Behavioural activation (BA)

BA is one of the most accepted and widely used behavioural treatments. Meta analyses found BA to be as effective as cognitive-behavioural therapy, and superior to control conditions (Cuijpers et al. Reference Cuijpers, Van Straten and Warmerdam2007; Mazzucchelli et al. Reference Mazzucchelli, Kane and Rees2009) and to medication (Ekers et al. Reference Ekers, Webster, Van Straten, Cuijpers, Richards and Gilbody2014). Dimidjian et al. (Reference Dimidjian, Hollon, Dobson, Schmaling, Kohlenberg, Addis and Jacobson2006) also showed that BA was superior to cognitive therapy and as effective as medication.

Problem-solving therapy (PST)

PST is a cognitive-behavioural treatment which emerged from behaviour modification research (D'Zurilla & Goldfried, Reference D'Zurilla and Goldfried1971; Nezu et al. Reference Nezu, Nezu and Perri1989). PTS was found to be superior to medication and as effective as other alternative psychosocial therapies and medication (Bell & D`Zurilla, Reference Bell and D'Zurilla2009). However, another meta-analysis showed that the effects of PST vary across studies, ranging from close to zero to very large effects (Cuijpers et al. Reference Cuijpers, Van Straten and Warmerdam2007).

Short-term dynamic psychotherapy (STDP)

STDP emerged from the work of Malan (Reference Malan1976), Mann (Reference Mann1973), Sifneos (Reference Sifneos1979), Davanloo (Reference Davanloo1980) and Luborsky (Reference Luborsky1984). Randomized clinical trials (RCTs) and meta-analyses have shown that STDP is as effective as medication (Barber et al. Reference Barber, Barrett, Gallop, Rynn and Rickels2012), more effective than control conditions, and as effective as alternative treatments at termination and follow-up assessments (Driessen et al. Reference Driessen, Cuijpers, de Maat, Abbass, de Jonghe and Dekker2010, Reference Driessen, Van, Don, Peen, Kool, Westra and Dekker2013, Reference Driessen, Hegelmaier, Abbass, Barber, Dekker, Van and Cuijpers2015a , Reference Driessen, Van, Peen, Don, Kool, Westra and Dekker b , Gibbons et al. Reference Gibbons, Thompson, Scott, Schauble, Mooney, Thompson and Crits-Christoph2012; Abbass et al. Reference Abbass, Kisely, Town, Leichsenring, Driessen, De Maat, Gerber, Dekker, Rabung, Rusalovska and Crowe2014; Fonagy, Reference Fonagy2015; Leichsenring et al. Reference Leichsenring, Luyten, Hilsenroth, Abbass, Barber, Keefe and Steinert2015). Furthermore, combined treatment of STDP and medication was found to be more effective than medication alone (Barber et al. Reference Barber, Muran, McCarthy, Keefe and Lambert2013; Fonagy, Reference Fonagy2015; Leichsenring et al. Reference Leichsenring, Luyten, Hilsenroth, Abbass, Barber, Keefe and Steinert2015).

Long-term dynamic psychotherapy (LTDP)

There is evidence supporting the effectiveness of LTDP and psychoanalysis (de Maat et al. Reference de Maat, de Jonghe, Schoevers and Dekker2009), especially in long-term follow-ups (Knekt et al. Reference Knekt, Lindfors, Härkänen, Välikoski, Virtala, Laaksonen and Renlund2008, Reference Knekt, Lindfors, Laaksonen, Renlund, Haaramo and Härkänen2011; Huber et al. Reference Huber, Henrich, Clarkin and Klug2013). There is also some indirect evidence that LTDP is superior to other psychotherapies in treating complex patients with chronic depression and previous treatment failures (Leichsenring & Rabung, Reference Leichsenring and Rabung2008, Reference Leichsenring and Rabung2010; Fonagy et al. in press). However, some of these conclusions have been challenged (e.g. Smit et al. Reference Smit, Huibers, Ioannidis, van Dyck, van Tilburg and Arntz2012).

