramine) and group self-control therapy in a sample of 32 depressed community volunteers (predominantly women). No significant differences were evident at the end of 12 weeks of active treatment, although combined treatment did work more rapidly and produced nonsignificantly greater change. Treatment gains were essentially maintained over a 3-month follow-up, with no indication of differential relapse. In a more recent study, adding self-control therapy enhanced the efficacy of a day-treatment program that included structured group therapy, occupational therapy, and medications as needed (van den Hout et al. 1995).

On the whole, these studies are too few and the findings too weak to warrant drawing any firm conclusions. Of the five trials, two used inexperienced therapists, and three included symptomatic volunteers; samples were small and medication doses marginal in the trial by Roth and colleagues. There was no indication of differential response as a function of subtype of depression, although this possibility was assessed in only one of the studies. Also, there was no indication of any preventive effect for behavior therapy (either alone or in combination) in any of the trials, although the brevity of the follow-ups and the reliance on cross-sectional assessments may have precluded finding such an effect even if it did exist. At the same time, response was more rapid to combined treatment than to behavior therapy alone in two of the three relevant trials. Moreover, the addition of social skills training did enhance interpersonal functioning in the only trial in which the effect of such training was assessed, findings that are in accord with those observed for medication combinations involving either marital therapy or IPT.

Finally, it should be noted that more purely behavioral interventions have been undergoing a renaissance in recent years. Jacobson and colleagues (1996) found that the behavioral activation component of cognitive therapy was as effective as the full treatment package. Spurred by these findings, they have developed a more purely contextual approach that emphasizes dealing directly with negative life events. Although it has yet to be tried in combination with medications, it does appear to faring well in comparison to drugs in a current ongoing trial (Hollon 2000).

Cognitive Therapy and Related Cognitive-Behavioral Approaches
A cognitive model of depression holds that the way a person interprets an event influences subsequent affect and behavior. Depressed patients are seen as holding irrational negative beliefs and falling prey to maladaptive information processing strategies in the face of negative life events (Beck 1967). Cognitive therapy (CT) is predicated on the notion that the systematic correction of these problems in thinking can reduce dysphoric affect and facilitate efforts to cope with stress (Beck et al. 1979).

Several of the trials that have suggested an advantage for combined treatment involving CT are less than fully informative, either because not all patients in the relevant conditions received medications (Miller et al. 1989, in an inpatient sample; Teasdale et al. 1984, in an outpatient sample) or because assignment to treatment was not fully random (Beutler et al. 1987, in a geriatric sample). Bowers (1990) similarly reported that combined treatment was better than pharmacotherapy alone in an inpatient sample, but the author failed to adjust for initial differences before treatment that favored the combined condition. Beck and colleagues (1985) found no evidence that adding amitriptyline enhanced the efficacy of CT in a sample of unipolar depressive outpatients. However, the sporadic nature of the contacts with the prescribing physician and the marginal medication dosages (up to 200 mg/day) reduce confidence in the adequacy of the pharmacotherapy. Covi and Lipman (1987) found no differences between combined treatment and CT in a sample of community volunteers, although both conditions were superior to brief dynamic psychotherapy. As in the previous study conducted by this group (Covi et al. 1974), it is unlikely that the brief group format provided an adequate representation of dynamic psychotherapy as it is typically practiced.

Several other trials provide a more representative picture of the efficacy of combined treatment involving CT with respect to acute treatment and its consequences. Blackburn and colleagues (1981) compared combined treatment with either pharmacotherapy alone (amitriptyline or clomipramine, in dosages up to 150 mg/day) or CT alone in both a general practice and a psychiatric outpatient setting. Both combined treatment and CT alone proved superior to pharmacotherapy alone among the general practice patients. However, the rate of response to pharmacotherapy was so poor (14%) as to raise questions about the adequacy of the pharmacotherapy provided (see Keller et al. 1982 for a critique of the prescription practices of general practitioners). In the psychiatric outpatient sample, there were no significant differences between combined treatment and the two single modalities, although such differences as were evident indicated that the combined condition was associated with a more favorable response. There also was no indication of either differential response as a function of endogenicity (Blackburn et al. 1981) or any differential change in purported mechanisms as a function of treatment (Blackburn and Bishop 1983). However, a 2-year naturalistic follow-up (treatment was continued at a reduced intensity for the first 6 months) indicated that treatment with CT during the acute phase (either alone or in combination) was associated with a significantly lower rate of relapse (including return to treatment) after treatment termination (Blackburn et al. 1986).

