Definition of TRD

TRD can be thought of as describing patients beyond a certain threshold in a course of serial treatment trials. There is, however, no universally accepted criterion for this threshold. A systematic review identified 150 randomized controlled trials, observational studies, and case series of patients defined as having “TRD” published between January 2016 and December 2017 ( ). The most common element of the definition used in these studies was a failure to respond to two different treatments, though this was the case in only just over half of the definitions (50.3%: ( ), see Fig. 1.1 ).

Number of failed trials required for inclusion in studies of patients with “treatment-resistant depression.”

(Data taken from Brown, S., Rittenbach, K., Cheung, S., McKean, G., MacMaster, F.P., Clement, F., 2019. Current and common definitions of treatment-resistant depression: findings from a systematic review and qualitative interviews. Can. J. Psychiatry 64(6), 380–387. https://doi.org/10.1177/0706743719828965 .)

However, there is a complete lack of consensus regarding how different classes of medication are considered, the length and dose of treatment trial required, how nonpharmacological interventions are considered, whether the trials needed to be in the current episode or not, what the target outcomes of trials is, how this should be assessed, and whether trials in past episodes count or not ( ; ). A systematic review has also highlighted the observation that only 17% of interventional studies recruited populations of patients meeting the most frequently specified criteria of two failed trails ( ). This leads to a major issue with regard to the evidence base for the management of TRD. A systematic review of all psychological and pharmacological augmentation strategies was only able to evaluate 28 randomized controlled trials (RCTs) including just 18 different treatments in participants meeting the TRD criteria of prior failure of two previous treatments, despite a wide range being widely used clinically ( ).

So how should TRD be defined? One possible approach is to examine whether there is evidence of factors that distinguish patients with TRD compared with other patients with depression. Considering the data from the largest randomized trial in depression, the STAR*D study, it has been suggested that there is a marked change in remission rates between second and third step treatments supporting a view that TRD should be defined on the basis of two failed treatments ( ) (see Fig. 1.2 ).

Remission rates to first-, second-, third-, or fourth-line pharmacological treatments.

(Data from Rush, A.J., Trivedi, M.H., Wisniewski, S.R., Nierenberg, A.A., Stewart, J.W., Warden, D., Niederehe, G., Thase, M.E., Lavori, P.W., Lebowitz, B.D., McGrath, P.J., Rosenbaum, J.F., Sackeim, H.A., Kupfer, D.J., Luther, J., Fava, M., 2006. Acute and longer-term outcomes in depressed outpatients requiring one or several treatment steps: a STAR*D report. Am. J. Psychiatry 163(11), 1905–1917.)

However, in a meta-analysis of over 4500 patients, a dichotomous or even a highly variable response to antidepressants does not seem to be the case, challenging the notion that there are distinct subgroups of treatment-responsive and treatment-resistant patients with depression ( ). Big data sets can be used to identify factors associated with proxy markers of TRD, for example, the need for using ECT, vagus nerve stimulation (VNS), or deep brain stimulation (DBS). One such study identified that such patients could be defined by having received three or more antidepressants, or one or more antipsychotics, in the preceding year ( ). However, the argument is somewhat circular in defining TRD on the basis of use of certain (neurostimulatory) treatments, and the derived criteria will vary between different health care environments whose treatment algorithms differ. The reality is that there are likely to be a multitude of reasons why some patients respond to medication and others do not, including pharmacokinetic and pharmacodynamic differences, and there are no clearly defined symptomatology-based, phenotypic, or biological boundaries between TRD and depression that is not treatment resistant ( ), precluding this as an avenue through which to define TRD—at least at the present time.

