The focus, advancement, and direction of any scientific discipline is critically and essentially dependent upon the basic conceptualization that it holds of its subject matter [1,2,3], and upon the subsequent taxonomy or listing which it develops of that subject matter [4, 5]. Chemistry has its periodic table. Zoology and botany list and categorize species. Astronomy has developed its star charts, galaxy taxonomy, and so forth.
In order to avoid unwarranted assumptions, the objects of systematic observation and experiment, theory development, assessment, and intervention within the science and practice of clinical psychology – it being a remedial discipline – may most simply be called clinical psychological problems (CPPs). CPPs have, over time, variously been conceived as instances of demonic possession , moral failings , deeply-rooted psychodynamic pathologies, observable patterns of maladaptive behaviour, or internal states fundamentally grounded in biology . Each alternative conceptualisation has entailed its own ad hoc or systematised taxonomy of psychological-level problems.
Since the publication of the third edition of the American Psychiatric Association’s Diagnostic and statistical manual of mental disorders (DSM)  – the first edition to be based heavily on a medical model of ‘mental disorders’ (closely paralleled by the WHO’s International Classification of Diseases) – the conception of CPPs as biologically-based internal states has come to dominate, and the DSM has become clinical psychology’s de facto problem taxonomy. This has led to and cemented the assumption that CPPs are solely and entirely mental disorders. This assumption has resulted in many widely recognised and irreconcilable problems for the discipline and profession of clinical psychology.
(The soon to be implemented latest iteration of WHO’s ICD classification – ICD-11 – acknowledges these problems more than DSM-5 does, and its response will be described shortly.)
The publication of the latest edition of this taxonomy, DSM-5 , has solved none of the problems alluded to [11,12,13,14,15,16], prompting the British Psychological Society’s Division of Clinical Psychology to release a consensus statement on psychiatric diagnosis – Position Statement on the Classification of Behaviour and Experience in Relation to Functional Psychiatric Diagnoses: Time for a Paradigm Shift  – which summarized DSM’s limitations and deficiencies for clinical psychology. It concluded that “the current classification system as outlined in DSM…has significant conceptual and empirical limitations, consequently there is a need for a paradigm shift in relation to the experiences that these diagnoses refer to, towards a conceptual system which is no longer based on a ‘disease’ model” (p.1).
However, beyond suggesting that “such an approach would need to be multifactorial, to contextualise distress and behaviour, and to acknowledge the complexity of the interactions involved” (p.3), and that it should be “in keeping with the core principles of formulation in clinical psychology” , the Division did not propose a precise focus or the content of such an alternative conceptual taxonomy that would satisfy this paradigm shift. It is insufficient to point up the limitations of a conceptual model of CPPs if a superior one cannot be proffered, and “at present there is no consensus on what an alternative, universal theory of what maintains and exacerbates psychological distress might look like” .
If we are to develop such a radical reconceptualization of CPPs, to foster “a true ‘Kuhnian’ revolution” ( p.1935) in clinical psychology, and to develop a subsequent taxonomy of such, then the nature and essence of these new CPPs may be discernible in two ways: (1) The precise nature of the recognised inadequacies and incompatibilities of the mental disorder model for clinical psychology can point us toward a more useful and relevant conceptualization of CPPs; And (2) what clinical psychologists actually address in their research and practice may be drawn and distilled to extract the true essence of CPPs.
So firstly, what can the problems and incompatibilities for clinical psychology of psychiatry’s DSM and its ‘mental disorders’ model teach us about the essence of CPPs?
Psychiatry is more biological-level; clinical psychology is more psychological-level
Psychiatry, as a branch of medicine, has a much greater focus on biological-level assessment, explanation, and intervention than has clinical psychology. Hence, its DSM lists.
‘mental disorders’, ‘diagnosed’ by the identification of a ‘syndrome’ of ‘symptoms’, which are assumed to be manifestations of a ‘pathological condition’ [19, 20].
This ‘nosology of diseases’  based not on empirical evidence, but on clinical authority and historical tradition , has been problematic for psychiatry itself [13, 23,24,25,26], let alone for clinical psychology. It has been plagued by such major problems as excessive rates of comorbidity [27, 28], which may be an indicator of arbitrary boundaries between its disorders [29, 30], by the broad heterogeneity within its diagnosed groups [21, 31,32,33], and by the fact that none of the putative underlying disease processes have been uncovered in the 35 years of research since DSM-III was published [25, 34,35,36,37]. The search for biological etiology has greatly disappointed [38, 39], suggesting that psychiatric diagnosis has oversimplified psychopathology .
DSM and the ICD, meantime, have been poor guides to even psychopharmacological treatment selection [41, 42], let alone to psychological therapy selection. Psychiatric drugs are frequently prescribed “outside their license”, as when chlorpromazine is administered for anxiety or insomnia, thioridazine (another antipsychotic) for alcohol withdrawal, and benzodiazepines for “pretty much everything…The classification of mental health conditions gives us a false sense of order…It has little or no relevance to psychotropic drug action” ( p.225).