Murphy and colleagues (1984) compared combined treatment with pharmacotherapy alone (nortriptyline maintained between plasma levels of 50 and 150 ng/mL), CT alone, and CT plus pill placebo in a sample of 95 depressed outpatients (predominantly women). Results after treatment termination (12 weeks) indicated an absence of significant differences among the conditions, although patients in the combined condition were somewhat more likely to show a full response. There was some indication that patients with an active coping style did better in CT, whereas patients with a more passive style did better in pharmacotherapy, which may have accounted for the slight nonsignificant advantage for combined treatment with respect to the probability of response (Simons et al. 1985). There was no indication of differential change in purported cognitive mechanisms as a function of differential treatment (Simons et al. 1984). Results from a 1-year naturalistic follow-up indicated that patients who had received CT during the acute-treatment phase (either alone or in combination with pill placebo) were less likely to return to treatment than were patients who had received medication only. Patients who had received combined treatment evidenced an intermediate rate that did not differ significantly from that of any of the other conditions (Simons et al. 1986).

Hollon and colleagues (1992) compared combined treatment with pharmacotherapy alone (imipramine in dosages up to 450 mg/day) or CT alone in a sample of 107 depressed outpatients (predominantly women). There were no significant differences after treatment termination (12 weeks), although such nonsignificant differences as were evident indicated that the combined condition was associated with a more favorable response. There were few indications of differential response as a function of subtype of depression, although the few patients who were dexamethasone nonsuppressors at initial screening did poorly in CT alone. Patients who received CT (either alone or in combination with medications) showed greater positive change on underlying cognitive propensities that confer risk for depression than did patients treated with pharmacotherapy alone (DeRubeis et al. 1990). A 2-year quasi-naturalistic follow-up indicated that patients who had received CT (either alone or in combination with medication) were less likely to exhibit a relapse after treatment termination than were patients who had received drugs alone and then had them discontinued (Evans et al. 1992). In fact, patients who had received CT were no more likely to exhibit a relapse after treatment termination than were patients who continued to be given study medications for the first year of the 2-year follow-up period.

Two large controlled trials from Germany compared combined treatment with either CT or drugs alone in both inpatient and outpatient samples (de Jong-Meyer and Hautzinger 1996). In the first trial, depressed patients without melancholia or with dysthymia were randomly assigned to either CT or pharmacotherapy or their combination (Hautzinger et al. 1996). Although differences between the groups were not significant, response rates were highest in the combined condition. In the second trial, depressed patients with melancholia were randomly assigned to either pharmacotherapy alone or combined treatment (de Jong-Meyer et al. 1996). Once again, treatment differences were not significant but did appear to favor combined treatment, at least in the outpatient sample.

Finally, two recent studies suggest that adding CT to ongoing medications can both reduce residual symptoms and prevent subsequent relapse or recurrence. In the first study, Paykel et al. (1999) recruited 158 depressed patients who were partially remitted after at least 2 months of drug treatment but were still experiencing residual symptoms. These patients were randomly assigned to either clinical management alone or the addition of 20 weeks of CT and the continuation of their medications. At the end of the continuation phase, CT was discontinued (but not the medications), and all patients were followed up for an additional year. The addition of CT not only reduced residual symptoms in medicated patients but also reduced subsequent relapse over the following year.

In a conceptually similar study, Fava and colleagues (1998a) first treated 45 depressed patients for up to 5 months with medications, then randomly added CT for residual symptoms for half the sample. Medications were tapered over the subsequent 20 weeks, and patients were followed up for an additional 2 years. Patients who received CT (in combination with medication) during the continuation phase were far less likely to experience the onset of a subsequent episode of depression (recurrence) over the subsequent follow-up period. Taken together, these studies suggest that adding CT to ongoing medications can both reduce residual symptoms and prevent the subsequent return of symptoms.

