Are antidepressants and psychotherapy equally effective in treating depression? A critical commentary

Deciding on the optimal treatment plan is crucial for individuals with depressive disorders. Specifically, pharmacotherapy (most commonly with antidepressant medications) and psychotherapy are far and away the two most heavily researched and available treatment modalities. Several comparator trials and meta-analyses have attempted to compare antidepressants and psychotherapy to determine which might provide more symptom relief, typically using clinician-rated symptom inventories such as the Hamilton Rating Scale for Depression (HRSD). By and large, the results of these meta-analyses indicate that short-term acute treatment of major depressive disorder yields equivalent results on average between pharmacotherapy and psychotherapy (Casacalenda et al., 2002; Cuijpers et al., 2008, 2013a,b; Huhn et al., 2014; Imel et al., 2008; Spielmans et al., 2011), and pharmacological treatment may be superior to psychotherapy in the treatment of dysthymia (Cuijpers et al., 2008, 2013b; Imel et al., 2008). Further, a meta-analysis of trials comparing psychotherapy to pill placebo (Cuijpers et al., 2014b) yielded an overall therapy–placebo effect size comparable to the effect sizes found in meta-analyses of antidepressant–placebo differences (Kirsch et al., 2008; Sugarman et al., 2014; Turner et al., 2008). Khan et al. (2012) found no significant differences between psychotherapy, antidepressants and alternative therapies, such as exercise and acupuncture, and argued that simply enrolling patients in an active therapeutic program may be more important than the specific type of intervention.

However, acceptance of this apparent equivalency depends on the assumption of ecological validity of trial design. That is, implying that patients will obtain equivalent results on average is valid only if the level of real-world therapeutic care is consistent with the clinical trials. There are several design features of clinical trials that might limit the generalizability of the findings and distort trial outcomes. This commentary will explore the role of intensive clinical management as it pertains to the improvement in clinical trials and concludes with recommendations for improving future comparator trials.

Individuals in the psychotherapy arms of clinical trials typically complete sessions once a week where they receive support, learn skills and monitor their symptoms, which is similar to what they are likely to encounter in actual clinical settings. In clinical trials of pharmacotherapy, patients complete an average of 0.82 follow-up visits lasting about half an hour for each week they are in the trial, corresponding to about five visits every six weeks (Iovieno et al., 2012). These visits serve the purpose of tracking clinical progress over the course of the trial through clinical interviews as well as monitoring for adverse effects and adjusting dosage, if permitted by the study protocol.

The standard level of clinical care in community settings for patients taking antidepressants is drastically less intense than in clinical trials. An analysis of a retrospective cohort of 193 151 adults from 1998 through 2005 revealed that about 40% of patients had a minimum of three provider contacts in the 12 weeks following initiation of antidepressant therapy, and fewer than 5% had seven or more contacts (Morrato et al., 2008). That is, 60% of patients completed less than one clinical follow-up per month, and almost no patients completed 10 visits over 12 weeks, which is the average for clinical trials (Iovieno et al., 2012). Another analysis examined 84 514 adults in the northeastern US between 2001 and 2003 following initiation of antidepressant treatment (Stettin et al., 2006). They found that the average number of outpatient medical visits during the first 12 weeks was 4.45, although only 0.97 of the visits were specifically with a mental health professional. Only 24% of individuals had a mental health appointment in the first 12 weeks. Moreover, this study was conducted with patients with private health insurance, and the authors note that the frequency of clinical monitoring is likely much lower in uninsured and Medicare populations.

In summary, the vast majority of patients do not receive the same level of clinical attention as in antidepressant trials. In fact, about 74% of patients on antidepressants receive their prescriptions from a general care practitioner rather than from a psychiatrist (Mojtabai & Olfson, 2008), and general care practitioners have been labeled as the “de facto” mental health system (Wennerstrom et al., 2015). However, these practitioners might not have the resources nor the expertise to provide the same level of clinical management as a dedicated mental health professional. Indeed, patients who receive their prescription from a general practitioner are more likely to stop taking their prescriptions within 30 days and to take the medication irregularly (Mojtabai & Olfson, 2008).

This less intensive clinical management compared to clinical trials is concerning because the therapeutic setting directly contributes to the placebo components of treatment (Rutherford & Roose, 2013), which appears to account for about half of the overall drug response in clinical trials (Kirsch & Sapirstein, 1998). Posternak & Zimmerman (2007) demonstrated that two extra follow-up visits during six-week double-blind trials resulted in an additional 1.6-point reduction on the HRSD for patients receiving antidepressants and 1.75 points for patients who received placebo. The additional two visits at weeks three and five accounted for approximately 40% of the overall placebo response between weeks two and six. The authors concluded that the therapeutic impact of multiple follow-up sessions is cumulative. Another meta-analysis of 146 clinical trials (Iovieno et al., 2012) found that the number of follow-up visits (especially those late in the trial) increased the chance of a favorable patient response to antidepressants. This relationship between follow-up visits and symptom reduction has been replicated in observational studies (Naudet et al., 2011) and in adolescents (Rutherford et al., 2011). Thus, follow-up visits appear to have a non-trivial impact on the overall level of improvement during clinical trials.