As part of the development of ICD-11, First et al.  surveyed 1764 mental health professionals, mainly psychiatrists, and found that the majority used ICD-10 or DSM-5 for administrative or billing purposes only. They rated such taxonomies as least useful for treatment selection and determining prognosis.
Psychiatry’s own response to these acknowledged problems has been to redouble its conceptual biological insistence. For example, the National Institutes of Mental Health are developing a Research Domain Criteria (RDoC) framework  which, even more than DSM, conceptualises mental illnesses as brain disorders (latent disease constructs), but which seeks to identify them through objective behavioural tests and neurobiological measures such as genetic tests and neuroimaging, rather than through topographical symptom checklists. This response to DSM’s failures has been described as a shift from the biopsychosocial model of mental disorders to a “bio-bio-bio model” . It is much more a framework for biologically-oriented research  than a clinical replacement for the ICD or DSM . While it tries to be more etiological and dimensional than those systems, its clinical usefulness lies well into the future .
The RDoC approach is therefore not at all a solution to clinical psychology’s problems with the conception of CPPs as diagnosable mental disorders, which are less to do with the technical limitations of diagnosis and more to do with a conceptual mismatch. The proposers and developers of the RDoC project have attempted to circumvent the problem that heterogeneous symptom profiles among diagnosed groups are likely to encompass a large number of biologically distinct entities . But clinical psychologists’ concerns are that these groups are likely to encompass a large number of psychologically distinct entities.
So even were the RDoC project to improve diagnostic reliability, validity, and clinical utility for psychiatrists, it would still offer no greater attraction to clinical psychology. Clinical psychologists as a group are not as biologically focused or trained, do not prescribe medications or administer ECT, and in fact in practice rarely and reluctantly diagnose [51,52,53]. Instead, they construct case formulations at a psychological level .
The idea that CPPs can and should be reducible to presumed underlying neurobiological conditions which are somehow more basic, real, or ‘scientific’ than psychological-level formulations is not helpful , not logical , and, for almost all CPPs, is theoretically premature . The psychological and the biological are different levels of analysis, assessment, and intervention , and any alignment of phenomena at these two levels is, by definition, correlational, not causal . It is no more likely that all CPPs will be reduced in the future to neurobiological conditions than that the geological study of earthquakes will be reduced to molecular theory ( p. 508).
While it is possible and desirable to theoretically unify the social sciences and biology, the notion of abandoning the principles, theories, vocabulary, and laws of the social sciences in favour of lower-level terms is a “preposterous” proposition , which would result in such theses as “A Comparison of Keats and Shelley from the Molecular Point of View” or “The Role of Oxygen Atoms in Supply-Side Economics”. Such “greedy reductionism” can arise when “in their zeal to explain too much too fast, scientists and philosophers often underestimate the complexities, trying to skip whole layers or levels of theory” ( p.82). “Mental disorders may be studied at different levels of analysis (e.g. molecular genetics, neurochemistry, cognitive neuroscience, personality, environment), and no level is inherently superior or fundamental to any other” ( p.856).
Clinical psychologists, when they operate within an adjacent level of analysis – in this case a psychiatric one – will lose a large, perhaps critical, amount of psychologically-relevant information. “Psychiatrists using the DSM diagnosis ‘major depression’ tend to mingle bereaved patients with both those afflicted by classic melancholia and those demoralized by circumstances” ( p.1854). So when clinical psychologists allow themselves to be diverted from the study and psychological-level formulation of CPPs to research into the treatment of DSM-diagnosed mental disorders, this means that a 19 year old survivor of 14 years of sexual abuse within her dysfunctional family, who is now sad and amotivational every day, will be regarded as experiencing precisely the same CPP as a 73 year old recently bereaved widower who is also sad and amotivational every day, because these people share some ‘symptoms’ – some topographical similarities. They will also find themselves in the same experimental or control group in a clinical trial of a particular cognitive therapy or antidepressant medication, and conclusions about efficacy will then be extended to other people with even more diverse CPPs, because they allegedly have the same mental disorder.
It is highly likely that some CPPs currently regarded or labelled as mental disorders are most usefully assessed, diagnosed, and treated within a medical model, but that some do not conform well to this level of analysis, and will respond better when assessed and addressed at a psychological level [59,60,61]. “Psychiatric diagnoses differ in the sorts of categories that best capture them” ( p.204). Some may be more categorical than others . There is some evidence, for example, that anorexia nervosa may be much less culture-bound and more heritable than bulimia nervosa , and so may be less socially constructed, more categorical, and a different ‘kind’ of thing. Schizophrenia and a simple reactive dog phobia are also likely to represent different classes of CPP in this light. The former more comfortably rests within a taxonomy of ‘mental disorders’ such as the DSM. A reactive dog phobia, on the other hand, may be more conceptually concordant with clinical psychology’s own parallel purely psychological-level taxonomy of CPPs.