What about from a regulatory perspective? While regulators do set out some outline criteria, what companies do is determined at least in part by business considerations. Intranasal esketamine has received a license in many parts of the world for augmentation of antidepressants in patients with TRD. This was on the basis of pivotal RCTs in which TRD was defined as a failure to respond to at least one antidepressant in the current episode assessed retrospectively, plus failure of a different antidepressant taken at an adequate dose for a total duration of at least 6 weeks observed prospectively ( ). No response was defined as less than a 25% improvement in symptoms and adequacy of dose of drug assessed using the Massachusetts General Hospital antidepressant treatment response questionnaire ( ). Currently, there are no other drugs licensed in Europe for TRD. An alternative approach was taken by the manufacturers of a long-acting formulation of quetiapine who obtained a license for its use in patients with “sub-optimal response” to antidepressants based on pivotal studies that recruited participants whose depression had failed to respond to just one or more antidepressants ( ). The decision between these two different positions is largely governed by marketing factors as well as a belief that the more previous treatments have been ineffective, the less likely a subsequent one is to work, based on the data from the STAR*D study ( ).

Ultimately, the definition of TRD is arbitrary. In an attempt to address this, a European-funded partnership with industry has drawn together a number of international experts in the management of depression from across the globe and, using a Delphi technique, attempted to arrive at a consensus definition of TRD particularly, but not exclusively, for regulatory studies ( ). There was strong consensus that the definition should be failure of at least two prior antidepressant treatments but only weak consensus that this should be based on treatments in the current episode or the last 2 years, whichever is longer. There was moderate consensus that an adequate trial of a medication was a minimum of 4 weeks at a minimally effective dose and that the minimum of two failed treatments needed to include two established treatments with different mechanisms of action, though how the latter is defined was not detailed. The consensus was that trials in TRD could include patients who had received psychotherapy or neurostimulation treatments, such as transcranial direct current stimulation (tDCS), repetitive transcranial magnetic stimulation (rTMS), and ECT, but not VNS or DBS. There was also consensus that trials should exclude comorbid personality disorders, other mental disorders, and active and severe substance misuse ( ). If this consensus is adopted, then it will lead to more consistency in studies allowing for comparisons of treatments, but the downside is that the data will be less generalizable given high rates of comorbidity of all types in TRD populations ( ; ).

Most definitions of TRD focus exclusively on unipolar depression, though some studies in TRD do include patients with a bipolar diagnosis ( ). There is much less consistency regarding the definition of treatment-resistant bipolar depression (TRBD). The most common definition in the recent literature is failure of one treatment known to be effective ( ). However, a group of experts in the management of bipolar disorder, using a Delphi technique, arrived at a consensus definition of TRBD as failure to reach sustained symptomatic remission for eight consecutive weeks after two different treatment trials, at adequate therapeutic doses, with at least two recommended monotherapy treatments or at least one monotherapy treatment and another combination treatment ( ). This has been operationalized in an ongoing trial as failure to respond to (or tolerate) two out of four treatments included in current guidelines for the treatment of bipolar depression ( ): quetiapine, olanzapine, lamotrigine, and lurasidone ( ).

Staging treatment-resistant depression

“Staging” TRD refers to operationalizing the differentiation of TRD from treatment-responsive depression and/or quantifying or categorizing the degree of treatment resistance. Several staging tools exist that have various pros and cons and are well reviewed by and . The original purpose of these tools varies, as some being more orientated toward use in research studies and others being more intended for clinical practice. See Table 1.1 .

