Mental health classifications in primary care: commentary, Dowrick

Response

To be fit for purpose, a classification system for mental health problems in primary care needs to do three important things: differentiate between normality and disease, clarify interactions between mind and body and indicate levels of severity and complexity, to guide decisions on whether to treat or refer.¹

There are currently four classification options available for use in primary mental healthcare: the Systematized Nomenclature of Medicine Clinical Terms (SNOMED CT), the fifth edition of the Diagnostic and Statistical Manual (DSM-5), the third edition of the International Classification for Primary Care (ICPC-3) and the 11th edition of the International Classification of Diseases (ICD-11). This commentary considers the strengths and limitations of each option, with particular focus on ICD-11, and suggests strategies to improve nosology and to enhance integration of primary and specialist mental healthcare.

SNOMED CT

SNOMED CT is a structured clinical vocabulary, currently embedded in the electronic record systems used by family doctors and other primary care professionals across the UK. It is also licensed, though less regularly used, in other countries including the USA and Canada. It is proposed as the most comprehensive and precise clinical health terminology in the world.² It provides multiple codes, roughly four times as many as DSM-5 and three times as many as ICD-11.

The main advantage of SNOMED CT for daily clinical practice in primary care is its ease of use, especially when embedded in primary care electronic records. However, its main limitation is that it is not really a diagnostic classification system at all. There are considerable uncertainties regarding its ontological commitment, that is to say what kind of entity is an instance of a given concept. In the electronic record, ‘Problem’ is the standard header, but there is no indication as to whether this refers to disorder (the closest term to diagnosis), or to observation, situation, findings or procedure. This offers considerable clinical latitude but comes with no expectation of a formal diagnosis. Nor are there any underlying descriptions of relevant symptoms and signs to guide clinicians towards a particular formulation. Clinical terminology, however structured and comprehensive, is not the same as clinical diagnosis.

DSM-5

DSM-5 is published by the American Psychiatric Association and serves as the principal authority for psychiatric diagnoses in the USA. In the context of gradual though as yet incomplete harmonisation with ICD-11,³ there are two elements of particular relevance to primary care.

The first is the well reported obfuscation of the boundaries between grief and depression, allowing a diagnosis of major depressive disorder to be made just 2 weeks after a bereavement. While this may enable patient access to treatments in the USA insurance-based healthcare system, there are major concerns about turning grief and other life stresses into mental disorders, and the consequent medical intrusion on personal emotions. It adds unnecessary medication and costs and distracts attention and resources from those who really need them. It also leads to considerable confusion for primary care diagnosticians.

Second, in relation to interactions between psyche and soma, the DSM-5 classification of somatic symptom disorder places emphasis on cognitive elements. Specifically, for a diagnosis of somatic symptom disorder, at least one of three psychological criteria should be present: (a) health anxiety, (b) disproportionate and persistent concerns about the medical seriousness of the symptoms and (c) excessive time and energy devoted to the symptoms or health concerns. This is substantially different from ICD-11.

ICPC-3

The third edition, endorsed by the World Organisation of Family Doctors (WONCA), now commands considerable attention worldwide. ICPC-3 is the preferred classification system for the WHO Primary Health Care team in its drive towards universal healthcare.

ICPC-3 is based on the content of the primary care consultation and includes both professional and patient perspectives. It covers reasons for encounter, including symptoms and complaints, episodes of care, functioning and social problems.⁴

Looking specifically at common mental health problems, ICPC-3 has codes for psychological symptoms or complaints, such as feeling anxious, nervous or tense or feeling sad. Episode of care codes for psychological diagnoses made by the clinician include anxiety disorder or anxiety state, bodily distress or somatisation disorder, depressive disorder, and mixed depressive and anxiety disorder. ICPC-3 also includes medically unexplained symptoms (MUS) as a code for a long-lasting symptom diagnosis.

ICD-11

ICD-11 has considerable potential value as a classification system in primary care. A simplified version has been provided by the WHO⁵ and for the most part appears fit for primary care purposes. However, there are two fields where substantial diagnostic uncertainty remains: the boundary between anxiety and depression; and the conceptualisation of somatoform disorders, which – despite their high prevalence – are relegated to an ‘other’ category in the WHO's primary care version.

Anxiety and depression

Given the ubiquity of symptoms of anxiety and depression seen in primary care settings, it is essential to have an accessible and coherent way of understanding their relationship. ICD-11 proposes a diagnosis of mixed depressive and anxiety disorder in cases where neither set of symptoms, considered separately, is sufficiently severe, numerous or persistent to justify a diagnosis of another depressive disorder or an anxiety or fear-related disorder.

An alternative proposal is to make a single compound diagnosis of anxious depression if both anxiety and depressive symptoms are present at case level for at least 2 weeks.

