Resolving the Personalisation Agenda in Psychological Therapy Through a Biomedical Approach

Abstract

Background/Objectives—The personalisation agenda—matching the correct psychological therapy to diverse and comorbid mental disorders—is an unanswered dilemma in the worldwide literature which has far reaching consequences for public health. This hypothesis article addresses the question: can a biomedical approach resolve the personalisation agenda? Methods—Narrative review drawing on clinical psychology, translational psychiatry, and biomedical science literature. Results—Diverse attempts to resolve the personalisation agenda have not yet succeeded. Randomised controlled trials are uniquely biased due to unwanted placebo effects; network meta-analysis cannot address adequately which psychological therapy to use; new methodologies have not yet produced data; and neuroscientific analysis cannot yet explain how trauma-based therapies work. However, a biomedical model which divides psychological therapy into low, medium and high intensity interventions can resolve the personalisation agenda. Conclusions—Combining low intensity (placebo), with medium intensity (cognitive behavioural techniques) and high intensity interventions (trauma-based therapies) are theoretically synergistic if combined with psychosocial treatments/exercise, and used in sequence in the correct order. A biomedical model based on recent advances in placebo studies and neuroplasticity can resolve the personalisation agenda, and improve outcomes for mental disorder.

1. Introduction

Psychological therapy is established globally as a generic evidence-based treatment for mental disorder, alongside psychotropic medications. In addition, psychosocial treatments based around exercise are now recognised as core evidence-based treatments for some mental disorders [1,2]. There is emerging consensus that these three treatment modalities are complementary not competing, i.e., that a biopsychosocial approach across the spectrum of mental disorder requires an integrated biopsychosocial management response [3,4]. However, with respect to psychological management, it has proved difficult to match diverse psychological therapies to specific mental disorders, and this dilemma has historically been described in the psychology literature as the “dodo bird effect” [5] or more prosaically as the “personalisation agenda”. Personalisation in psychological therapy is defined as optimising treatment outcomes by tailoring the therapy to the specific needs of the individual [6].

The difficulties in resolving the personalisation agenda through evidence-based analysis have generated a substantial literature, but there remain many challenges in targeting the right therapy to the right patient: failure to do so is wasteful of a limited resource and sometimes dangerous to the individual patient. There are four main reasons why the personalisation agenda has not been resolved: these are summarised in Figure 1 and analysed in detail in Section 3 below.

This hypothesis paper examines the personalisation agenda from a fresh perspective. It re-examines existing literature to hypothesise that a biomedical approach, based on accurate diagnosis and categorisation of mental disorders, can address the personalisation agenda and potentially improve outcomes of psychological therapy. The hypothesis is developed through the methodology of narrative review.

The article is in four sections. Firstly, it quantifies the nature of the personalisation problem and its consequences. Secondly, it examines the limitations of the randomised controlled trial in mental disorders. Thirdly, it analyses alternative attempts to resolve the personalisation agenda and their limitations. The fourth section draws on knowledge derived from the first three sections to posit a new and alternative biomedical solution to address the personalisation agenda.

This article will be of interest to a broad range of biomedical scientists because of its focus on research methodology, the interaction of placebo responses with applied statistics, and how theoretical models could be translated into real-world practice.

2. Background Information and Definitions

Precise definitions in mental disorders are important, and in this article the generic term “mental disorder” is used. This is defined in the Diagnostic and Statistical Manual (DSM) as “a syndrome characterised by clinically significant disturbance in an individual’s cognition, emotion regulation, or behaviour that reflects a dysfunction in the psychological, biological or development process underlying mental functioning” [7]. The terms “mental health problems” and mental health difficulties” are avoided as they are not defined and lack scientific credibility.

A biomedical approach subdivides mental disorder into categories, agreed by international consensus, so that evidence-based treatment can be targeted at each component based on knowledge of underlying pathophysiology [8]. Diagnoses should be made by qualified practitioners after detailed assessment based on knowledge of medical and psychological comorbidities, and how to do this is explicit in clinical practice guidelines for different mental disorders, for example, NICE Guidelines on depression CG222 [9], or the CANMAT guidelines on schizophrenia [10].

A biopsychosocial model has traditionally been used as a diagnostic tool to identify biomedical, psychological and social factors that may contribute to the pathogenesis of mental disorder. In some mental disorders, particularly depression, there is evidence that a biopsychosocial approach to management, whereby health workers cooperate in a united model of action, can improve outcomes [4].