The oldest staging tool is the Antidepressant Treatment History Form (ATHF) first published in 1990 ( ; ), with several updates, most recently in 2019, in a short-form version (ATHF-SF) ( ). The tool was originally introduced to assess degree of resistance in patients receiving ECT and is oriented toward use in research studies. In its original form, the tool assesses each treatment trial on a scale of 1–5, with scores of 1 and 2 indicating an inadequate trial and 3 and above being adequate, with higher scores indicating higher doses and/or augmentation of a base antidepressant. The tool is most typically used to describe the total number of “failed” (irrespective of adequacy) and “failed adequate” trials a patient has received. It is possible to sum the scores across treatment trials and describe the highest score for any trial administered, though these are rarely reported ( ). The new ATHF-SR simplifies matters by defining a trial simply as adequate or not. For a trial of pharmacotherapy to be counted as adequate it needs to have been taken for at least 4 weeks at a dose specified in the ATHF-SR. This new version also collects information regarding neurostimulatory and psychotherapeutic treatment trials and defines what counts as adequate trials for these. As a result, the number of failed and number of failed adequate trials can be totaled or separated out by drug class and treatment modality. It is the only commonly used staging tool that has separate and specific criteria for psychotic and bipolar depression. For example, for a trial of an antidepressant to count as adequate in a patient with psychotic depression, it must have been combined with an adequate dose of an antipsychotic. The 2019 paper includes data collection tools, checklist, and instructions in supplementary material that is freely available for use.

The “Massachusetts General Hospital Staging model (MGH-s)” or “Antidepressant Treatment Response Questionnaire (ATRQ)” was first published by and updated in 2011 ( ). It is similar but somewhat simpler to use than the ATHF. Apart from ECT, it does not include any treatment other than pharmacotherapy. An adequate trial of an antidepressant is deemed to be at least 6 weeks (rather than 4 as in the ATHF) with the minimum dose for trials defined in the tool. Each failed trial in the current episode scores 1 point. As with the ATHF, there is no implicit hierarchy of efficacy of antidepressants nor preference for between- versus within-class switches. However, unlike the ATHF-SF, an account is taken of optimization of treatment trials with an extra 0.5 points added if the trial extended to at least 10 weeks and if the dose is increased to or above a specified “maximum” dose. Augmentation is dealt with by adding a further 0.5 points for any antidepressant that was augmented with at least one other drug. Only examples of possible augmentation agents are given, meaning that when used in research studies, these need to be operationally defined ( ), and it is unclear why augmentation with a medication supported by positive RCT evidence scores the same as the antidepressant trial being extended from 6 to 10 weeks. A failed trial of ECT adds a further 3 points, though it is unclear as to how this score has been arrived at. Nevertheless, despite these shortcomings, the ATRQ relatively straight forwardly produces a continuous score which might help differentiate degrees of treatment resistance.

The Thase and Rush staging model (TRSM) was developed as a guide for clinical practice ( ). Rather than attempting to provide a continuous score for degree of treatment resistance, patients are categorized into one of five “stages.” Stage I includes individuals whose depression has not responded to at least one adequate trial of an antidepressant; Stage II if nonresponse to at least two antidepressant from two distinct pharmacological classes; Stage III as Stage II plus failure of a trial of a tricyclic antidepressant (TCA); Stage IV as Stage III plus a failure of a trial of a monoamine oxidase inhibitor (MAOI); and Stage V as Stage IV plus failure of a course of bilateral ECT. The TRSM is both simple to use and implies a treatment algorithm for patients who don’t respond to first line treatment. However, what constitutes an adequate trials is not defined and the TRSM has also been criticized on a number of grounds ( ), including the implication that failure to respond to two different classes of antidepressants is harder to treat than failure to respond to two drugs of the same class, despite the evidence suggesting little difference in response rates when comparing between versus within class switches ( ). It also implies a hierarchy of efficacy of different classes of antidepressants for which there is little evidence, and does not take into account augmentation strategies for which there is an evidence base ( ).

Like the TRSM, the European Staging Model (ESM) does not produce a continuous score, but rather categorizes patients into three groups: “nonresponders,” defined as having had a nonresponse to one adequate trial of an antidepressant or ECT; those with “TRD,” defined as failure of response to two or more antidepressant trials from different classes; and those with “chronic resistant depression (CRD)” who have been resistance to several antidepressants, including augmentation strategies, over a period of at least 12 months ( ). The TRD subgroup is subdivided by total duration of failed trails from 12 to 16 weeks for TRD1 to 36 weeks to 1 year for TRD5. As with the TRSM, the ESM has a number of weaknesses ( ). It is far from clear why total duration of treatment trials takes on such an important element in categorizing degree of resistance. While it is true that duration of untreated episode of depression is inversely related to the probability of remission ( ), it is also clear that probability of remission decreases with each treatment step ( ) though this is not taken into account. Similarly, the use of augmentation agents is only included in the definition of CRD and not the various degrees/stages of TRD.