It will be a long time – if ever – before a complicated bereavement is fully explained by reference to a particular neural bundle, or treated solely with a localized electrical zap or a ‘complicated bereavement pill’. Clinical psychology and biological psychiatry are different disciplines, operating at adjacent but different levels of analysis, and neither should subsume the other.
Our new conception of CPPs, and its subsequent taxonomy, will therefore centre on psychological-level states and processes – involving cognitions, emotions, behaviours, and situations or stimuli – and not on biological-level ones.
Mental disorders are social constructions; they have no essence
Another major problem with equating CPPs with mental disorders is that this subsumption represents relegation to a less developed, less theoretically robust, less therapeutically relevant level of analysis. This is an inevitable consequence of the fact that, whereas our theoretical knowledge of the processes, functions, and mechanisms underlying CPPs has grown greatly, DSM’s listing of mental disorders began and has remained stolidly atheoretical [8, 32, 64, 65]. DSM has made no claims about underlying mechanisms, functional processes, pathophysiology, etiology, and hence treatment implications of its mental disorders, and is therefore a “weak medical model” .
This deliberate policy was originally so as to accommodate a large number of theoretical orientations from a range of professions or disciplines [32, 35], but also more recently because, as previously described, the medical model has largely failed to further our understanding of the heterogeneous assortment of disorders the DSM lists . The sluggish pace of discovery in psychiatry has been attributed, in part, to the limited validity and the arbitrariness of traditional diagnoses .
So, whereas a clinical psychologist will see a CPP involving problematic social anxiety, for example, as a psychological-level persisting negative process that requires case formulation and specific subsequent psychological-level intervention, according to DSM a Social Anxiety Disorder is a state or condition identified (but not explained) by its symptoms. How do we know that Bill has a Social Anxiety Disorder? He shows enough symptoms. What caused these symptoms? His Social Anxiety Disorder. There is no evidence that the mental disorder or mental illness called “Social Anxiety Disorder (Social Phobia)” (DSM300.23) actually exists. It has no ‘essence’. There are no reliable or validated biological markers or measures outside clinical psychological judgement that can detect this illness. It is defined by its effects, which are attributed to its existence . It is an assumption – a convention – constructed for sociological or political reasons, just as the disease model of alcohol problems and the chemical imbalance theory of depression were. They were developed in an attempt to reduce stigma or encourage people to take their antidepressants. But no evidence supportive of their veracity has emerged since.
So the concept of ‘mental disorders’ is inadequate to supplant that of CPPs because it is descriptive only – not at all explanatory. But such disorders are therefore also inevitably vaguely and arbitrarily defined and demarcated. “Diagnostic criteria [in psychiatry] shift and sway like in no other area of medicine” . The DSM meanders between at least seven different criteria in distinguishing non-problems from problems-deserving-therapy (mental disorders). At different times DSM specifies: (i) A certain symptom cluster. Three or more symptoms from a field of seven are required to diagnose an Antisocial Personality Disorder. (ii) A certain level of distress. In OCD “the obsessions or compulsions… cause clinically significant distress”. (iii) A level of dysfunction. This is required for a diagnosis of Specific Phobia. (iv) A certain type of etiology. This marks a Posttraumatic Stress Disorder (PTSD) or a Substance-Induced Sexual Dysfunction. (v) A statistical deviation. To diagnose a Female Orgasmic Disorder requires a “marked infrequency” of orgasms. (vi) The chemistry involved. This determines an Alcohol-Related Disorder. And (vii) duration is a criterion determining the presence of an Acute Stress Disorder or Dysthymia . One or two clear conceptual criteria to distinguish CPPs from ‘normal problems in life’ would be much preferred.
When arbitrary categories are forced onto dimensional phenomena like symptoms, then both reliability and validity have been shown to suffer [68,69,70]. Not only have no biological markers for the common mental disorders been uncovered, but this arbitrarily interchangeable collection of criteria for determining their presence means that not a single mental disorder has been established as a discrete categorical entity, as opposed to a dimensional outlier [71, 72].
This conceptual vacuum has left ‘mental disorders’ as merely social constructions , ‘open concepts’ , or ‘practical kinds’ [15, 40] without a true defining essence , and has resulted in intractable and interminable debates among psychologists and psychiatrists as to what ultimately defines a mental disorder . Hence we have seen the inclusion and then exclusion in 1974 of homosexuality , and historical arguments over ‘childhood masturbation disorder’ and ‘lack of vaginal orgasm’ . Right up to today, clinicians and researchers have criticized the seemingly unsystematic and arbitrary addition in DSM-5 of “bizarre new illnesses” such as Excoriation Disorder (skin picking), Major Depressive Disorder 2 weeks after a bereavement, or Somatic Symptom Disorder when one is adjudged ‘too upset’ when told of a cancer diagnosis . New diagnoses have mostly arisen when a few influential insiders have decided that a new category would be clinically meaningful and handy, and have lobbied for its inclusion .