For patients, interacting with physicians and study personnel can be rich opportunities for social interactions and developing relationships, contrasting the isolation and relative lack of activity that can often accompany depression. Further, discussing their symptoms, improvements and experience of side effects in a supportive medical environment during clinician-rated measures, such as the HRSD, can be therapeutic in itself and potentially instill hope and their faith in the treatment (Rutherford & Roose, 2013). Attending several follow-up visits during the course of clinical trials enhances support and clinical improvement. Receiving support and developing therapeutic relationships with clinicians have been identified as major common factors of improvement in mental health treatment (Lambert & Ogles, 2004), and may obscure the comparison with psychotherapy. These factors may partially account for why individuals given a pill placebo experience about 80% of the improvement as those given antidepressants in clinical trials (Kirsch et al., 2008). Simply put, results from pharmacotherapy arms of clinical trials describe improvement based on the combined effects of medication and intensive clinical management. If patients do not experience a similar level of clinical attention in actual practice, as is the case with the 74% of patients who receive prescriptions through a general practitioner, they are unlikely to experience optimal outcomes.

Providing less clinical contact in real-world settings limits the ecological validity of comparator trials between antidepressants and psychotherapy. This issue has been acknowledged by researchers on the topic but has not received adequate discussion in the literature. Cuijpers et al. (2008) noted in their meta-analysis that “unrealistic monitoring” of patients through weekly visits “may have distorted the actual differences between psychological and pharmacologic treatments” and that their results “should be considered with caution (p. 1684)”. However, the results from this study are cited in other manuscripts from this group as evidence that psychotherapy and pharmacotherapy are about equally efficacious in treating depression, with no mention of this limitation (Cuijpers et al., 2013b, 2014b). Spielmans et al. (2011) also noted in their meta-analysis that the frequency of face-to-face contact in most studies was much larger than in realistic clinical practice, and challenged the generalizability of comparator trials.

A recent commentary (Severus et al., 2012) suggested ways in which clinical trials could more closely mirror real world settings. Although the authors correctly posited that frequent contact with clinicians and close monitoring of symptoms could lead to beneficial effects on their own, they make no mention that this type of management is not representative of actual clinical practice in most settings. Another example of the questionable ecological validity of pharmaceutical trials is the highly cited Sequenced Treatment Alternatives to Relive Depression (STAR*D) study (Trivedi et al., 2006). This study was specifically designed to generalize to clinical practice by allowing for dose titration, switching of drugs for patients who did not respond to initial treatment and not excluding patients with comorbid medical and psychiatric conditions. However, the study protocol also utilized an intense frequency of clinical monitoring, with patients completing an average of 4.8 study visits during the first 12 weeks of treatment. Their protocol called for “didactic instruction, ongoing support and guidance by the clinical research coordinator, the use of structured evaluation of symptoms and side effects at each visit and a centralized treatment monitoring and feedback system (p. 30)”. Thus, the results might not generalize to practices that do not have as frequent clinical monitoring or patient opportunities for support.

Based on the available evidence, it appears that the level of improvement in antidepressant arms of clinical trials represents an overestimate compared to real-world clinical settings due to the enhanced patient management in these trials. If patients would like to take an antidepressant, it should be in an environment where they can receive intensive clinical management akin to clinical trials, and preferably paired with psychotherapy (Cuijpers et al., 2014a). Being in a clinical trial of psychotherapy may provide some extra attention, such as completing clinician-rated interviews like the HRSD in addition to therapy sessions. However, the degree of difference between the clinical trials and real-world settings is substantially lower than in pharmacotherapy arms, because the provision of supportive attention is an essential specific component of psychotherapy (Lambert & Ogles, 2004). This argument challenges interpretation of data indicating that pharmacotherapy and psychotherapy are equivalently effective in acute management of depression, as is often cited in the literature. The following is a brief list of recommendations to improve the ecological validity of future comparator trials:

1. Reduce frequency of clinical management for pharmacotherapy. As described, the standard level of clinical attention for patients taking antidepressants is much lower in real-world settings, with the average patient completing only about one visit with a mental health professional during the first 12 weeks of treatment (Stettin et al., 2006). Clinical trials have not yet evaluated whether this level of care is a viable treatment option or equivalent to psychotherapy. Although this infrequent clinical monitoring is inconsistent with recommended treatment guidelines for psychiatric treatment, it is clear that it has become a standard model of practice in many settings. Thus, a potential trial design could include randomizing patients to either receive an antidepressant prescription or 12 weekly psychotherapy sessions. Symptom measurements would only be collected at baseline and at the end of the study, as collecting these outcome data may one of the mechanisms of improvement. This model of treatment would provide a similar level of care that most patients receive in the real world, and could be ethically justified as such. To ensure the safety of the patients in these trials, emergency resources could be made available for patients with more severe psychopathology who might require more intense management.