The simplest staging tool developed for clinical usage is that published by . This proposes two stages based on the findings from the STAR*D study: Stage I and Stage II TRD. Stage I is defined as “failure of 2 adequate dose-duration antidepressants or psychotherapy from different classes (either in combination or succession) in the current episode,” while Stage II is defined as “failure of 3 or more adequate antidepressant or psychotherapy trials from different classes … .” This proposal was born out of an attempt to consider where in the treatment algorithm newer, more expensive, invasive and/or less safe treatments might be considered (see later).

The Maudsley Staging Model (MSM) takes a slightly different approach to the other tools ( ). It produces a continuous score using a simple scoring system, making it usable in routine clinical practice. What makes the tool different is that it incorporates duration and baseline severity of the current episode into the score. The total score is made up of a score of 1–3 for duration, 1–5 for baseline severity, 1–5 for number of antidepressant failures (a score of 1 for 1–2 failures and a score of 5 for more than 10 failures), with an additional 1 point if any augmentation has been used and 1 for a trial of ECT, giving a total of between 3 and 15. While there are criticisms of the MSM, including that the importance of augmentation is minimized, other neurostimulation and psychotherapy is not considered and that duration of episode and number of trialed medication are not independent variables ( ), there is evidence that the MSM has utility in being able to predict patient outcomes. This was originally shown in a sample of 62 patients followed up naturalistically over an average of 2.5 years ( ) and then replicated in a sample of 643 patients assessed in the Netherlands Study of Depression and Anxiety at baseline and after 12 months ( ). A Dutch extension of the MSM has been proposed ( ) with additional items of degree of functional impairment (rated on the Global Assessment of Function scale and scored 0–3), comorbid anxiety (scored 0, 0.5, or 1), comorbid personality disorder (scored 0, 0.5, or 1), psychosocial stressors (scoring 0 or 1), use of pharmacological augmentation (scored 0–3), use of psychotherapy (scored 0–2), and intensity of treatment (i.e., outpatient, day patient or inpatient, scored 0–2). This extension of the MSM is reported as outperforming the original MSM in prediction of future depressive symptomatology ( ).

In terms of thresholds, there is some evidence of differences in the performance of the tools. In a Swedish registry study of 127,108 patients on antidepressants, 19% were identified as having TRD using the ATRQ, 15% by the MSM and 9.5% by the ESM ( ). There was little in the way of differences in the characteristics of the patients identified by each tool.

There is no consensus as to which staging tool should be used in different circumstances. While tools such as the ATHF and ATRQ can aid systematic assessment of past treatments, staging tools were used to define TRD in only just over a third of studies (38.1%), the most commonly used being the TRSM, in a systematic review ( ). The ATHF continues to be used in large studies of patients with significant degrees of treatment resistance, such as the RECOVER study of VNS ( ) and, as previously mentioned, the ATRQ is also used in a number of past and current large RCTs of various treatments in patients with TRD ( ). The European project to build consensus around the definition of TRD previously described ( ) recommended use of the MSM on the basis of its predictive ability, though it should be noted that this European project is led by clinicians from the Maudsley and the other staging tools split opinion.