Unlike mental disorders, CPPs must be ‘natural kinds’ of things in the world [77, 78] with a defined essence , that delineates a category else they revert, like mental disorders, to being merely dimensional variants of normality [80, 81]. Such a capitulation, in which we merely choose to call something a mental disorder or a CPP when it has gone far enough to bother us, is a surrender to the opposite, postmodern extreme [82, 83] in which problems may be explained by a “medical narrative” no more or less legitimately than by a learning theory-based narrative, or by “socio-political, spiritual or paranormal explanations” . When a conceptual over-reach leads to the abandonment of the scientific method altogether, the need for a ‘paradigm shift’ is urgently indicated.
It is generally recognized in all fields of scientific endeavour that it is preferable that concepts be essentialist rather than undefined or arbitrary , and categorical rather than dimensional  or merely nominal . For example, it is epistemologically and clinically much more useful if disorders can be conceptualized in terms of pathological processes rather than being solely descriptive . After a century of successful, progressive research and practice in clinical psychology, we are more than confident that CPPs exist in the real world , and are qualitatively different from normality, but depend on a client’s, a therapist’s, or society’s evaluation. That is, that CPPs are a form of ‘harmful dysfunction’  where the harm is a judgement, but the dysfunction is objective .
Though CPPs are almost universally and intuitively regarded as qualitatively different from normality, much theoretically important research has focussed on dimensional constructs and their relationships with (even barely valid) categories such as mental disorders. This has occurred because each approach (dimensional or categorical) has its advantages in different contexts . When testing hypotheses, dimensional measures in research retain more information than categories, especially for phenomena that are distributed fairly continuously, and with unclear boundaries .
Categories are also highly dependent on appropriate cutpoints. We know that cutpoints for most mental disorders are fuzzy and somewhat arbitrary. This is why they are barely categorical. And this is also why most dimensional models of psychopathology focus on the personality disorders . They have even more ‘fuzzy boundaries’ than the other mental disorders.
Hence, recognising the problems that DSM-5 and ICD-10 have had with arbitrary thresholds, large category overlaps, and low clinical utility, the new ICD-11 has adopted a dimensional approach to personality disorder classification with 5 trait qualifiers: Negative Affectivity, Detachment, Dissociality, Disinhibition, and Anankastia . Whereas DSM-5 and ICD-10 require a quota of criteria to be met (e.g. 5 of 9) to define a disorder, ICD-11 diagnoses Personality Disorder via global evaluation of personality functioning, where the clinician may specify 5 evidence-based trait dimensions that contribute to the unique expression of personality disturbance.
It has been claimed that WHO’s ICD-11 is less entirely based on tradition and authority than the DSM  as it is more empirically derived, in that dimensional approaches correspond better to the observed data than do purely categorical ones . Continuous (dimensional) measures of psychopathology have been found to increase both the reliability and the validity of assessments over discrete (categorical) measures . Subsequent to this, Reed, Sharan et al.  found the reliability of ICD-11’s guidelines to be superior to that previously reported for equivalent ICD-10 guidelines, and Reed, Keeley, et al.  found clinicians rated the clinical utility of ICD-11’s diagnostic guidelines very positively with regard to ease of use, but still poorly for treatment selection.
Note that the dimensions of personality functions here refer to processes rather than static traits or features, and therefore this assessment of functioning more closely approximates a functional analysis than a diagnostic checklist. And it therefore can potentially more closely lead to therapeutic conclusions . “[T]reatment should target what the Personality Disorder does to the patient (i.e., severity), as we cannot change what it is (i.e., traits).”
These changes of direction embodied in the soon to be implemented ICD-11 are a response to the recognition that psychotherapeutic interventions are often transdiagnostic . However, they offer only a clinically useful dimensional overlay to a categorical ‘mental disorder’ conceptualisation – not a paradigm shift. “Advances in psychiatric research in general, and progress in nosological science in particular, will continue to be iterative….but no evidence has suggested that genetic or other biological information will lead to a paradigm shift in diagnostic classification in the immediate future” ( p.7).
However, clinicians and clinical researchers generally require a categorical approach  because they need to constantly decide whether to treat or not, whether to hospitalise or not, or which drug or psychotherapy to apply, or not. ICD-11, as well as DSM-5 must be predominantly categorical for administrative and treatment selection purposes. Such decisions are categorical, even if the data underlying them is dimensional .
For our purposes, the cutpoints and criteria for the determination of the presence of a CPP must be clearer, better validated, and more essentialist than those for the diagnosis of mental disorders. Dimensional measures give us indications as to the statistical significance of a relationship or an outcome, but categorical approaches tell us more about their clinical significance – whether a qualitative difference has been created or detected, or merely a possibly-trivial change in one psychometric score.
The usual structure of research projects in this field has been to explore the relationships between dimensional constructs and factors, such as ‘anxiety sensitivity’, and categorical problems, such as generalised anxiety disorder (GAD)  even though GAD as a construct has poor validity, dubious reliability, and an arbitrary cutpoint. Hence, it has been noted that the cycle of progress that is meant to occur between dimensional research in psychiatry and clinical diagnostic accuracy has not eventuated [86, 93].
