(Reprinted with permission from Depression and Anxiety 2011; 28:263–266)

Depression is a self-limiting disorder that tends to be recurrent. This means that although any given episode typically will remit on its own even in the absence of treatment, anyone who has ever had one episode of depression will likely have another (or more) and a sizable minority will become chronic.^[1] Although we tend to think of depression as an eminently treatable disorder (about two-thirds of all patients will respond to any given treatment), rates of remission are not all that impressive; only about one patient in three will remit following a brief course of treatment lasting several months and only about two in three will remit following an extended course of treatment lasting up to a year and involving exposure to several sequential types of treatment.^[2]

The majority of depressed patients are treated pharmacologically (if they are treated at all) and this represents a change over the last several decades. Since the advent of the relatively safe and easily managed specific serotonin reuptake inhibitors (SSRIs) in the late 1980’s, primary-care physicians have become more comfortable prescribing antidepressant medications (ADMs) and the number of people treated and the proportion treated with medications has mushroomed.^[3] That more people are treated is a good thing as the major problem with depression before that time was that too many depressions went untreated. Nonetheless, reliance on ADMs as the first (and often only) line of treatment may be less than optimal.

There is no question that ADMs work, but there is considerable reason to think that their efficacy and specificity have been overstated. A recent reanalysis of the FDA database for all registered trials for a dozen of the more recently approved medications (including the bulk of the SSRIs) found evidence of substantial publication bias.^[4] Trials that produced positive findings almost invariably found their way into the literature, whereas trials that did not either were never published or were published in a manner that made the target medication appear to be efficacious when it was not. There also is considerable evidence that severity moderates the likelihood that active medication will separate from pill-placebo such that over half the patients with major depressive disorder (MDD) derive little pharmacological benefit from taking ADM.^[5] This is some thing that the pharmaceutical companies understand all too clearly; for decades their trials have screened out patients with less severe depressions, although they are not at all reluctant to market to such patients. Finally, there is no evidence that having taken ADM does anything to reduce risk for subsequent symptom return and standard practice has evolved over the last several decades to keep most patients who respond on medication indefinitely.^[6]

Early studies suggested that cognitive therapy (CT) was at least as efficacious as medications, but most of these trials were conducted by advocates for CT and questions could be raised about the adequacy of the pharmacotherapy provided. The NIMH Treatment of Depression Collaborative Research Program (TDCRP) was the first placebo-controlled comparison of drugs and psychotherapy in the treatment of MDD; there were no differences among patients with less severe depressions,^[7] but among patients with more severe depressions both ADM and interpersonal psychotherapy (IPT) were superior to pill-placebo but CT was not.^[8] This was the only study in which ADM was superior to CT, but it had an outsized impact on the perception of treatment efficacy and practice guidelines published by the American Psychiatric Association suggested that ADMs should be provided for more severe depressions.^[1]

However, there were differences (albeit nonsignificant) across the sites in the TDCRP that suggested that CT did as well as medications at the site with prior experience with that psychosocial intervention.^[6] A subsequent attempt at replication that ensured that both modalities were adequately implemented (the Penn/Vandy study) found that CT was as efficacious as ADM and that both were superior to pill-placebo in the treatment of patients with more severe depressions.^[9] As in the TDCRP, there were differences between the sites that appeared to reflect therapist competence; the less experienced cognitive therapists at Vanderbilt required additional training and supervision during the study proper to rise to the level of the more experienced cognitive therapists at Penn in terms of patient outcomes. ADM was more efficacious than CT for patients with Axis II disorders, whereas CT was superior to ADM for those without.^[10] Number of prior exposures to ADM predicted poorer response to medications but not to CT,^[11] and patients with greater numbers of adverse life events or who were unemployed did better in CT than on ADM.^[12] These findings suggest that competently implemented CT can be as efficacious as ADM for more severely depressed patients and point to patient characteristics that can be used to select the better treatment for a given patient. Less severely depressed or complicated patients appear to respond to any reasonable nonspecific treatment.