The terminology used for placebos is as adopted by international consensus [11].

Psychological trauma is defined in the Diagnostic and Statistical Manual (DSM-V) [12] p. 250. The DSM was the first classificatory system to identify post-traumatic stress disorder as a diagnosis in its own right, rather than trauma being an aetiological risk factor or trigger for other mental disorders.

Neuroplasticity is defined as the ability of the central nervous system to change its activity in response to intrinsic or extrinsic stimuli by reorganising its structure, functions or connections [13], and the translational psychiatry literature subdivides neuroplasticity into neurodevelopmental neuroplasticity and adult neuroplasticity [14]. At a cellular level, neuroplasticity is subdivided into ionotropic neuroplasticity, which is rapid in action associated with flow of ions across neuronal membranes and metabotropic neuroplasticity, which relies on stimulation of neurotropic growth factors within the glutamate/GABA/glutamine cycle such as BDNF (brain-derived neurotropic factor) [15].

2.1. The Consequences of Failing to Resolve the Personalisation Agenda

The difficulties in resolving the personalisation agenda have resulted in a crisis in psychological therapies [16]. By themselves, psychological therapies reduce symptoms, improve quality of life and reduce risk of relapse for many common mental disorders such as anxiety and depression [17]. However, large-scale reductions in population prevalence of mental disorder are less clearly demonstrated because of limitations in access, variation in uptake and high dropout rates [18]. Further, there is little evidence that psychological therapies have significantly impacted the 750,000 global deaths by suicide reported annually [19].

There is evidence from the World Health Organisation’s MH Gap programme [20] that psychological therapy of all genres is unavailable to almost a billion people [21], particularly in the global south, where comorbidity surveys produce a conservative estimate of prevalence of mental disorder of 10–20% [22,23]. This highlights the extensive evidence that the majority of all mental disorders are associated with and may even be caused by relative poverty and health inequalities [24,25]. However, where mental disorder is closely associated with social disenfranchisement and relative poverty, there is little evidence that using psychosocial interventions produces better outcomes than other genres of psychological therapy. Psychosocial interventions, which are usually classified under the broad rubric of “psychological interventions”, are not clearly defined but are listed in Figure 2. Some mental disorders related to relative poverty may therefore have a political rather than a clinical solution [24].

Addressing the crisis in psychological therapies therefore requires effective psychological therapies targeted at specific conditions and available to different communities in a culturally acceptable way [3,21]. It requires psychological therapies that improve outcomes, are relatively easy to administer and accessible to those who need them most, including refugees and asylum seekers [3]. Further, national epidemiological surveys [23,26] provide strong evidence that mental disorder is commonly accompanied by comorbidities—for example, depression and substance misuse—so psychological therapies usually need to address comorbidities too in order to improve long-term outcomes. The mismatch between demand and supply will continue to challenge health commissioners, but resolving the personalisation agenda would go some way to ameliorating the crisis through more efficient and targeted use of resources, and greater clarity of roles amongst health professionals.

There are three main consequences of the difficulties in resolving the personalisation agenda.

Firstly, treatment is not targeted at patients who require it most. In the United Kingdom for example, there has been a huge demand for counselling for minor mental disorders from the Improved Access to Psychological Therapies (IAPT) service [27], at the expense of psychological therapies being available to those with severe and enduring mental disorders [28]. In lower- and middle-income countries, psychological therapies are widely unavailable, so the voluntary sector, traditional healers, shamans, traditional Chinese medicine, and complementary therapists have stepped into the vacuum to provide varying degrees of unevaluated placebo-based assistance.

Secondly, skilled and experienced psychological therapists adopt a ‘one-size-fits all’ approach for all “mental health problems”. They are trained in cognitive behavioural techniques and may concurrently help their patients by adopting a pluralistic approach. A pluralistic approach utilises common factors [29] combined with cognitive behavioural techniques and a trauma-based focus when required. Such combinations may be used in a phased way, or even within the same therapy session. This approach usually helps the individual in front of the therapist, but is insufficient to resolve the personalisation agenda, nor help the millions of mentally disordered patients worldwide who cannot access skilled and experienced therapy.

Thirdly, there has been a growth of “patient choice”, which has not been subject to an evidence-based approach or health economic analysis, and therefore varies widely in the degree to which it is supported in clinical practice guidelines [30]. Whilst patient choice promotes patient empowerment, it is not clear how a patient makes an informed decision regarding which therapy to choose when the therapist cannot themselves decide based on available evidence. It is not clear what happens if the patient chooses a therapy that the therapist is not trained to deliver, nor if a patient chooses a therapy that is very expensive, such as psychodynamic psychotherapy.