Cognitive-behavioral therapy (CBT)

CBT, pioneered by Beck et al. (Reference Beck, Rush, Shaw and Emery1979), became the most widely studied treatment for depression. Robust evidence supports its superiority over control conditions (DeRubeis et al. Reference DeRubeis, Hollon, Amsterdam, Shelton, Young, Salomon, O'Reardon, Lovett, Gladis, Brown and Gallop2005; Tolin, Reference Tolin2010; Cuijpers et al. Reference Cuijpers, Berking, Andersson, Quigley, Kleiboer and Dobson2013a ; Marcus et al. Reference Marcus, O'Connell, Norris and Sawaqdeh2014), and some studies found that combined treatment is superior to medication alone (Cuijpers et al. Reference Cuijpers, Berking, Andersson, Quigley, Kleiboer and Dobson2013a ; Köhler et al. Reference Köhler, Hoffmann, Unger, Steinacher, Dierstein and Fydrich2013; Hollon et al. Reference Hollon, DeRubeis, Fawcett, Amsterdam, Shelton, Zajecka and Gallop2014). When comparing CBT to alternative treatments, there are mixed findings (Hofmann et al. Reference Hofmann, Asnaani, Vonk, Sawyer and Fang2012). While some meta-analyses reported the superiority of CBT (e.g., Tolin, Reference Tolin2010; Marcus et al. Reference Marcus, O'Connell, Norris and Sawaqdeh2014), they have been criticised (Leichsenring & Rabung, Reference Leichsenring and Rabung2010; Baardseth et al. Reference Baardseth, Goldberg, Pace, Wislocki, Frost, Siddiqui and Wampold2013), and others found that it is as effective as alternative treatments (e.g. Leichsenring, Reference Leichsenring2001; Cuijpers et al. Reference Cuijpers, Van Straten, Andersson and Van Oppen2008, Reference Cuijpers, Berking, Andersson, Quigley, Kleiboer and Dobson2013a ; Beltman et al. Reference Beltman, Voshaar and Speckens2010; Driessen et al. Reference Driessen, Van, Don, Peen, Kool, Westra and Dekker2013, Reference Driessen, Van, Peen, Don, Kool, Westra and Dekker2015b ).

Interpersonal therapy (IPT)

Interpersonal psychotherapy (IPT) emerged from Sullivan's (Reference Sullivan1953) interpersonal theory and was later adapted for depression (Klerman et al. Reference Klerman, Weissman, Rounsaville and Chevron1984). In a large RCT, Elkin et al. (Reference Elkin, Shea, Watkins, Imber, Sotsky, Collins and Parloff1989) found that IPT is equally effective to medication and CBT, with some evidence for increased efficacy of IPT for severely depressed patients. There is also preliminary evidence suggesting the superiority of IPT over alternative psychotherapies (Cuijpers et al. Reference Cuijpers, Van Straten, Andersson and Van Oppen2008, Reference Cuijpers, Geraedts, van Oppen, Andersson, Markowitz and van Straten2011), specifically supportive counselling (Barth et al. Reference Barth, Munder, Gerger, Nüesch, Trelle, Znoj, Jüni and Cuijpers2013) and CBT (de Mello et al. Reference de Mello, de Jesus Mari, Bacaltchuk, Verdeli and Neugebauer2005).

Emotion-focused therapy (EFT)

EFT emerged from the humanistic, experiential and client-centred therapeutic approaches (Greenberg & Watson, Reference Greenberg and Watson2006). Experiential therapies were found slightly more effective than non-treatment and control conditions (ES = 0.12; Elliott et al. Reference Elliott, Greenberg, Lietaer and Lambert2004); however, effects largely varied in magnitude and direction. Studies found that adding emotion-focused techniques to client-centred therapy increased the efficacy of treatment (Greenberg & Watson, Reference Greenberg and Watson1998; Goldman et al. Reference Goldman, Greenberg and Angus2006), and that EFT was superior to client-centred therapy in long-term follow-ups (Ellison et al. Reference Ellison, Greenberg, Goldman and Angus2009). Watson et al. (Reference Watson, Gordon, Stermac, Kalogerakos and Steckley2003) found in an RCT that EFT equal to CBT but superior in reduction of interpersonal problems. Overall, while preliminary evidence supports the efficacy of EFT, there is still a need for future research comparing EFT to placebo, control and alternative psychotherapies.

