ICD-11 – fielding for diagnostic success

Dr Geoffrey Reeds, WHO, Switzerland, outlined the different evaluative field studies that were being undertaken for ICD-11 - ecological and web-based field studies.

An alternative view to the future of psychiatric diagnosis offers interesting insights on improving current psychiatric practice

Diagnosis in the natural setting

Ecological implementation field studies evaluate the clinical utility and usability of the proposed ICD-11 diagnostic guidelines in natural conditions i.e. the setting in which they are intended for use. Such field studies evaluate the reliability of diagnoses that account for the greatest portion of disease burden and mental health services utilization in clinical settings established through International Field Study Centers. Essentially, the ease and utility of diagnoses made by the referring/treating clinician and a participating/non-treating clinician using ICD-11 were independently compared. A total of 28 sites in 13 countries participated.

Considering a range of mental health disorders, there was generally a substantial level of agreement between clinicians in diagnosing most conditions using ICD-11. Compared with IDC-10 field trials in which more complex and detailed diagnostic instruments were used, there appeared (with three notable exceptions; acute and transient psychosis, panic disorder and post-traumatic stress disorder) to be an incremental increase in diagnostic agreement in using ICD-11. Thus it appears that a more clinician-friendly ICD presentation is not inherently less reliable.

A more clinician-friendly ICD presentation is not inherently less reliable

Diagnosis on the web

Case-controlled field studies were completed through the internet via the Global Clinical Practice Network (GCPN) to examine the accuracy and consistency of diagnostic judgements using ICD-11 versus ICD-10, using standardized material. Professor Wolfgang Gaebel, Dusseldorf, Germany, explained how the German Society for Psychiatry, Psychosomatics and Psychotherapy (DGPPN) conducted a comprehensive field study (in collaboration with 4 other German medical societies) to support development of ICD-11 in mental health.

Over a 6-month period, two separate, web-based studies evaluated the consistency and utility in diagnosis (WP1, GCPN platform) and code assignment (WP2, ICD-11 FiT tool) of ICD-11 versus ICD-10.  Not unexpectedly, ICD-11 exceeded ICD-10 in the consistency of diagnosis due to new clinical descriptions and diagnostic guidelines; it was also significantly quicker and simpler to use. However, consistency in code assignment favored ICD-10. The time taken to assign a code was shorter with ICD-10 versus ICD-11, although utility assessments again showed advantages for ICD-11. These results suggest that there is further room for improvement in ICD-11.

In addition, Professor Gaebel suggested that these results emphasize the importance of extensive dissemination and implementation activities and the need for thorough training when ICD-11 is released. He also advised that the results presented were preliminary and should be interpreted with caution due to the low numbers recruited to the studies over the relatively short time period available. For the GCPN part, continued recruitment is ongoing to achieve the numbers needed for accurate hypothesis testing.

Predicting diagnosis in psychiatry – beyond ICD-11

What about the future and diagnostic guidelines beyond ICD-11? Professor Mario Maj, Naples, Italy, presented his personal perspectives of ICD-11 and the future of diagnosis in psychiatry.

As he explained, psychiatry consists of patterns of related, reported experiences – symptoms – and observed behaviors – signs – which permit inferences to be made about the further disease course and management. However, it is no longer assumed that these patterns of mental disorders are ‘natural’ discreet entities. In his opinion, the IDC-11 chapter on mental disorders is likely, therefore, to be a collection of intrinsically different constructs.

It is no longer assumed that mental disorders are ‘natural’ discreet entities

Co-morbidity is intrinsic to mental illness

Professor Maj then gave the following reasons for his position. The notion that patterns of mental disorder are independent of each other has proved to be invalid. Co-morbidity is not an artefact of our current diagnostic system, it is an intrinsic feature of psychopathology, he stated.

Arbitrary thresholds

Additionally, there is no clear boundary between ‘normal’ and ‘sick.’ Research on all of the main patterns of mental disorder suggest continuity between the full diagnosable forms of illness and some experiences, behaviors or traits which appear to be relatively common within the population as a whole. Thus, the threshold for a clinical diagnosis has to be drawn arbitrarily mostly on the basis of severity or functional impairment and validated on the basis of its predictive value in terms of outcomes and choice of treatment.

The threshold for a clinical diagnosis has to be drawn arbitrarily mostly on the basis of severity or functional impairment and validated on the basis of its predictive value in terms of outcomes and choice of treatment

Therefore, a simple deterministic etiological model, such as applies to infectious disease, cannot be extended to mental disorders. Instead, these are the product of the complex interaction of a variety of vulnerability and predictive factors. As he pointed out, most genetic variants are non-specifically associated with a range of patterns of mental disorder and even episodic and persistent cases of the same pattern of mental disorder many have different vulnerability factors e.g. episodic depression is apparently more strongly heritable while persistent depression is more strongly linked to childhood maltreatment.

If this is the case, the relationship between a given pattern of mental disorder and the response to a given treatment is likely to involve an element of chance. Thus, the current identification of a given pattern of mental disorder, be it ICD-10 or DSM, while it has clear therapeutic implications, is insufficient to guide the choice of treatment.

A simple deterministic etiological model cannot be extended to mental disorders

ICD categories clinically useful?

Professor Maj suggested that the principal clinical utility of a diagnostic system is in communication between clinicians, determining implementation characteristics (e.g. ease of use) in clinical practice, and usefulness in selecting interventions and in making clinical management decisions. However, he questioned whether ICD categories are the most clinically useful means of describing the patterns seen by psychiatrists.

Alternatives to ICD

At present, there are two main alternatives to ICD categorization. The first states that the domain of psychopathology can be more efficiently described in terms of dimensions, while the second believes that the neurobiological underpinnings of psychopathology should be the major drivers of psychiatric diagnosis and classification. Either or both approaches, in order to emerge in the future as real alternatives to the ICD, will have to show at least equivalent clinical utility. But are they really alternatives to ICD?

Dimensional and neurobiological alternatives - elements of each will gradually be incorporated into the further characterization of individual cases

The ICD, although a useful guide to clinical management, is not sufficient for that purpose. Some important domains are missing, such as degree of severity, antecedent and concomitant variables. It may be that these dimensional and neurobiological alternatives will not inform the basis of a reclassification of psychopathology, but that elements of each will gradually be incorporated into the further characterization of individual cases.

Technology and listening to patients can happen now

The good news is that neither of these developments require revision of ICD-11 - they can commence now.  Professor Maj concluded by stating that we should promote the construction and validation of tools to guide the clinician systematically in the characterization of each individual case. Furthermore, there needs to be a special focus on the assessment of psychopathological dimensions and an exploration of relevant antecedent and concomitant variables; such elements of the approaches that were presented as alternatives to the ICD can be incorporated into current practice.