2.2. The Limitations of the Randomised Controlled Trial in Resolving the Personalisation Agenda

To achieve parity of esteem, regulatory bodies such as the U.S. Food and Drug Administration (FDA) have sought to apply the same rigorous standards of evidence-based medicine to the treatment of mental disorders as to physical health conditions [31]. However, placebo responses tend to be higher in trials of mental disorders, which can reduce observed effect sizes and complicate interpretation of treatment efficacy [32]. Traditional randomised controlled trials (RCTs) used in psychological therapy outcome research often compare an active psychological intervention against control conditions that may themselves generate therapeutic effects—such as supportive contact or expectancy—blurring the distinction between “active” and “placebo” treatments. In fact, placebo responses are now recognised as clinically relevant and particularly pronounced in mental disorders, where the brain appears to be especially sensitive to contextual and relational factors [33,34].

Common control conditions such as waiting list controls introduce further complexity, and delayed access to the intervention raises ethical and methodological concerns [34]. Research participants may experience inconsistent placebo effects while waiting and then during the trial intervention.

Genres of psychological therapy vary widely, but a consistent finding over decades is that “common factors” involving the empathy and skill of the therapist are as important in determining outcome as the type of therapy involved [29]. The therapeutic alliance, motivation of the patient, and the social environment are factors that cannot be ignored This suggests that placebo influences are important to outcomes, although the limited nature of RCTs in this field has not been able to tease out the contribution of placebo to outcome in psychological therapies, a significant factor in the failure to resolve the personalisation agenda.

Cognitive behavioural therapy (CBT) and its variants are the most researched and most commonly used psychological therapies [35] in both “severe and enduring” [28] and less severe forms of mental disorder [36]. The mode of action of CBT was described 20 years ago [37], and since then a number of variants of CBT have been developed that provide a better match for specific mental disorders—for example, dialectic behavioural therapy for borderline personality disorder [38]. However, in relating these to the personalisation agenda, RCT methodology has not been able to establish why CBT techniques work less well in some mental disorders, e.g., mental disorders related to trauma, or in some clinical settings, e.g., forensic settings. The alternative genres of psychological therapy used vary widely from eye movement desensitisation and reprocessing (EMDR) [39]—which takes minutes to deliver—through to classical psychodynamic therapy based on the ideas of Jung and Freud—which may take years to deliver.

In summary, historically the RCT when applied to psychological therapy has been confounded by bias and variable placebo influences such that it has only addressed to a limited extent the multiple variables inherent in personalisation of therapies.

2.3. The Limitations of Alternative Research Approaches in Resolving the Personalisation Agenda

Due to the limitations of the RCT, there have been alternative research attempts to address the personalisation agenda, with variable success. The research effort, led by clinical and academic psychology, can be broadly divided into four categories: new/prospective randomised controlled trial (RCT) research; re-analysis of existing data through network meta-analysis (NMA); novel research methodologies including artificial intelligence; and neuroscientific analysis of mental disorders based on a new understanding of neuroplasticity in mental disorders. These four domains of research and their limitations are now examined.

(A).

New RCT research

Having identified the crisis in psychological therapies more than a decade ago [16,18], a large programme of new research, based on RCTs, has been developed. Ten areas of research have been identified [3], which are summarised in Figure 3.

Since this research agenda was set out in 2018 [3], there has been little RCT research published to address the questions set out in Figure 3. RCT research in this area is difficult to conduct, both in terms of recruitment and research design, and definitive answers to the 10 research questions need to be addressed using techniques which go beyond the traditional RCT, because of its particular limitations in mental health research [40].

(B).

Re-analysis of existing data through network meta-analysis

NMA is a powerful statistical technique to compare treatment outcomes in medical and mental disorders [41] where there is more than one treatment modality. It amalgamates comparable RCTs through systematic review, makes direct and indirect comparisons of data, and where Bayesian statistics can be applied, has more statistical power to differentiate between several treatments than pairwise meta-analysis [41]. In psychological therapy research, NMA has been used since 2012 [42] to create hierarchies of different genres of therapy, including comparing CBT with trauma-based therapies, and has been thought to be able to address current gaps in research including the personalisation agenda by re-analysis of existing RCTs [43].