Utility and issues of the treatment-resistant depression categorization

An agreed consistent definition of TRD and its assessment/staging would be of great value in terms of research, allowing for comparison and meta-analyses of treatment trials, and for regulatory purposes ( ; ). Given the very significant impact of number of previous failed treatments on likelihood of response ( ) it is potentially inappropriate, for example, to compare different treatments in a network meta-analysis when they have been examined in populations defined in different ways. In addition, inherent in all definitions of TRD is the concept that the management of depression involves serial acute treatment trials ( ). In some health care settings, it is clear that many patients receive suboptimal treatment, for example being continued on antidepressants to which they are not responding for long periods, without consideration of alternative or additional pharmacological options ( ). In such situations, considering further serial options may be extremely valuable ( ). Staging tools such as TRSM can help clinicians to sequence such treatment. Similarly, a staging tool such as the MSM which has predictive utility ( ; ) may be of value in clinical prognostication.

However, there are a number of issues with the definition and concept of TRD. There is a lack of consensus regarding how adequate treatment trials are defined, how outcomes to trials are assessed and how to consider trials of treatments of different modalities. Conceptually, TRD implies that it is possible to define patients with depression as either having a treatment-responsive or treatment nonresponsive illness. This dichotomization of patients is not supported by the evidence base ( ). Rather, the situation appears to be that both groups of patients are extremely heterogeneous.

Several novel psychotherapeutic, pharmacological and neurostimulatory options have recently joined the treatment armamentarium, with more in development. This is a boon, but forces the question of when and for whom to choose these newer, often more expensive, and/or invasive treatments ( ; ; )? The situation varies between different health care settings, but the TRD threshold is not clearly the point at which management of depression radically changes. For example, in the United Kingdom, primary care physicians find it hard on occasion to access specialist mental health services ( ) meaning that patients are potentially treated with three or four antidepressants as monotherapy before augmentation and novel options are considered. To help guide clinicians as to when to consider novel treatments, coupled their staging of TRD with possible next step treatment options. For Stage I TRD they suggested rTMS, Esketamine, nitrous oxide, buprenorphine or ECT in some case. For Stage II TRD they suggested ECT, VNS or DBS. Given that Stage II TRD in their model is reached after a failure of just three failed trials, this does not sit well with health care for example in the United Kingdom. As a result, a group of UK experts from primary, secondary and tertiary care proposed the concept of “multi-therapy resistant major depressive disorder (MTR-MDD)” ( ). The aim of this group was to propose a threshold around which it might be appropriate to consider more expensive, less evidence supported and/or more invasive treatments. The idea was to find a balance between identifying as early as possible the point when such “nonstandard” treatments should be considered against ensuring that more standard, safe, well evidence-based treatments have been at least considered. The MTR-MDD criteria were set as “a failure to respond, achieve remission, maintain a response/remission, tolerate, refused or contra-indication of … .” at least: two trials of a structured evidence-supported psychotherapy; four adequate trials of antidepressants; two adequate trials of augmentation (specifically including lithium, quetiapine, or aripiprazole); and a course of eight ECT treatments. Such a threshold is way above that of conventional TRD but it may be more appropriate for considering treatments such as VNS ( ). However, other treatments, for example augmentation with modafinil, might be considered well before the full MTR-MDD criteria are met ( ). This illustrates that rather than TRD being a single important point for clinical decision making, the reality is that different decisions will be made at various points along the treatment pathway, often well beyond the TRD threshold.

Why is an alternative concept needed?

While clinicians generally view TRD as being a failure to respond to two antidepressants they generally do not use or apply this definition in practice ( ; ). Clinicians tend to use terms and phrases such as “recurrent or relapsing depression,” “complex depression,” “difficult-to-treat depression,” “multiple episodes of depression,” “chronic depression,” and “struggled with depression throughout their lives” ( ). Clinicians also identify challenges in defining TRD ( ). Firstly, the number of failed treatments is confounded by time given observations that increasing duration of depressive episode is associated with decreasing chances of remission ( ). Secondly, it is not clear what the target level of improvement is for each treatment trial. The issue of patients with “partially responsive depression (PRD)” has also been identified as an issue that needs consideration by the pharmaceutical regulators ( ). Thirdly, patients defined as having TRD often have comorbidities which can influence the chance of improvement with treatment, something not taken into account in most definitions of TRD.