Strengths and weaknesses of meta-analyses

In this paper, we base many of our arguments on data from meta-analyses. While meta-analyses are an important source of information, providing reviews and summaries of large numbers of studies, we are also aware of their limitations. Specifically, the quality of a meta-analysis derives in part from the quality and thoroughness of the studies included in it. Considering quality in determining the conclusion from a meta-analysis is not a simple proposition (Kocsis et al. Reference Kocsis, Gerber, Milrod, Roose, Barber, Thase and Leon2010; Gerber et al. Reference Gerber, Kocsis, Milrod, Roose, Barber, Thase and Leon2011; Thoma et al. Reference Thoma, McKay, Gerber, Milrod, Edwards and Kocsis2012). One can include study quality as a moderator in one's meta-analysis, as some researchers have done; however, this is not common practice and may reflect the authors’ biases. More specifically, if a meta-analysis is not all-inclusive, it may be biased in one direction and could suffer from the author's ‘allegiance’ (see Leichsenring & Rabung, Reference Leichsenring and Rabung2010; Baardseth et al. Reference Baardseth, Goldberg, Pace, Wislocki, Frost, Siddiqui and Wampold2013). Thus, taking those limitations into account, we surmise that meta-analyses provide a rich and valuable source of information, although the specifics require close scrutiny. One important question for the field is whether to prefer the results of a meta-analysis which includes several average quality studies to those of several high-quality trustworthy specific RCTs which suggests the superiority of treatment A over a control condition (Hollon, personal communication, 2015).

Remission rates

The ultimate goal of all treatments of depression is remission. It is often thought that if a treatment was given in an appropriate dose and duration and it did not end with remission, it cannot be considered successful (John Rush et al. Reference John Rush, Trivedi, Wisniewski, Nierenberg, Stewart, Warden and Fava2006; Hollon et al. Reference Hollon, DeRubeis, Fawcett, Amsterdam, Shelton, Zajecka and Gallop2014). However, in depression, it is possible for response and remission to occur without treatment with estimated high rates of spontaneous remission ranging from 23% within 3 months to 53% within 12 months (Whiteford et al. Reference Whiteford, Harris, McKeon, Baxter, Pennell, Barendregt and Wang2013). Therefore, efficacious treatments are expected to result in significantly higher remission rates and/or a much quicker recovery time.

Despite the well-established benefits of treating depression over no treatment, findings from RCTs and naturalistic studies show that remission rates are far from 100% across all treatment modalities. Specifically, one of the largest RCTs (STAR*D), which tested the efficacy of sequential treatment allocation of medication and cognitive therapy, reported an overall remission rate of 27.5–32.9% (Trivedi et al. Reference Trivedi, Rush, Wisniewski, Nierenberg, Warden and Ritz2006), with decreasing rates from 32.9% in step 1 to 14.7% in step 4 (John Rush et al. Reference John Rush, Trivedi, Wisniewski, Nierenberg, Stewart, Warden and Fava2006). These findings clearly call for a need to improve our treatment modalities.

In fact, in face of the low remission rates, NIMH director Insel (Reference Insel2009, Reference Insel2006), emphasised a pressing need to develop new approaches to diagnosing and treating depression. As a result, most NIMH funding resources have shifted towards researching new neurobiologically based approaches to mental illnesses using NIMH's new diagnostic system – the Research Domain Criteria (RDoC) (e.g. Cuthbert & Insel, Reference Cuthbert and Insel2013). In the context of treating depression, the NIMH has been moving towards the development of personalised interventions based on the specific needs and circumstances of each individual patient.