Thus, although differences favoring combined treatment over either single modality with respect to the magnitude of change typically were not significant, they were generally apparent. This suggests the presence of a modest but robust advantage for combination treatment involving CT and pharmacotherapy, about which we will have more to say later. Although there was no indication that pharmacotherapy worked more rapidly than CT alone, there were consistent indications that CT prevented relapse (and possibly recurrence) after treatment termination, an advantage retained by combined treatment. Although they are far from conclusive (such a pattern could be an artifact of differential retention), these findings suggest yet another way in which combined treatment may enhance the breadth of response relative to either drugs or psychotherapy alone.

One other study deserves special mention. As previously described, a recent multisite study found that patients with chronic depression were more likely to respond to combined treatment than to either pharmacotherapy or cognitive-behavioral therapy alone (Keller et al. 2000). In that study, patients with chronic depression were randomly assigned to either pharmacotherapy (nefazodone) or Cognitive Behavioral Analysis System of Psychotherapy (CBASP) or their combination. CBASP is the first psychotherapy developed specifically for chronic depression. In this approach, patients are taught to focus on the consequences of their behavior in problematic situations and to use a specific set of social problem-solving strategies to address problems in interpersonal relationships (McCullough 2000). In essence, it appears to combine the focus on interpersonal interactions of IPT with the greater structure and direction of CT. The most striking aspect of this study was the clear advantage provided by combined treatment. Nearly 75% of all patients responded to combined treatment, relative to only about 50% for each of the single modalities. Patients treated with medications did respond more rapidly than patients treated with psychotherapy alone. It remains to be seen whether these findings will be replicated, although they were robust across the dozen participating sites. Similarly, it also remains to be seen whether the advantage noted for combined treatment is specific to either chronic depression or the particular treatments studied, but the findings have generated considerable excitement in the field.

Summary
Although few studies document any clear advantage for combined treatment over either single modality with respect to the magnitude of acute symptom reduction in nonchronic populations, it would be premature to conclude that such an effect does not exist. Most studies have found a modest advantage for combined treatment, an advantage that falls short of conventional levels of significance but would be of interest clinically. As Kazdin and Bass (1989) noted, few of these studies have had sufficient power to detect the kind of modest increments that might be expected when combining two effective interventions. Clearly, additional studies with larger samples are needed before this issue can be resolved. Similarly, the relative paucity of indices of differential response may reflect the fact that the majority of studies in the literature have not been large enough or have not paid sufficient attention to the assessment of individual differences to detect such patient-by-treatment interactions if they did exist (Smith and Sechrest 1991). The recent finding of a clear advantage for combined treatment over either single modality in the treatment of chronic depression suggests the utility of conducting larger studies in targeted samples.

There is consistent evidence that combined treatment enhances the breadth of response. That is, whenever one modality does something that the other does not, and vice versa, combined treatment appears to retain the advantages of both. Drugs appear to work more rapidly than most types of psychotherapy (CT excluded) and may depend less on the skill of the practitioner for their effect. Conversely, psychotherapy appears to do more to enhance social functioning (particularly marital therapy or IPT) or reduce subsequent risk (particularly CT) than does pharmacotherapy alone. In essence, pharmacotherapy appears to provide rapid, reliable relief from acute distress, and psychotherapy appears to provide broad and enduring change, with combined treatment retaining the specific benefits of each.

Evidence for Continuation- and Maintenance-Phase Efficacy
Relapse refers to the return of symptoms associated with the treated episode after initial remission, whereas recurrence refers to the onset of a wholly new episode after recovery (Frank et al. 1991). Continuing medications for patients who are in remission but not yet fully recovered appears to prevent relapse (Prien and Kupfer 1986), and maintenance pharmacotherapy after recovery appears to prevent recurrence among both bipolar patients and unipolar patients with recurrent depression (Consensus Development Panel 1985).

There is currently no indication that pharmacotherapy does anything to reduce subsequent risk once medications are withdrawn (Hollon et al. 1990). Advocates of the psychotherapies have long argued that these approaches can produce lasting change, and, as previously described, there are some indications that this might be the case. Continuation and maintenance psychotherapy after acute response or recovery have been less often studied. Only a handful of such trials are currently available in the literature, most involving IPT.