Experts in the treatment of mood disorders also recognize problems with the TRD concept in a publication written by representatives of the World Federation of Societies for Biological Psychiatry, Australasian Society for Bipolar and Depressive Disorders, International Society for Affective Disorders, Collegium Internationale Neuro-Psychopharmacologium, and the Canadian Network for Mood and Anxiety Treatments ( ). A key issue is that there are multiple reasons for failure to respond to antidepressants including “biological resistance, diagnostic error, limitations of current therapies, psychosocial variables, a past history of exposure to childhood maltreatment or abuse, job satisfaction, personality disorders, co-morbid mental and physical disorders, substance use or non-adherence to treatment,” while only a subset of individuals do not respond to pharmacokinetic or pharmacodynamic issues related to a medication which may be resolved by substitution or addition of another ( ). The authors therefore recommend that in practice it is necessary to “deconstruct” an individual’s depression to understand all the various individual predisposing, precipitating and perpetuating factors related to their depression and why it is nonresponsive to treatment. Only by doing so may it be possible to construct an individualized care plan.

Encapsulating these ideas is the proposal from that instead of considering whether a patient has “TRD” or not, a better term and concept is of “difficult-to-treat depression (DTD),” drawing on models of care for chronic waxing and waning physical health conditions such as rheumatoid arthritis ( ) or the recovery model as applied, for example, to schizophrenia ( ). Semantically, the term “TRD” raises the question as to who or what is “resistant” and runs the risk of leading to blame and/or nihilism. The term “difficult to treat” is proposed since it recognizes that there is a problem that needs addressing and implies that to address this a collaboration between patient, family and health care team is required to address it ( ; ).

Conceptually, the DTD model takes a chronic, rather than acute, illness approach with two fundamental differences with the TRD model of care. Firstly, the TRD model assumes that a patient remains unwell following a series of acute treatment trials because of nonresponse to each in turn. However, DTD recognizes multiple routes for there to be a poor treatment outcome. In addition to a lack of response or remission, there can be a situation of an inability to sustain this following initial improvement through relapse or recurrence of depression. Another clinically challenging situation is when a series of treatment trials are ended not through nonresponse but intolerance. STAR*D data demonstrates that with each increasing treatment step, there is not only a reduction in the rate of remission, but also an increase in rates of relapse and intolerance (see Fig. 1.2 ). It is critical that all three routes to a poor outcome to treatment are considered, not just one. Secondly, an implicit assumption of the TRD model is that depression is an acute illness and that the goal of treatment is remission. This is not an unreasonable position given that lack of full remission of symptoms is associated with both ongoing psychosocial functional impairment ( ; ) and a very significantly increased risk of relapse ( ). Whenever possible, clinicians should strive for full remission. However, the reality is that this is not possible for all people suffering from depression as a result of the multitude of their predisposing, precipitating and perpetuating etiological factors ( ). The DTD model recognizes this reality and addresses some of the issues of applying a TRD model in clinical practice, such as a “try and try again approach” without consideration of the individual patient ( ) and the risk of missing diagnostic and therapeutic opportunities ( ).

Definition

A group of international experts has published a consensus statement on the definition, assessment and management of DTD ( ). This group proposed a definition of DTD, or “suspected DTD” as depression that continues to cause significant burden despite usual treatment efforts. This definition is intentionally broad in scope due to match the heterogeneity of clinical scenarios. There are two main elements to the definition: what constitutes “significant burden” and “usual treatment efforts.” In terms of burden, DTD recognizes that this may arise from (a) a failure in achieving response or remission acutely; (b) failure in sustaining response or remission; (c) a lack of functional restoration despite good symptomatic control; (d) unacceptable tolerability or nonadherence or rejection of the treatment option. Critically burden is defined by the patient. What constitutes “usual treatment efforts” depends on the professional treating the patient, the health care environment, local treatment guidelines etc. What is perceived as “difficult” in one setting may not be in another. It may appear inappropriate that a patient is deemed to have DTD in one setting but not another. However, the purpose of the term is to bring into play a different way of thinking about the illness management, including review of whether a referral to another setting, where it may not be perceived as so difficult, is in order. Also note that “treatment” relates to all and any forms of treatment for depression. It is acknowledged that the DTD “label” will encompass a very heterogeneous group of individuals ( ).