When dimensional measures are used in research on mental disorders, the richer data has provided even more evidence that DSM diagnoses correlate poorly with these measures, and that they better predict certain psychological processes than psychopathological conditions. For example, Melville et al.  found that problem behaviours, rated or scored dimensionally, among adults with intellectual disabilities, loaded in a factor analysis within an emotion dysregulation/problem behaviour dimension, not within depressive, anxiety, organic, or psychosis dimensions or factors.
Similarly, when the relationship between the dimensional construct ‘dispositional negativity’ and adverse outcomes such as ‘emotional disorders’ is studied , a “dynamic cascade of processes” – presumably amenable to functional analysis – is uncovered; not a simple relationship with one or two diagnosed mental disorders. Further, when dimensional psychopathology measures are factor analysed across a population, a strong general psychopathology factor emerges. Carragher et al.  concluded from this that transdiagnostic treatment approaches are indicated and warranted, and the domain of psychopathology should be restructured in an empirically-based manner, as has recently commenced through the HiTOP consortium .
So dimensional approaches to CPPs or to mental disorders can be very valuable, especially in research. But it would be preferable that any psychological-level alternative to the categorical diagnosis of mental disorders be a more evidence-based and essentialist categorical conception of CPPs.
Therefore, any new conception of CPPs must, by contrast with DSM’s mental disorders, be theory-rich, evidence-based, problematic-by-judgement, real-by-nature (essentialist), categorical (qualitatively distinct from normality) according to only one or two reliable and valid criteria, and must recognise various conceptual kinds of psychological problem.
Treatment-relevant case formulation versus nominative diagnosis
With further regard to the DSM system’s natural taxonomic inadequacies for clinical psychologists, because the vast majority of psychologists seek to intervene at a psychological level (i.e. in situations, thoughts, feelings, and behaviours), they are much more interested in developing a process- or functional- or mechanism-focused case formulation than a symptom-derived diagnosis. Clinical practice is predominantly theory-based rather than manual-prescriptive .
Whereas psychiatric practice is more and more dominated by the process of deriving DSM diagnoses, [32, 97], an examination of the UK’s Generic Professional Practice Guidelines for psychologists, or the APS’s College of Clinical Psychologists Course Approval Guidelines, shows that “training programs for clinical psychologists emphasize formulation rather than diagnosis” ( (p.448). Formulation is fundamental to clinical psychology in the same way that diagnosis is fundamental to psychiatry , and DSM diagnosis is often irrelevant to psychological practice [20, 54].
Among the advantages of the case formulation approach identified by the BPS’s Division of Clinical Psychology [17, 98] are much greater treatment-relevance, strengthening of the therapeutic alliance, normalization of problems, providing a sense of hope, reducing blame, and increasing collaboration and empathy.
With regard to treatment-relevance and clinical utility, even for psychiatrists the DSM “describes a collection of disorders, not an integrated system of psychopathology” ( (p.147). Many existing diagnoses encompass multiple pathological processes . DSM’s search for reliability of diagnosis at the cost of theoretical integration and validity  plus its high rates of comorbidity, high frequency of “Other Specified/Unspecified” (previously “Not Otherwise Specified”) diagnoses, and divergent and overlapping criteria sets make for little guidance in choice of treatments .
But especially for clinical psychologists, who are more concerned with psychological-level case formulation, in most cases a DSM diagnosis tells one little about etiology, course, choice of treatment, or treatment response [52, 99,100,101,102,103]. “Identifying a disorder by its symptoms does not translate into understanding it. Clinicians need some heuristic concept of its nature, grasped in terms of cause or mechanism” ( p.1845).
A clinical psychologist basing treatment on a DSM diagnosis in place of a case formulation is like a dietician assessing the adequacy of a person’s diet by taking a height measurement, instead of interviewing the person. Height measurement is a much more reliable, consistent, brief, and precise process, but far too much validity and treatment relevance are lost.
Therefore, the new conception and taxonomy of CPPs must be formulation-relevant and treatment-relevant, and not just a listing of surface-symptom-defined diagnoses.
Processes or mechanisms versus topographical description
Tucker  admitted that, by contrast with the rest of modern medicine, in psychiatry “we are still doing pattern recognition” (p.159). This approach falls down when the same pattern or topography can be established by diverse processes, or when different topographies can come from the same process . These phenomena have come to be known as the problems of multifinality and divergent trajectories . Multifinality  refers to how one general transdiagnostic risk factor or process can result in several different disorders, as when stress  can contribute to the development of several different CPPs involving anxiety, depression, or alcohol abuse. Divergent trajectories occur when more proximal or moderating variables, such as attentional biases, result in OCD in one person, but a sleep disorder in another .