However, NMA has not been able to definitively distinguish between different genres of psychological therapy and offers little advantage over pairwise meta-analysis. The results are dependent on which RCTs are included in comparing therapies head-to-head, and this is subject to publication bias and missing outcome data [41].

If NMA is of limited utility in comparing psychological therapies internally with each other, it has advanced the literature in comparing psychological therapies with other treatment modalities: in depression NMA shows that combinations of interventions (psychological therapy + exercise + psychotropic medication) produce lasting benefits [1,44,45].

In conclusion, NMA provides strong evidence for biopsychosocial management of mental disorder, requiring clinicians to work more closely together with a shared philosophy of care, but does not resolve the questions arising from the personalisation agenda.

(C).

Novel research methodologies including artificial intelligence

As RCTs offer a low level of evidence into long-term effectiveness due to individual variability and contextual influences, research methodology in clinical psychology has evolved in response. Innovations include pragmatic trials, adaptive designs, stepped-care models, and hybrid trials that assess both efficacy and implementation in real-world settings [46]. Interested readers are directed to comprehensive recommendations for future research design and practice into psychological therapies [40]. Additionally, mixed-methods approaches, including qualitative research, are increasingly used to capture patient experience and therapeutic process [47].

Research into personalisation for psychological therapies is evolving. For example, recent developments in algorithmic personalisation through machine learning [48] show some promise in predicting which patient will benefit from which therapy [49]. However, this approach is in its infancy and has provided only limited evidence so far in resolving the personalisation agenda [49].

In summary, advances in methodology to better elucidate psychotherapy outcomes are evolving, but have not yet been translated into data that resolves the personalisation agenda.

(D).

Neuroscientific analysis based on neuro-radiological advances

The pathophysiology of mental disorder is now understood as a dysfunction in circuits linking the frontal cortex, the limbic system, and the hypothalamic–pituitary–adrenal (HPA) axis, areas of the brain that are particularly susceptible to stress and neuroplastic change [13].

The last three decades have seen rapid advances in understanding of mental disorders based on the “window to the brain” that neuro-radiological techniques have provided. Functional magnetic resonance imaging can aid in identifying mental disorders in vivo, synthesising data from resting state and task-directed scans. Mental disorders that arise from abnormalities in neurodevelopmental neuroplasticity (e.g., autism spectrum disorder (ASD), attention deficit hyperactivity disorder (ADHD), and early onset schizophrenia) can be distinguished by characteristic neuro-radiological appearance from mental disorders arising in later life (such as adult depression and late onset schizophrenia). This data has directed attention away from biochemical hypotheses of mental disorder [50], whose relevance has been challenged [51], towards both adult and neurodevelopmental neuroplastic abnormalities of the brain connectome.

Furthermore, neuroscientific analysis has illuminated how some psychological interventions work in correcting fronto-limbic circuits that are demonstrably dysfunctional in mental disorder [52]. In particular, the mode of action of placebo effects and CBT is now delineated in terms of stimulation of neuroplasticity [37,53], as follows:

Placebo—Recent biomedical advances in placebo research suggest that placebo works in mental disorders through a combination of empathy on the part of the therapist [54,55], and conditioning, expectancy and cognitive appraisal on the part of the recipient [56,57]. The benefits of placebo effects persist [58], suggesting an underlying mechanism of metabotropic neuroplasticity. The neuro-radiological correlates of placebo in depression and other mental disorders have been delineated by the EMBARC series of studies [59] and related research.

CBT—works by helping individuals reduce subjective stress by connecting thoughts and feelings, eliminating avoidance behaviours and influencing reward networks, achieved by challenging faulty schema and basic assumptions [60]. Neuro-radiological correlates show CBT and related techniques are associated with reduced activation of the amygdalahippocampal cortical regions involved in negative emotions, coupled with increased activation of frontal regions involved in cognitive control of negative emotions [37]. CBT therefore stimulates neuroplasticity in similar fronto-limbic areas that are influenced by placebo effects. The overlap between placebo and CBT is striking in that they exert their beneficial effects in mental disorders via similar fronto-limbic pathways. This raises the possibility that CBT is a placebo effect ‘on steroids’, with the placebo message amplified and magnified over a course of CBT, which uses both ionotropic and metabotropic mechanisms to stimulate beneficial neuroplasticity.