The rationale for the proposed diagnosis of anxious depression emerges from research conducted during field trials. In a study conducted in primary care centres in Brazil, China, Mexico, Pakistan and Spain, and based on either perceived psychological distress or distressing somatic symptoms, practitioners referred patients to a research assistant who administered a computer-guided diagnostic interview. Complete data were obtained for 2279 participants. Anxious depression was the most common diagnosis (48%), compared with generalised anxiety disorder (42%) and mixed anxiety and depressive disorder (45%). One-third of those diagnosed with anxious depression had anxiety lasting less than 3 months, but these participants reported as much disability and suicidal ideation as those with longer duration of symptoms.⁶

However, there is concern that the two-week criterion for caseness runs the same risk as the DSM-5 criterion for major depressive disorder, in blurring the boundary between psychiatric disorder and normal reaction to adverse life events. There is also potential for confusion among primary care practitioners, simply because it is different from ICD-11. And the presence of differing criteria may impede plans for effective integration between primary and mental health services.

Bodily distress or bodily stress?

The ICD-11 diagnosis of bodily distress disorder (BDD) relates to bodily symptoms that are distressing and involve excessive attention and, perhaps, repeated contact with healthcare. With BDD, if a medical condition is present, the attention is excessive in relation to its nature and progression. Bodily symptoms and associated distress are persistent and associated with significant impairment. Although BDD will typically involve multiple bodily symptoms that may vary over time, there may be a single symptom – usually pain or fatigue – that is associated with the other features of the disorder.

Clinical academics focusing on primary care have recommended the different diagnosis of bodily stress syndrome (BSS) on the basis that – in primary care – single unexplained somatic symptoms are so common. BSS specifies the presence of at least three symptoms not explained by known medical pathology and which are associated with distress or impairment. It also eliminates the requirement that the primary care practitioner makes a judgement about whether the attention devoted to the symptoms is ‘excessive’. BSS is differentiated from health anxiety, with the latter characterised by persistent and/or intrusive ideas or fears of having an illness that cannot be stopped (or only stopped with great difficulty) or intense preoccupation with minor bodily sensations or problems that are misinterpreted as signs of serious disease.

The rationale for the proposed diagnosis of BSS is also based on evidence from field trials conducted in primary care. Unlike DSM-5, where emphasis is on cognitive elements, the focus here is on the substantial overlap between somatic and psychological symptoms.

A study involving 587 individuals diagnosed by primary care practitioners with either BSS or health anxiety identified a mean of 11 symptoms per individual: 70% had both BSS and health anxiety, 79% had co-occurring diagnoses of anxiety, depression or both, and 56% had a diagnosis of anxious depression. Levels of disability were high, with a mean WHO-DAS score of 13.⁷ In a subsequent statistical analysis of 797 individuals with somatic complaints, depressive and anxious symptoms, two bi-factor models fitted the data. The first model had all symptoms loaded on a general factor, along with one of three specific depression, anxiety and somatic factors. The second had a general factor and two specific anxious depression and somatic factors.⁸

These studies conclude that mood and anxiety disorders are likely in the presence of multiple somatic symptoms, and suggest that depressive, anxious and somatic symptoms are best understood as different presentations of a common latent phenomenon.

While the arguments for common latent phenomena are well made, there remain concerns about the proposed nomenclature. As with the diagnosis of anxious depression, the presence of differing criteria between standard and primary care versions of ICD-11 may be confusing to primary care practitioners, and adversely affect opportunities for integration between primary care and specialist services. Moreover, what BSS, BDD and DSM-5's somatic symptom disorder all have in common is that they confer diagnostic certainty on a set of problems that are characterised by uncertainty. They also place the locus of responsibility for the problem with the patient. Many primary care colleagues internationally prefer to retain the term ‘MUS’ for two reasons: it is a working hypothesis and is not diagnostically prescriptive; and it locates the problem – and hence the responsibility for its resolution or mitigation – as one to be shared between patient and practitioner.⁹

Prospects

Although SNOMED CT is currently dominant as a clinical vocabulary (at least in the UK), its lack of underlying diagnostic precision means it serves no useful purpose as a coherent classification system. Indeed, it may obstruct the emergence of such a system, if practitioners and policymakers continue with the assumption that one already exists.

There is an urgent need for convergence between ICPC-3 and ICD-11. These classification systems already have a great deal in common. Remaining differences, although sometimes keenly contested, are resolvable, given willingness to discuss and negotiate. In the UK setting, there are opportunities for conversation between the Royal Colleges of Psychiatrists and General Practitioners. Internationally, the WHO may usefully enable further dialogue between, for example, WONCA and the World Psychiatric Association. The result should be both a more effective nosology and greater integration of care between primary and specialist mental health services.