The DSM system has ignored these issues and focused on final symptoms and their topography, saying nothing about mechanisms. We need to comprehend psychopathological disorders “not simply by their outward show but by the causal processes and generative mechanisms known to provoke them” ( p.1855).
Major problems such as treatment irrelevance and excessive unexplained comorbidities have resulted [28,29,30]. In the clinical psychological literature these comorbidities, such as between Major Depressive Disorder, Generalized Anxiety Disorder, and Dysthymia , are assumed to reflect the many similarities of inputs, symptoms, and processes among the various models of particular CPPs.
All of these factors have led to the recent development of transdiagnostic models of psychopathology [109,110,111] which seek to identify fundamental processes underlying multiple, often comorbid, psychopathologies [105, 112].
Mechanisms across the disorders, such as negative attentional bias , experiential avoidance , safety behaviours, or rumination  have been studied, and transdiagnostic treatment programs that target these processes rather than individual diagnoses have then been developed [4, 109, 116, 117]. This naturally follows the finding that more than half of patients who present with depressive disorders also have elevated comorbid anxiety symptoms, and that, when psychotherapy for depression is undertaken, anxiety can be significantly ameliorated . Very few studies have examined this issue because the literature is strictly structured around individual psychiatric disorders .
Current transdiagnostic approaches circumvent the problem of the plethora of manualized treatment programs for a growing number of specific diagnoses , so that the training of therapists and development of treatment packages can be more parsimonious .
The taxonomic arm of this empirically-based transdiagnostic movement – the Hierarchical Taxonomy Of Psychopathology (HiTOP) consortium – grew out of the psychological study of individual differences . Its rich vein of studies [121,122,123] establishing an alternative dimensional organization of psychopathology helps to overcome such problems with traditional nosologies as the issue of arbitrary thresholds and subsequent loss of information, ensuing reliability problems, diagnostic heterogeneity, theoretically disruptive high comorbidities , and exclusion of undiagnosable ‘subthreshold’ people with serious CPPs [49, 69, 124].
The emergent HiTOP dimensions form a hierarchy with five levels (symptoms, syndromes, subfactors, etc.), and can thus help explain why disorders from different classes respond to the same treatment (e.g. social anxiety responding to antidepressants) . In this way it is a critical part of the transdiagnostic movement.
But the HiTOP hierarchical dimensional models of classification, though guided by research  are still the result of a consensus among the consortium , require interpretation by human experts , and the approach suffers from all the problems of a dimensional taxonomy. It has not, to date, been used clinically, as the consortium has yet to develop meaningful cut-off points for pathology . It can still only offer a dimensional elaboration, based on symptom measurements, on top of a categorical ‘disorder’ model , because it still does not implicate proximal causes for, and the ‘essence’ of, CPPs. It is a descriptive phenotypic model, and does not directly incorporate etiology and underlying mechanisms . It shares many of the same constructs with the categorical model frameworks  such as a focus on ‘mental disorders’.
For example, when Nolen-Hoeksema and Watkins  have suggested ways to explain multifinality and divergent trajectories in terms of distal, proximal, and moderating causes or risk factors (p. 592), they have done so via a flow chart resulting in (DSM-type) ‘Disorder A’, ‘Disorder B’, and ‘Disorder C’. Despite emphasising that we need more focus on the precise mechanisms involved (p.591), the transdiagnostic movement still regards an Anxiety Disorder as the same CPP whether it has arisen through a mechanism of avoidance or of rumination. The necessary ‘paradigm shift’ would see these two situations as different CPPs.
As psychological interventions increasingly target mechanisms, such as specific cognitive dysfunctions, rather than symptom-based mental disorders, a new comprehensive conceptual framework to assemble the results of psychotherapy research will be required .. The transdiagnostic movement has not to date offered a ‘paradigm shift’; only a useful extra (dimensional) layer, such as allowing for variables such as ‘neuroticism’ or ‘extraversion’ in treatment selection , on a categorical ‘mental disorder’ conceptual system.
But also, these transdiagnostic processes and properties are dimensional responses to problems with the categorical assumptions of DSM . The assumption is that the heterogeneous disorders in DSM are made up of dysfunctional versions of processes that vary along continua in the general population [45, 126]. For example, attentional bias toward negative information is common in people without . Within the transdiagnostic movement to date, such a bias cannot be regarded as essential or diagnostic. So there remain problems of cut-off points, a quality distinction between different problems rather than a quantity distinction, and the very definition or essence of CPPs. A categorical conception of CPPs is preferable [61, 84, 86]. It is much preferable that CPPs, unlike mental disorders, display an essence – that they be more than just ‘worse than normal’.
It has been argued that dimensional data can lead to actionable ‘diagnoses’ in medicine , so why not in clinical psychology? For example diagnoses are determined, and treatments initiated, from blood pressure measurements and fasting glucose levels using indicative ranges of scores. However, even in medicine, this is regarded as second best. It is much preferable to uncover some clear, qualitatively distinct pathology such as an infection or a lesion, than to find that a score looks too high or too low. Is it better to treat every adult person under a height of 4′6″ with growth hormone, or to reserve this treatment for people who are not producing their own growth hormone?