Trauma-based therapies—Despite these speculative advances in understanding the mode of action of placebo and CBT, the main limitation of neuroscientific analysis is that it is not yet known how other genres of psychological therapy work, including trauma-based therapies [3]. Until this is established, neuroscientific analysis by itself will not be able to resolve questions arising from the personalisation agenda.

2.4. Applying a Biomedical Model to Resolve the Personalisation Agenda

The biomedical model is drawn theoretically from advances in network meta-analysis (NMA), knowledge of placebos, and the role of neuroplasticity in the pathophysiology of mental disorder.

2.4.1. Basic Principles of the Biomedical Model

There are four basic principles of a biomedical model:

(A).

Treatment of mental disorder can be based on accurate, standardised diagnosis of both core mental disorders and comorbidities.

(B).

Comorbid mental disorders may require contemporaneous or sequential treatment with three levels of intensity psychological treatment modalities: low, medium and high intensity.

(C).

Placebo has a core role as a low-intensity treatment in all psychological therapies [29,33,58] and shares a similar neuroplastic mechanism of action with other psychological therapies [61].

(D).

Most mental disorders respond to placebo plus cognitive behavioural techniques [28]. However, mental disorders associated with trauma only respond partially to this approach and require high-intensity trauma-based therapy for good long-term outcome. The mechanism of action of trauma-based therapies is via neuroplasticity [61].

2.4.2. Applying a Biomedical Model in Clinical Practice

Three levels of psychological interventions can be commissioned to treat mental disorders cross-culturally irrespective of age. Low-intensity treatment comprises the placebo effect, which can be used by any health professional and works through expectation, cognitive appraisal, and stimulation of neuroplasticity in fronto-limbic areas [56]. Medium intensity treatments (CBT and variants) share this mechanism of action and reinforce it: they work in the majority of mental disorders but not in mental disorders secondary to trauma. High-intensity treatments are trauma-based and should only be administered by skilled therapists.

The biomedical model can be applied to patient groups where outcomes of therapy are currently poor. In forensic settings, virtually all patients are traumatised, so the biomedical model predicts that CBT approaches in isolation are not going to have limited benefits until the trauma is addressed. Similarly, patients with medically unexplained symptoms [62] and chronic pai-n [63] frequently have unrecognised trauma which they are well defended against disclosing: high intensity treatments are required first, so that they then may benefit from medium intensity treatments. Neurodevelopmental disorders such as ASD, ADHD and personality disorder do not require expensive high-intensity treatments, as CBT approaches are usually effective.

2.4.3. Why a Biomedical Model Can Potentially Resolve the Personalisation Agenda

The three levels of intensity are complementary not competing, and may be used simultaneously or in sequence in the same patient. There is overlap between low intensity and medium intensity treatments, but high intensity treatments are reserved for patients whose mental disorder is related to trauma: the biomedical model posits that these patients do not respond to low or medium intensity interventions until the trauma is resolved. Applying a biomedical model has the potential to transform the management of mental disorders in three ways.

Firstly, mental disorders can be treated generically with a combination of exercise, targeted psychological therapy, and psychotropic medication. Specifically, these modalities all work by correcting aberrant circuitry [64,65,66] in fronto-limbic areas through neuroplasticity, and may be synergistic when used in combination [67].

Secondly, the presence of trauma becomes a key discriminator in deciding personalised care: trauma needs to be addressed first through high-intensity treatment before benefits can be derived from medium and low-intensity treatments.

Thirdly, resolving the personalisation agenda through a biomedical model enables shared decision making [68] to replace patient choice in engaging patients with therapy.

3. Discussion

This is the first article to propose a biomedical solution to the personalisation agenda. The article proposes a novel theoretical framework—an “ideal type” [69] or structured model- to resolve the personalisation agenda. If further research confirms the effectiveness of this model, there will be a need to develop practical recommendations for its implementation.

A biomedical model can resolve the personalisation agenda because it is predicated on a diagnosis (or more commonly diagnoses), which then directs the therapist to the optimal genres of psychotherapy based on underlying pathophysiology.

Applying a biomedical model to psychological therapies potentially represents a clash of cultures [70]. The “biomedical model” in psychology and sociology literature has for many years been portrayed as reductionist [71], with diagnostic labelling creating stigma [72]. The biomedical model has been conflated with medical hegemony in decision making and “over-medicalisation” [73], which includes prescribing psychotropic medication when the better approach is psychosocial or psychotherapeutic management. However, this article concludes that applying a biomedical model, as defined above, is the only way to resolve the personalisation agenda within the foreseeable future, and to address the research questions posed by the Lancet Psychiatry Commission [3] shown in Figure 3. Furthermore, a biomedical approach to personalisation does not preclude a biopsychosocial approach to management of mental disorder where evidence suggests a combined tripartite approach improves outcome. For example, outcomes in depression are improved if psychological therapy is combined with exercise [74] and, where necessary, psychotropic medication [3,44].