2. Conduct trials in a single-blind manner. An additional recommendation for future comparator trials between pharmacotherapy and psychotherapy is to conduct trials in a single-blind manner. That is, some comparator trials include a pill placebo condition (e.g. DeRubeis et al., 2005), whereas others trials simply include a randomization to active medication or psychotherapy (e.g. David et al., 2008). Pill placebos are a crucial component of evaluating the specific effects of a pharmacological treatment, but are not necessary in the comparison between pharmacotherapy and psychotherapy. In comparator trials, it is neither possible nor necessary to blind patients to whether they are receiving psychotherapeutic or pharmacologic intervention. However, with a pill placebo control, participants receiving medications are blinded as to whether they are receiving an active treatment, while those receiving active psychotherapy are not blinded. Inclusion of a pill placebo condition creates a bias against the active medication in these trials, as effect sizes are typically larger for comparisons between two antidepressants than between an antidepressant and a placebo (Sneed et al., 2008), likely because patients are assured that they received an active medication. This design feature also limits the ecological validity of the studies, as patients in real-world settings do not need to wonder whether they are receiving an antidepressant or a placebo.

Assessors collecting outcome data (such as interviews used to determine HRSD scores) should always be blinded to participant assignment to reduce bias, hence the term “single-blind”. Importantly, assessors should complete surveys for possible unblinding at the end of the study to ensure this lack of bias. Including patient-reported outcome measures such as the Beck Depression Inventory in addition to (or instead of) clinician-rated measures could also reduce assessor bias. Additionally, completing clinician-rated measures such as the HRSD through patient interviews may provide some of the essential components of psychotherapy, as described previously. In particular, reporting side effects for antidepressants should be conducted through patient surveys, as interviews with clinicians result in a severe underreport of symptoms (Zimmerman et al., 2010), especially for sexual dysfunction (Khin et al., 2015; Serretti & Chiesa, 2009).

3. Include long-term follow-up. A final recommendation is for all comparator trials to include long-term naturalistic follow-ups. Primary endpoints for acute treatment studies usually occur anywhere between 6 and 26 weeks (Spielmans et al., 2011), but long-term follow-ups provide a more comprehensive picture of which modalities may yield optimal outcomes. Spielmans et al. (2011) found that despite the finding of short-term equivalency in these trials between “bona fide” psychotherapies and antidepressants, long-term follow-up after 18–40 weeks indicated superiority for psychotherapy. Cuijpers et al. (2013a) found that patients treated with cognitive behavioral therapy had superior outcomes compared to antidepressant treatment that was discontinued during follow-up periods ranging from 6 to 18 months. Outcomes were roughly equivalent when comparing cognitive behavioral therapy to continued pharmacotherapy. Imel et al. (2008) also found that psychotherapy yielded superior results compared to pharmacotherapy at long-term follow-up with an average interval of 15 months. Moreover, length of follow-up was a significant moderator such that the advantage of psychotherapy over medication was superior at longer follow-up intervals. Despite the apparent consensus amongst these analyses, these long-term follow-ups are included for a minority of comparator trials, and more data would increase the reliability of these findings.

The long-term superiority for psychotherapy could perhaps be due to the skill-building components of psychotherapy, as suggested by Imel et al. (2008), or due to a higher risk for relapse following antidepressant treatment. While it certainly makes logical sense that the specific skills developed in psychotherapy could be evident in preventing relapse after acute treatment, additional evidence suggests that antidepressants may have potentially iatrogenic effects in long-term treatment. For example, one study (Babyak et al., 2000) examined exercise, sertraline and combination therapy in the treatment of major depression. Although all three groups had approximately equivalent outcomes at the 4-month study endpoint, the exercise alone group had the lowest relapse rate 6 months later. That is, the benefits of exercise appeared to be reduced by the addition of sertraline treatment. A similar finding was found in a study of cognitive-behavioral therapy, imipramine and their combination in the treatment of panic disorder (Barlow et al., 2000). Six months after treatment discontinuation, individuals in the psychotherapy plus placebo group had significantly higher response rates than the psychotherapy plus imipramine group. Again, in this study, the addition of pharmacotherapy appeared to detract from the long-term benefits of other treatments. Further studies and replication could help to provide more evidence to guide long-term treatment recommendations.

In conclusion, the prevailing wisdom that antidepressants and psychotherapy yield equivalent outcomes in acute treatment of depression may be flawed due to the lack of ecological validity in trial design. For patients, taking a medication without receiving regular clinical management may detract from the potential benefit of the treatment. A primary suggestion to better mirror clinical practice in comparator trials is to reduce the frequency of clinical management, which includes enhanced attention, support and behavioral activation, some of the key mechanisms for improvement in psychotherapy. Additionally, more long-term follow-ups will help in determining optimal treatment recommendations. Importantly, this design feature does not invalidate the differences between antidepressants and pill placebo, because both the groups receive equivalent amounts of clinical management in these trials and should receive the same benefits from this attention. Finally, although the current commentary focused on the treatment of depression, the issue of unrealistically intense clinical management is prevalent across clinical trials for many mental disorders (e.g. Barlow et al., 2000; Foa et al., 2005).

Declaration of interest

The author reports no conflicts of interest. The author alone is responsible for the content and writing of this article.