DTD model of care

A model of care for DTD is described in detail in . Essentially, if depression continues to cause significant burden despite standard first- and second-line treatment, then it may be considered “suspected DTD.” Full remission may be achieved through increasing the dose, or switching, an antidepressant, or augmentation with another medication, psychotherapy, neurostimulatory treatment, psychosocial intervention or self-help strategy. The depression has proved to no longer be “difficult to treat.” However, the further down the treatment algorithm the clinician and patient get without full remission and restoration of function, the more a DTD approach may be warranted.

The first step in managing suspected DTD is an assessment encompassing a review of the differential diagnosis, an identification and active treatment of all comorbidities (physical, psychiatric, substance misuse and iatrogenic), and a thorough assessment of the patient’s individual etiological factors and reasons for nonresponse. The overall notion is to address any tractable reasons for the depression being difficult to treat in the hope that this turns the illness into a more easily treated depression. For example identification that the underlying illness is bipolar disorder and hence a drug like quetiapine may be more appropriate than an antidepressant ( ), or identifying that a patient is perimenopausal management of which can make depression more treatable ( ). Similarly, if a poor outcome results from ongoing interpersonal conflicts, then a course of interpersonal psychotherapy may be in order ( ).

A model of care based on the TRD concept has a single goal: remission. While optimization of symptom control and achieving remission if possible remains critical, the DTD includes two additional treatment goals: reducing risk and impact of relapse; and optimizing psychosocial function and quality of life, with this last goal in particular being defined by the patient themselves ( ). The DTD also describes several principles of care including enhancing engagement and shared decision making within an integrated service pathway while also encouraging self-management techniques. Measurement-based care, rating not only depressive symptoms but also psychosocial functioning and quality of life is advocated ( ). While the patient’s treatment history is an important prognostic indicator, other factors are also critical in the understanding and management of DTD ( ). With the partial exception of the Dutch modification of the MSM, these are not included in any of the current TRD staging tools. Further research is required to explore the optimal way of objectively assessing degree of DTD. Critical the model recommends reassessments and consideration of treatment direction at frequent intervals at which the differential diagnosis and etiological factors are reviewed, rescreening for comorbidities is conducted and the use of all treatment modalities considered: pharmacological, neurostimulatory, psychotherapeutic, psychosocial, self-help etc. The aim is to leave no stone unturned in considering options to advance the three goals of treatment, seeking specialist input as necessary to achieve this.

Utility

Since some patients do not response acutely, and others do not sustain their responses despite multiple treatments ( ; ), how long should clinicians go on recommending further acute treatment trials? If all available options and combinations are tried this would take more than a lifetime. Clearly, we should neither undertreat nor overtreat. Patients with clear acute episodic depression and those earlier in the treatment pathway may be appropriately managed using a TRD approach of repeated acute treatment trials. However, this approach may be less appropriate for patients with more persistent depression, those who are not entirely symptom free between episodes, or any for whom finding a tolerable treatment is proving challenging. When is it parsimonious to reconsider the differential diagnosis and case formulation, and to contemplate a different approach—perhaps to address psychosocial stressors or to develop ways to cope with or adjust to some of their symptoms? The DTD model is proposed as a way of trying to address these questions. The model also has advantages in emphasizing the “shared care” between the patient, their carers and family, and health care professionals, and the need to take a broad holistic approach to management. Recognizing that some patients have depression that is difficult to treat is acknowledging a clinical reality but also emphasizing that things may be difficult but that this does not mean impossible.