3.1. Training Requirements

To implement a biomedical model for psychological therapies, all health workers should be trained to deliver low intensity interventions as part of communication skills training. All mental health workers should be trained in common factors and how to deliver medium intensity interventions. Treating trauma through high intensity interventions requires higher training and supervision by accredited psychological therapists, although EMDR is a shortcut that can help some patients where skilled high intensity psychological therapy is unavailable.

Psychological therapists who have been using the terms “mental health problems” and “patient choice” may need training in shared decision making.

3.2. Further Research

There are three priorities for further research.

Firstly, when a new model or intervention is introduced, the normal response is “Do a large-scale RCT to evaluate it”. However, in this case the RCT is part of the problem not the solution, as discussed in Section 2 above. Therefore, mixed methods research [75] and recently developed research designs [40] specific to psychotherapies will be required to evaluate long-term outcomes of a biomedical model for resolving the personalisation agenda.

Secondly, the biomedical model is based on the necessity for high-intensity interventions for mental disorders associated with trauma, rather than the severity of the mental disorder dictating the need for high-intensity interventions per se. Most severe and enduring mental disorders are better treated with medium-intensity interventions [28]. Further research is required into the neuropsychology of trauma and how trauma affects expectation.

A third research priority is to evaluate the impact of applying a biomedical approach as an early intervention in neurodevelopmental mental disorders.

3.3. Limitations

There are several limitations to this article:

• This article is written from a Western perspective on how mental disorder should be classified and treated, although it does seek to accommodate the World Health Organisation perspective [76] on cross-cultural issues. It makes only passing reference to the role that artificial intelligence and recent advances in psychotherapy research design may play in future in personalising therapy [49].
• Dividing psychological therapies into three levels of intensity may be too simplistic when considering the complexity of the human condition. The article has posited a heuristic model that provides a pragmatic problem-solving approach to an enduring dilemma, but a biomedical approach is only one of the possible approaches [40].
• Much of the data in support of applying a biomedical model for psychological therapy is derived from research into depression. Depression is the most common mental disorder supported by a larger research literature than other mental disorders. Around a third of the references below refer specifically to psychological management of depression, rather than mental disorder more widely. It has been assumed that data can be extrapolated from depression to all mental disorders, based on the generic contribution of dysfunctional and corrective neuroplasticity in mental disorders. This requires analysis by further research using the correct methodologies.
• The biomedical model is based on the hypothesis that trauma-based therapies share their mode of action with other genres of psychological therapy and work via neuroplasticity. Whilst there is some evidence to support this hypothesis [61], the hypothesis has not been confirmed [3] by empirical research.
• There has been no consideration of the ethics of using placebo as a low-intensity treatment for mental disorders. The general public associates placebo with sugar pills, and whilst no subterfuge can be involved in the use of placebo [77], it is not known if telling patients in advance that placebo is being used will dilute its beneficial effects.

4. Conclusions

1.

The personalisation agenda can potentially be resolved by a biomedical approach, which divides psychological therapies into low, medium and high intensity interventions.

2.

Most mental disorders can be treated by generic mental health workers with low- and medium-intensity interventions, which overlap and are based on expectation and cognitive appraisal. High intensity treatments are reserved for mental disorders caused by trauma.

3.

Psychological interventions are complementary not competing, patients with comorbidities may require all three approaches for optimal management and outcome.

4.

The biomedical model posits that it is important to identify psychological trauma when diagnosing mental disorders, as the presence of trauma changes management.

5.

Long-term outcomes of psychological therapy should be evaluated by mixed methods research, given the unique limitations of the RCT in this field.

6.

For the Mental Health Gap programme, where skilled psychological therapy is not available, the combination of low intensity psychological intervention plus exercise may substitute in some mental disorders: EMDR is the most affordable high intensity intervention for trauma.

7.

Resolving the personalisation agenda will not by itself resolve the crisis in psychological therapies: NMA provides evidence that biopsychosocial networks of neuroplasticity-inducing treatments are required to address the relative poverty and comorbidities that drive much mental disorder.