In an early continuation trial, Klerman and colleagues (Klerman et al. 1974; Weissman et al. 1974) crossed three levels of medication (amitriptyline in dosages up to 150 mg/day versus pill placebo versus no pill) with two levels of psychotherapy (high versus low contact). The sample consisted of 150 depressed female outpatients whose depression had remitted with 4-6 weeks of pharmacotherapy. High-contact psychotherapy involved an early forerunner of IPT. Results across the 8-month continuation phase indicated that the principal effect of drugs was prevention of relapse (Klerman et al. 1974) and that of psychotherapy was improvement of social functioning, although the latter effect was evident only after months of treatment (Weissman et al. 1974). Combined treatment resulted in a relapse rate equal to that of medication alone and improvement in social functioning equivalent to that provided by IPT alone, suggesting once again that combined treatment can provide an incremental increase in the breadth of response relative to either single modality.

In a subsequent maintenance study by Frank et al. (1990), 230 outpatients (predominantly women) with recurrent unipolar depression were treated to recovery with combined pharmacotherapy and IPT. Then, those patients who showed a full and sustained response were randomly assigned to up to 3 years of maintenance treatment in one of five conditions. Treatment conditions included combined treatment, pharmacotherapy alone (up to 300 mg/day of imipramine), IPT alone (monthly visits), IPT plus pill placebo, and pill placebo alone. Both maintenance medication and maintenance IPT were associated with reduced rates of recurrence, although the former was considerably (but not significantly) more effective than the latter. Aside from having a modest, nonsignificant advantage during the first year of maintenance, combined treatment was no more effective than pharmacotherapy alone in preventing recurrence. Combined treatment did, however, reduce attrition: only 8% of the patients who received IPT in addition to medications dropped out of treatment as opposed to 21% of the patients maintained on medications alone.

Reynolds and colleagues (1999) conducted a placebo-controlled maintenance-phase trial that further suggested an advantage for combined treatment over either drugs or IPT alone with elderly depressed patients. In that trial, patients ages 60 years and over were first treated to recovery with a combination of IPT plus nortriptyline and then were randomly assigned to maintenance treatment with combined treatment, IPT alone, drugs alone, or placebo. Recurrence rates over the next 3 years favored combined treatment over either single modality (a trend in the case of IPT), which were each, in turn, superior to placebo. Thus, combined treatment was superior not only to IPT alone (as it had been in the study by Frank and colleagues [1990]) but also to pharmacotherapy alone. Whether this relative advantage is specific to elderly patients (or robust at any age) remains to be seen.

Finally, Jarrett and colleagues (1998) found that patients who continued in CT were less likely to relapse than if they discontinued after responding to acute treatment. It remains unclear whether such continuation treatment would prove useful for patients who are also continued on medications. Blackburn and Moore (1997) found that it made little difference whether patients were treated or maintained with medications or CT. However, drugs and CT were provided sequentially, and no patients were treated with the combination simultaneously.

In aggregate, these findings suggest that although medications alone may be adequate to suppress relapse or recurrence (geriatric patients being a possible exception), combined treatment may enhance the breadth of response just as it does when provided during the acute phase. Moreover, the effect of IPT on social adjustment appears to take time to develop, which suggests that this modality may have an effect on symptomatic expression over the long run that has yet to be detected in existing trials. On the whole, these studies appear to be consistent with those of acute-phase response: medications appear to produce reliable and robust effects on depressive symptoms, whereas psychotherapy appears to enhance the breadth and stability of response.

Side Effects
Side effects associated with combined treatment are largely those associated with pharmacotherapy. Such side effects are common but can typically be managed by altering the treatment regimen or changing the medication. Adverse reactions are rare and tend to occur early in treatment, if they occur at all. However, little is known about the long-term safety of the newer heterocyclic agents and serotonin reuptake inhibitors likely to be selected as the initial medication of choice. If psychotherapy can reduce risk for subsequent recurrence, then combined treatment during the acute phase might be preferred to long-term maintenance medication (Munoz et al. 1994).

There are some indications that adding psychotherapy may facilitate patients’ willingness to tolerate side effects and thereby enhance compliance. Cochran (1984) found that bipolar depressive patients treated with CT were more compliant with standard lithium maintenance therapy than were patients treated with pharmacotherapy alone. Similarly, recurrent unipolar depressive outpatients were less likely to drop out of maintenance pharmacotherapy if they also received IPT than if they did not (Frank et al. 1990). Concurrent psychotherapy provides an opportunity for patient and therapist to explore concerns that may not always be raised in the course of pharmacological management.