Hence, the new conception of CPPs will focus on mechanisms and processes, not states or conditions. But beyond the current transdiagnostic movement, it will regard the operation of these processes as essential, definitive, and ‘diagnostic’. Thus a categorical conception will emerge, not a merely dimensional one.
Some CPPs are clearly not mental disorders
In examining what clinical psychologists actually address in research and practice, a stark example of the non-equivalence of CPPs and mental disorders can be found in the fact that clinical psychologists address relationship, marital, and family problems using the exact same assessment and treatment models as for, for example, anxiety or depression problems. Such interpersonal situations clearly cannot be conceptualised as internal mental disorders, and so DSM has relegated “relational pathology” to a terse footnoted ‘V’-code listing, an omission long lamented [127, 128]. As a bizarre and unfortunate consequence, when a clinical psychologist sees a couple or family in the Australian Medicare system they are not eligible for a fee rebate unless one attending party has been given a mental disorder diagnosis by the referring medical practitioner and is being treated for this. ‘No blame’ relationship therapy will not be rebated.
A similarly bizarre and unfortunate result of the conceptual medicalization of CPPs arises with parenting problems. Patterson  has described how parent-child interactions frequently directly reinforce deviant behaviour, and he has outlined the role of parent-child discipline practices in the development and maintenance of aggressive behaviour in children. These insights led to the development of the most empirically supported treatment for such problems – Parent Management Training . But, again, to be eligible for a rebate in Australia, not only must the child have a diagnosis of, for example, Conduct Disorder, but the child must attend each consultation. The assumption is that the problem resides within the child, as would a lesion or infection, and so the mental disorder must be in attendance for treatment to be conferred.
However, perhaps the largest class of CPPs effectively addressed by clinicians, but barely researched because they are not ‘mental disorders’, lies in the third to half of all people who seek clinical psychological help but cannot be given a clear diagnosis because their problems do not fit criteria and categories neatly . They may be ‘subthreshold’ , or ‘subclinical’ , or situation-specific (such as being evidenced only at work). High levels of distress commonly occur in the absence of a diagnosable condition , as when one or two symptoms occur very strongly, but three or more are required for a diagnosis . Should clinical psychologists turn away people presenting with such CPPs because they do not have a diagnosed mental disorder?
Therefore, not all CPPs are internal mental disorders detectable and definable by a certain intensity of symptom presentation. They are more likely to be particular sorts of psychological-level processes, which can occur between people as well as within them.
Social consequences of seeing all CPPs as mental disorders
Promotion of the disease model of CPPs has often occurred in an attempt to ameliorate the serious stigma consequent upon the ‘moral failing’, sinfulness, or demonic possession models of CPPs [136, 137]. The medical model has been advanced as a simple solution to the “brain or blame” dilemma or the “chemistry or character” dichotomy as to whether a person’s mental suffering is real, or they should be told to pull themselves together .
Clinicians’ models affect the community’s beliefs and hence sympathetic or stigmatizing attitudes . Psychiatry was aware of this when it promoted the disease model of alcoholism in an attempt to reduce stigma and punitive responses, and increase treatment takeup and compliance . Support groups have used the fact that the concept of mental illness has been arbitrarily defined to agitate for problems such as depression or alcohol dependence to be regarded as diseases, hoping to reduce stigmatization and increase service or research funding .
However, in many areas this strategy has backfired, and the personal and social consequences of a psychopathological label have proven to be negative, fatalistic, adverse, and stigmatizing . The “disease like any other” campaigns to convince the public that mental disorders are non-volitional biological illnesses for which sufferers do not deserve blame and discrimination have been “an unequivocal failure in reducing stigma” ( p.852). For example, with regard to attitudes to depression and schizophrenia, Schomerus et al.  found that (a) belief in the biomedical model has increased, (b) acceptance of medical treatment has increased, but (c) attitudes toward people with mental disorders has not improved.
A diagnosis of a mental disorder can often be a cause of disempowerment and social exclusion , and may label the person rather than the problem . A diagnosis of mental illness is known to negatively affect self-identity, attract stigma , result in a negative prognosis, and engender isolation [144, 145]. People who believe that mental distress is a kind of biological illness are more likely to see psychiatric patients as dangerous and unpredictable [146, 147]. They may blame less, but will fear and avoid patients more , and will assume a worse prognosis [147, 149].
Such deleterious consequences are exacerbated by DSM’s assertion that all of the following are examples of the one kind of thing. They are all equally ‘mental disorders’: Mild Tobacco Use Disorder, Schizophrenia, Female Orgasmic Disorder, Delirium, Restless Legs Syndrome, Alzheimer’s Disease, a Spider Phobia, and Opioid Intoxication. Admitting to sadness 2 weeks after one’s spouse has died can put one in the same class, conceptually, as a paranoid schizophrenic, a smoker, a person suffering a panic disorder, or a violent psychopath.