Most critically, patients who responded to CT were only about half as likely to relapse (or recur) following treatment termination than were patients who responded to ADM and no more likely than patients continued on medications.^[13] This is a relatively robust finding that has been found across multiple prior trials comparing cognitive behavioral interventions to medications across a number of different nonpsychotic disorders.^[14] This suggests that CT has an enduring effect that reduces risk for subsequent symptom return that is more than purely palliative. It is more expensive to treat someone to remission using CT than ADM, but its enduring effect may make CT more cost-efficient than ADM over time. Patients who best learn and can perform the basic strategies inherent in CT are least likely to relapse.^[15] The same applies to those patients who show sudden gains, a pattern of rapid response following insight into the role of cognition.^[16] With respect to treatment process, therapist competence with CT predicted subsequent symptom change on a session-by-session basis.^[17] There also are indications that cognitive behavioral strategies can be used to prevent the onset of depressions in at-risk adolescents.^[18]

A recent study from Seattle suggests that a more purely behavioral intervention called behavioral activation (BA) may be as efficacious as ADM and as enduring as CT in the treatment of MDD. In that trial, no differences were evident among patients with less severe depressions, but among patients with more severe depressions both BA and ADM were superior to pill-placebo where as CT was not.^[19] This was a pattern of findings reminiscent of those found in the TDCRP (with BA substituted for IPT). CT therapists in the Seattle trial also were no more expert with that approach at the start of the trial than the therapists at Vanderbilt and did not have the advantage of additional training and supervision during the study proper. CT did particularly poorly with more severely depressed patients with a history of difficulty in interpersonal relationships.^[20] This is reminiscent of the poorer response shown to CT by Axis II patients in the Penn/Vandy study and suggests that less than fully expert cognitive therapists are most likely to get in trouble with more severe and complicated patients. Nonetheless, patients who responded to CT were less likely to relapse (or recur) following treatment termination than patients who responded to ADM and no more likely to relapse than patients continued on medications, and patients who responded to BA did nearly as well.^[21] This again suggests that the cognitive and behavioral interventions may be more enduring and therefore more cost-efficacious than medication treatment over time. Other psychosocial approaches have not been adequately tested with respect to whether they also have enduring effects,^[14] but IPT does appear to have a benign effect on the quality of relationships not found for medications.^[6]

These findings suggest that ADM may not always be the best choice for a first-line treatment depression. Given that the efficacy of ADM has been overstated in the published literature, that at least half of all MDD patients derive little specific pharmacological benefit from medication treatment, and that drugs provide no lasting benefit after their use is discontinued, a case can be made that the cognitive or behavioral interventions are at least as viable as medications as the preferred first-line treatments of choice for depression. That certainly would appear to be the case for patients with less severe depression for whom medications provide no specific pharmacological effect but who would benefit from the enduring effects of the psychosocial interventions without the side effects of medications. In fact, nonspecific processes common to any reasonable intervention may be sufficient to produce response in patients with less severe depressions; the psychosocial interventions are no more likely to separate from ‘‘placebo’’ in such patients than are the ADMs.^[22] For such patients, something appears to be better than nothing and it may no more matter that the therapist be sufficiently expert to produce a specific effect with respect to acute response than that ADM be pharmacologically active. This could mean that most therapists could do a credible job with most patients and that only the more severe (or chronic or comorbid) patients would require an expert therapist.

The situation with respect to patients with more severe depressions is somewhat more complex. The efficacy of CT appears to depend to a considerable extent upon the competence of the therapist, whereas the pharmacologically active component of medication treatment appears to offset a number of problematic behaviors on the part of the prescribing clinician.^[23] In the hands of an expert cognitive therapist, CT may be preferred to ADM because of its enduring effect and freedom from side effects. When a highly competent CT therapist is not readily available then ADM likely is to be preferred. Moreover, there may be advantages to staying on ADM that go beyond the prevention of depression; in the Penn/Vandy study we found that paroxetine had nearly twice the ‘‘true’’ drug effect on neuroticism that it had on depression.^[24] This suggests that ADM may have a specific effect on stress reactivity that makes life better even when depression is not at issue. Patients who are prone to stress reactivity may prefer to start and stay on SSRI medications if it also helps to deal with a life-long propensity toward neuroticism and its associated apprehension and irritability. CT had a similar but smaller effect on neuroticism and may take longer (if ever) to produce an effect of comparable magnitude.

BA should be easier to master than CT and therefore easier to disseminate. Nonetheless, it has yet to be tested outside of the site at which it was developed. Few treatments turn out to be as efficacious as they first appear to be in the initial trials at their centers of origin and it remains to be seen how BA fares at other sites in subsequent studies. It will be especially interesting to see if the initial indications of an enduring effect for BA hold up across multiple replications as they have for CT. That being said, if BA does prove to be as robustly efficacious as medications and as enduring as CT but easier to disseminate, then it might prove to be the optimal first line choice for the treatment of depression regardless of the severity or complexity of the depression.