Drop-out rates

Premature discontinuation, also referred to as dropout, is a major concern as it impedes the effectiveness of treating depression (Cahill et al. Reference Cahill, Barkham, Hardy, Rees, Shapiro, Stiles and Macaskill2003). While dropout rates vary, premature discontinuation is a problem shared by all psychotherapies. A recent meta-analysis found a mean rate of 17.5% dropout in therapies for depression, with a large variance across studies ranging between 0 and 50% (Cooper & Conklin, Reference Cooper and Conklin2015). Another meta-analysis (Flückiger et al. Reference Flückiger, Del Re, Munder, Heer and Wampold2014) found 29.2% drop-out rate in evidence-based treatments (EBT) for acute depression and anxiety disorders (compared with 32.7% rate in treatment-as-usual control conditions (TAU)). Meta-analyses comparing dropout rates in different psychotherapies have yielded mixed findings, from equal dropout rates across treatments (Cooper & Conklin, Reference Cooper and Conklin2015), to significant differences between psychotherapies, with the highest rate in CBT and the lowest rate in PST (Cuijpers et al. Reference Cuijpers, Van Straten, Andersson and Van Oppen2008). Taken together, these findings demonstrate the strong need to develop new treatment models tailored to patients’ needs.

Enduring effects of psychotherapy

While ample evidence supports the efficacy of psychotherapies for depression at the endpoint of treatment, much less is known about their enduring effects. Some researchers found that different psychotherapies are effective in preventing relapse/recurrence (Hollon et al. Reference Hollon, Stewart and Strunk2006; Hollon & Ponniah, Reference Hollon and Ponniah2010; Steinert et al. Reference Steinert, Hofmann, Kruse and Leichsenring2014; Biesheuvel-Leliefeld et al. Reference Biesheuvel-Leliefeld, Kok, Bockting, Cuijpers, Hollon, van Marwijk and Smit2015) and that their enduring effects are as effective as a continuous use of medications (Cuijpers et al. Reference Cuijpers, Hollon, van Straten, Bockting, Berking and Andersson2013b ). In contrast, others have shown that the superiority of EBT for acute depression and anxiety disorders over TAU is minimal on 0–4-month follow-ups, and does not extend at 8–12-month follow-ups (Flückiger et al. Reference Flückiger, Del Re, Munder, Heer and Wampold2014).

These mixed findings indicate the need to study the long-term effects of psychotherapy (Lambert, Reference Lambert and Lambert2013). The lack of evidence may largely be due to the major limitations of follow-up research. First, in some countries, there is no detailed information about patients’ treatment utilisation and researchers need to rely on self-reports, which are often biased. Second, patients’ frequent reluctance to participate in follow-ups often results in a large amount of missing data.

Treatment resistant depression

Accumulating evidence has shown that depression is a pervasive and chronic disorder and that many patients do not benefit from standard treatments. Researchers have identified a 43% recurrence of episodes among depressed young adults and 5–19% of suicide attempts among depressed patients aged 5–30 (Rohde et al. Reference Rohde, Lewinsohn, Klein, Seeley and Gau2012). These findings have led investigators to focus on ‘treatment-resistant’ forms of depression (Taylor et al. Reference Taylor, Carlyle, McPherson, Rost, Thomas and Fonagy2012). Studies attempting to identify patient characteristics, which are more highly correlated with – or predictive of – high dropout rates and/or poorer outcome, have resulted in mixed findings. Some researchers found comorbidity of personality disorders and social racial minority status to be predictive of higher dropout rates (Cooper & Conklin, Reference Cooper and Conklin2015). Others reported that these characteristics did not moderate the superiority of EBT over TAU, but rather identified additional moderators such as comorbidity of substance abuse (Flückiger et al. Reference Flückiger, Del Re, Munder, Heer and Wampold2014). These contrasting findings demonstrate the need to develop more effective ways to treat non-responders (Barrett et al. Reference Barrett, Chua, Crits-Christoph, Gibbons and Thompson2008).