These consequences of problem assessment and problem formulation are not inevitable. It has been claimed that a psychological case formulation or functional analysis approach both avoids the problem of stigmatization  and the abdication of responsibility  of a mental disorder diagnosis. This provides further reason that the new conception of CPPs needs to be psychological-level and formulation-based.
Research on CPPs versus mental disorders
Although in practice clinical psychologists formulate much more than they diagnose, almost all research in the discipline ignores this fact. To be considered methodologically sound, and hence to qualify for funding, almost all psychotherapy research must be undertaken with formally diagnosed subjects with the intention of ‘curing’ them of their mental disorders by removing their symptoms. However in real-world clinical practice case formulation guides treatment, which targets psychological processes, not symptom profiles. Treatment outcome measured by “escape from diagnosis” is in this light arbitrary, misleading, and inadequate.
Research trials have typically treated highly selected groups with a single diagnosis, while in clinical practice patients have many comorbidities and atypical symptom profiles [54, 119]. Clinicians are more likely to apply several interventions, and will base this on the individual case formulation they have developed, on the assumption that each technique is targeting something different. When experimental subjects are merely diagnosed and then randomly allocated to comparative treatment groups, they will have an undetermined distribution of relevant underlying mechanisms . A ‘package’ approach ignores basic psychological science and the individual needs of individual clients, is atheoretical, and alienates research from clinical practice .
Important comparative studies on various CBTs for depression, such as cognitive therapy (CT) versus behavioural activation (e.g. [152, 153]), or for anxiety problems, such as exposure therapy (ET) versus CT (e.g. [154, 155]), have not been able to find consistent differences between comparative treatments . Michelson et al. , for example, were unable to separate the benefits of cognitive, behavioural, and psychophysiological treatments for agoraphobia, though all three were superior to a wait-list control. This is unsurprising, though, when subjects are DSM-diagnosed and then randomly allocated to groups, as though they all have the same CPP. It assumes a diagnosis-to-treatment-selection link. This is an example of theory governing the nature of research. However, the medical model of DSM is so entrenched that many researchers would not even see this difference ( p.157).
“There are undoubtedly many functionally distinct subtypes of patients currently mixed together in popular diagnostic systems” ( p.971). For example, agoraphobics may have a classically conditioned fear of separation, or a fear of panic attacks. Further, this latter fear may in turn be of medical catastrophes or of social embarrassment . Over a quarter century ago, Wolpe [157, 158] warned that such neglect of individual differences in the dysfunctional processes that occur within a diagnostic group puts us in danger of making a mockery out of group treatment outcome research.
A major motivation of the HiTOP consortium has been the fact that randomized controlled trials (RCTs) rarely show superiority among thoughtfully conceived treatment packages , and that research has found that many interventions can be beneficial with a host of problems regarded as distinct categorically .
This problem has been thoroughly outlined by Smith, McCarthy, and Zapolski , who have pointed out that assessing the effect of CT versus ET on a DSM-defined ‘Depression’ group is an example of assessing the relationship of a construct or variable with another multidimensional construct or measure (such as PTSD or Neuroticism) which has multiple (diagnostic) criteria. The resultant composite correlation will be an average of the correlations with each of the dimensions or criteria, each of which could correlate quite weakly with the others.
The power of RCTs is seriously compromised when the groups that subjects are randomized into are vaguely or spuriously defined. “With heterogeneous treatment effects, the ATE [average treatment effect] is only as good as the study sample from which it was obtained” . This is why researchers have begun to focus on transdiagnostic mechanisms of intervention . “Diagnostic heterogeneity compels the clinician to go beyond the assigned diagnosis and generate individual-level formulations that are not codified in the diagnostic scheme” ( p.6).
It will be of much greater benefit when we are able to assemble research results into clinical guidelines not on ‘the treatment of Depression’ or ‘of Bulimia Nervosa’, but on psychological interventions with CPPs A and B, defined by mechanisms, which may cross diagnoses or differ within a diagnosis. For example, we know that targeting specific mediating cognitive processes in a social phobia is more effective than standardized generic cognitive-behavioural treatment , because the mental disorder ‘Social Anxiety Disorder’ can encompass a number of (mechanism-defined) CPPs.
A new conception of CPPs must therefore be: (a) A psychological-level one (i.e. involving cognitions, behaviours, emotions, and situations); (b) Psychologically theoretically rich and evidence-based; not a postmodern ‘categories-by-convention-only’ model. It must define an essence. If it comprises a ‘harmful dysfunction’ (Wakefield, 1992), then its harmfulness must be a matter of subjective judgement, but its dysfunction must be defined objectively; And according to only one or two criteria, not a hodgepodge of them; (c) Categorical, rather than merely dimensional; (d) Encompassing of all problems currently appropriately and successfully addressed by clinical psychologists; not merely diagnosed mental disorders; And (e) better at avoiding the stigma and responsibility-confusion problems which have been exacerbated rather than ameliorated by the disease model.