Scientific validation of the ICD-11 CDDR

Abstract

The process of scientific validation of the ICD-11 Clinical Descriptions and Diagnostic Requirements (CDDR) for Mental Disorders has spanned more than 10 years, being remarkably comprehensive and inclusive as well as truly international, with the involvement of many hundreds of clinicians and researchers from all regions of the world.

The field trials of the ICD-11 CDDR – contrary to those of the ICD-10 Clinical Descriptions and Diagnostic Guidelines (CDDG) and the DSM-5 diagnostic criteria – have been genuinely “developmental” (as opposed to “summative” or “evaluative”) in their nature¹. That is, they have been conducted before the finalization of the relevant sets of criteria or guidelines – thus allowing the identification of conceptual or terminological problems in the proposed texts, the correction of those texts, and the further testing of the revised versions – rather than being conducted or concluded after the finalization of the relevant sets of criteria or guidelines, thus just providing information to clinicians about what they could expect from those products. As we will see, several amendments to the CDDR were actually implemented as a consequence of this design.

The CDDR field trials can be subdivided into two main groups: a) Internet-based trials, implemented through the Global Clinical Practice Network (involving, at the time when the trials were conducted, more than 15,000 mental health and primary health care professionals from more than 150 countries), which used a case vignette methodology to assess the effects of specific differences between the CDDR and the ICD-10 CDDG on the participants’ clinical decision making; b) clinic-based (or ecological) trials, assessing the reliability and clinical utility of the CDDR in real clinical contexts. The clinic-based trials differed from the DSM-5 field trials in that they used a joint-rater design (with two clinicians jointly interviewing each patient) rather than a test-retest design (with two clinicians separately interviewing each patient at different time points), thus controlling for information variance and more specifically testing the reliability of the proposed guidelines (rather than testing more generally the reliability of the relevant psychiatric diagnoses)².

Among the Internet-based CDDR field trials, of special interest have been those focusing on disorders specifically associated with stress³, and on feeding and eating disorders⁴. A case-control field trial on the former grouping of disorders³, conducted with 1,738 mental health professionals from 76 countries, found that several changes introduced in the ICD-11 – including the addition of complex post-traumatic stress disorder (complex PTSD) and prolonged grief disorder – resulted in significantly improved diagnostic decisions. However, the trial also identified some problems with the proposed CDDR text (including difficulties with interpretation of the “re-experiencing” criterion for the PTSD diagnosis, and in differentiating prolonged grief disorder from normal bereavement), which led to a revision and further validation of the text.

Similarly, a case-control field trial on feeding and eating disorders⁴, conducted with 2,288 mental health professionals representing all world regions, found that the changes introduced in the ICD-11 CDDR improved the diagnostic accuracy and clinical utility compared to the ICD-10 CDDG. However, the trial also identified difficulties in determining whether a person with a diagnosis of anorexia nervosa was recovered, as well as problems in the identification of binge eating episodes, which led to a refinement of the definition of recovery for anorexia nervosa and to the specification that the subjective experience of loss of control over eating and related distress is a pathognomonic feature of binge eating even when the amount of food consumed is not objectively large.

Among the clinic-based CDDR field trials, the largest one – conducted among 1,806 patients in 13 countries and focusing on mental disorders which account for the greatest proportion of global disease burden (schizophrenia and other primary psychotic disorders, mood disorders, anxiety and fear-related disorders, and disorders specifically associated with stress) found that intraclass kappa coefficients for diagnoses ranged from 0.45 (for dysthymic disorder) to 0.88 (for social anxiety disorder), and could be considered moderate to almost perfect for all diagnoses, with an overall reliability superior for the CDDR compared to the ICD-10 CDDG². The same trial also found that the CDDR were perceived as easy to use, corresponding accurately to patients’ presentations, clear and understandable, providing an appropriate level of detail, taking about the same or less time than clinicians’ usual practice, and providing useful guidance about the distinction of each disorder from normality and from other disorders⁵.

The somewhat long period of time elapsed between the production of the first draft and the finalization of the CDDR has also allowed an extensive and detailed validation of some of the categories newly introduced in the ICD-11. Emblematic in this respect has been the validation of the new categories of prolonged grief disorder and complex PTSD.

According to a PubMed search performed on February 8, 2024, the papers with original data published since 2013 on prolonged grief disorder have been 57. Overall, they documented the construct validity of the new category, its differentiation from other disorders (e.g., depression and PTSD); its association with marked functional impairment (beyond the effects of concomitant disorders); and its higher consistency with patterns of prolonged grief in longitudinal studies compared to DSM-5 persistent complex bereavement disorder. This evidence has led to the inclusion of prolonged grief disorder in the DSM-5-TR⁶.

A PubMed search performed on the same date detected 199 papers with original data published since 2013 on complex PTSD. Overall, they documented the validity of the new category, and its differentiation from PTSD in a wide range of cultures, in children and adolescents as well as in adults, and across several traumatized populations⁷. This research has been facilitated by the development and validation – prompted by the circulation of the CDDR drafts – of a new specific assessment instrument, the International Trauma Questionnaire.

A further significant difference between the ICD-11 CDDR and the DSM-5 diagnostic criteria which has been tested internationally is that concerning the classification of severe irritability in children and adolescents. A study conducted with 196 clinicians from 48 countries⁸ found that the formulation proposed in the CDDR (using chronic irritability-anger as a specifier for the diagnosis of oppositional defiant disorder) led to a more accurate identification of severe irritability and a better differentiation from boundary presentations compared to both the DSM-5 solution (introducing the new category of disruptive mood dysregulation disorder) and the ICD-10 classification (listing oppositional defiant disorder as one of several conduct disorders without attention to irritability). Participants using the DSM-5 often failed to apply the diagnosis of disruptive mood dysregulation disorder when it was required, whereas they more often applied a psychiatric diagnosis to irritability that was normative in relation to the developmental stage.

One more innovative aspect of the process of validation of the ICD-11 CDDR has been the systematic involvement of experts by experience, through an international study (INCLUDE) conducted in India, the UK and the US⁹. This study collected users’ input on five diagnoses: depressive episode, generalized anxiety disorder, schizophrenia, bipolar I disorder, and personality disorder. Overall, the CDDR were in many cases perceived as useful and relevant to lived experience.

Of course, a currently missing key element in the scientific validation of the CDDR is their performance in ordinary clinical practice. The ongoing efforts to translate them in as many languages as possible, to widely disseminate them, and to accelerate their endorsement and implementation by national governments, will be crucial in this respect, and a regular update of the text (every two years)⁷, to be performed on this basis, is already being planned.