A suggestion for an alternative stepwise sequential model of allocation of treatment for depression

This paper reviewed the significant progress in the development and research of EBT for depression in the past few decades. However, evidence of high dropout rates, large non-responsive populations and mixed findings regarding the enduring effects of psychotherapy suggest that there is still much to be learned about the treatment of depression.

We believe that there is a need to develop a stepwise sequential approach for a more cost-effective and applicable treatment modalities that will integrate different evidence-based therapies while taking into account patients’ characteristics and needs. Similar to treatment modalities in advanced individualised medicine, we encourage clinicians to use their clinical judgment and to make decisions based on session-to-session feedback and continuous awareness of the patient's progress. Medication can be incorporated into each phase of treatment; however, this is not the focus of our model at this point. Our suggestion is preliminary, stating that more research needs to be done in order to determine the specific use of medication (see Fig. 1 for the outline of the model).

Fig. 1. A suggestion for an integrative stepwise sequential model for treatment of depression. Note: BA, behavioural activation therapy; PST, problem-solving therapy; CBT, cognitive-behavioural therapy; IPT, interpersonal therapy; EFT, emotion focused therapy; STDP, short-term dynamic psychotherapy; LTDP, long-term dynamic psychotherapy.

Step 1: BA or PST. The treatment process should begin with a time-limited focused therapy such as BA or PST. Since these treatments are brief, cost-effective and have proven to be successful for large depressed populations, it would be reasonable to use them as a first line of treatment.

Step 2: CBT, IPT, EFT or STDP: In case of non-response, we suggest moving to the second phase – short-term treatments: CBT, IPT, EPT or STDP. Given the evidence from RCTs and meta-analyses, we believe that there is a great likelihood that these treatments will be equally effective and appropriate for patients who have failed to improve in briefer and more symptom-focused treatments. There is, however, some evidence that specific subgroups of patients benefit more from specific treatments (Barber & Muenz, Reference Barber and Muenz1996; Sotsky et al. Reference Sotsky, Glass, Shea, Pilkonis, Collins, Elkin and Oliveri2006; Barber et al. Reference Barber, Barrett, Gallop, Rynn and Rickels2012; Beutler et al. Reference Beutler, Forrester, Gallagher-Thompson, Thompson and Tomlins2012; DeRubeis et al. Reference DeRubeis, Cohen, Forand, Fournier, Gelfand and Lorenzo-Luaces2014). If replicated, these findings should be included in the model.

Step 3: LTDP or Long-term CBT: Complex patients, who suffer from chronic depression and have shown two treatment failures in steps 1 and 2, will be referred to long-term CBT or LTDP. This suggestion is based on the promising evidence supporting the efficacy of LTDP for complex patients who experienced treatment failures and impressive RCT findings showing high remission rates (80.1%) after 18 months of combined therapy (cognitive therapy + medication) (Hollon et al. Reference Hollon, DeRubeis, Fawcett, Amsterdam, Shelton, Zajecka and Gallop2014).

Important considerations

The suggested model should be adjusted according to the therapist's clinical judgment, the patient's preferences and progress and the limitations of the clinical settings (such as the availability of therapists from a specific approach). Additionally, we expect that computer-based treatments will be included in this model in the future, given the increasing evidence supporting their efficacy and their advantages (high cost-effectiveness and high accessibility to a range of populations).

Lastly, we expect that future adjustments will be applied to the model based on the accumulating evidence on moderators and predictors of treatment success. Some preliminary findings have shown that specific populations benefit from specific treatments (e.g., Barber & Muenz, Reference Barber and Muenz1996; Sotsky et al. Reference Sotsky, Glass, Shea, Pilkonis, Collins, Elkin and Oliveri2006; DeRubeis et al. Reference DeRubeis, Cohen, Forand, Fournier, Gelfand and Lorenzo-Luaces2014). For example, Barber et al. (Reference Barber, Barrett, Gallop, Rynn and Rickels2012) found that depressed minority men (mostly African-American) benefited more from short-term supportive-expressive therapy than placebo or medication. We recommend that future studies focus on replicating these previous findings and identifying more significant factors affecting treatment success.