Changes from ICD-10 to ICD-11 and future directions in psychiatric classification


Abstract

This article provides a brief overview of the changes from ICD-10 to ICD-11 regarding the classification of mental, behavioral, or neurodevelopmental disorders. These changes include a new chapter structure, new diagnostic categories, changes in diagnostic criteria, and steps towards dimensionality. Additionally, we review evaluative field studies of ICD-11, which provide preliminary evidence for higher reliability and clinical utility of ICD-11 compared with ICD-10. Despite the extensive revision process, changes from ICD-10 to ICD-11 were relatively modest in that both systems are categorical, classifying mental phenomena based on self-reported or clinically observable symptoms. Other recent approaches to psychiatric nosology and classification (eg, neurobiology-based or hierarchical) are discussed. To meet the needs of different user groups, we propose expanding the stepwise approach to diagnosis introduced for some diagnostic categories in ICD-11, which includes categorical and dimensional elements.


Este artículo entrega una breve descripción de los cambios de la CIE-10 a la CIE-11 con respecto a la clasificación de los trastornos mentales, conductuales o trastornos del neurodesarrollo. Estos cambios incluyen una nueva estructura del capítulo, nuevas categorías diagnósticas, cambios en los criterios diagnósticos y pasos hacia un enfoque dimensional. Además, se revisan los estudios de campo de evaluación para la CIE-11, que proporcionan evidencia preliminar de una mayor confiabilidad y utilidad clínica de la CIE-11 en comparación con la CIE-10. A pesar del extenso proceso de revisión, los cambios de la CIE-10 a la CIE-11 fueron relativamente pocos en el sentido de que ambos sistemas son categoriales y clasifican los fenómenos mentales en base a síntomas auto-reportados o que sean clínicamente observables. Se discuten otros enfoques recientes de la nosología y de la clasificación psiquiátrica (por ejemplo, basados en la neurobiología o de acuerdo a jerarquías). Para satisfacer las necesidades de diferentes grupos de usuarios, se propone expandir el enfoque gradual del diagnóstico introducido para algunas categorías diagnósticas de la CIE-11, que incluye elementos categoriales y dimensionales.

Cet article propose un aperçu des évolutions entre la CIM-10 et la CIM-11 concernant la classification des troubles mentaux, comportementaux ou neurodéveloppementaux. Un nouveau chapitre, de nouvelles catégories diagnostiques, des critères diagnostiques modifiés et des étapes dimensionnelles ont été ajoutés. De plus, nous examinons les données préliminaires issues d’études de terrain d’évaluation de la CIM-11, en faveur d’une plus grande fiabilité et utilité de cette dernière comparée à la CIM-10. Les modifications de la CIM-10 vers la CIM-11 sont relativement modestes malgré une révision étendue, les deux systèmes restant catégoriels et les troubles mentaux étant classés d’après des symptômes auto-rapportés ou cliniquement observables. Nous analysons d’autres approches récentes de la nosologie et de la classification psychiatriques (selon la neurobiologie ou hiérarchiquement par exemple). Certaines catégories diagnostiques de la CIM-11 pourraient bénéficier selon nous de cette méthode progressive en incluant des éléments catégoriels et dimensionnels.

Keywords:

• International Classification of Diseases
• diagnosis
• mental
• behavioral or neurodevelopmental
disorder
• diagnostic guideline

Introduction


The development of the Mental, Behavioral or Neurodevelopmental Disorders (MBND) chapter of the ICD-11 was the largest and most participative process in the history of mental health disorder classification. The three major aims for this process were global applicability, scientific validity, and clinical utility. ^1,2 In 2007, the WHO Department of Mental Health and Substance Abuse assigned the International Advisory Group for the Revision of the ICD-10 Mental and Behavioural Disorders. ³ This advisory group, together with the WHO, established working groups in which experts from all continents reviewed the available evidence and proposed changes to specific parts of the ICD-10 Mental and Behavioural Disorders chapter. These proposals were discussed in a collaborative process with various stakeholders (eg, mental health professionals and users of mental health services), resulting in a beta-draft of the ICD-11 MBND chapter. From 2015, the WHO made the ICD-11 MBND beta draft publicly available on the internet for review and comments. ⁴ Additionally, feedback from mental health practitioners was obtained via formative field studies. ^5,6 In May 2019, the 72nd World Health Assembly voted to adopt ICD-11 , which will be implemented by the WHO member states from January 1, 2022.


In this article, we first present a brief summary of changes regarding the classification of mental, behavioral, or neurodevelopmental disorders from ICD-10 to ICD-11 . In this summary, we review, with examples, changes in the chapter structure, new diagnostic categories, changes in diagnostic criteria and dimensional approaches in ICD-11 . Second, we review findings from a series of field studies evaluating how well the ICD-11 functions when applied by health professionals. Third, we discuss new approaches in psychiatric nosology and we propose expanding dimensional additions to categorical diagnoses to a broader range of diagnostic categories in ICD-11 .


Changes from ICD-10 
to IICD-11

Chapter structure


The ICD-11 MBND chapter contains 21 disorder groupings compared with 11 disorder groupings in ICD-10 . Table I displays an overview of the disorder groupings in ICD-10 and ICD-11 . Sleep-wake disorders and conditions related to sexual health were separated from the ICD-11 MBND chapter and cross-listed from the new sleep-wake disorders and conditions related to sexual health chapters. Principles for ordering disorder groupings in ICD-11 were shared etiology, pathophysiology, and phenomenology. Additionally, the aim of the WHO and American Psychiatric Association to harmonize the structure of ICD-11 and DSM-5 influenced the chapter structure of ICD-11 . ² A central difference between ICD-11 and ICD-10 regarding chapter structure is the omission of a separate disorder grouping for mental and behavioral disorders with onset during childhood and adolescence. The disorders previously pooled in this grouping were moved to other disorder groupings in the ICD-11 MBND chapter, highlighting developmental continuity across the lifespan. ¹ 


New diagnostic categories in ICD-11 and changes 
in diagnostic criteria


Several diagnostic categories were added in ICD-11 . Table II displays brief descriptions of these new diagnostic categories. The introduction of some new diagnostic categories in ICD-11 has been controversially discussed. ^7,8,9 For instance, there were concerns over the pathologization of grief, computer gaming, and compulsive sexual behavior.


In addition to the introduction of new diagnostic categories, there were also changes in the diagnostic criteria for previously existing diagnoses. For example, the diagnostic threshold for Post-Traumatic Stress Disorder (PTSD) was raised in ICD-11 by defining three core symptoms that should be present in all cases: re-experiencing the traumatic event as vivid intrusive memories, flashbacks, or nightmares; avoidance of thoughts and memories of the event, situations or people reminiscent of the event; persistent perceptions of heightened current threat. There is some evidence indicating that the prevalence of ICD-11 PTSD is lower than the prevalence of ICD-10 PTSD, ^10,11 whereby the ICD-11 criteria seem to identify the more severe cases of PTSD. ¹² Regarding the prevalence of new diagnostic categories, preliminary evidence suggests that the prevalence of the ICD-11 Prolonged Grief Disorder might be almost three-fold higher than the prevalence of DSM-5 Persistent Complex Bereavement Disorder (18.0% compared with 6.4%). ¹³ In sum, it is unclear how the introduction of ICD-11 will influence the prevalence rate of mental disorders as a whole. To prevent pathologization of normal behavior, the ICD-11 Clinical Descriptions and Diagnostic Guidelines (CDDG), which describe the main clinical features for each disorder, focus on defining the boundary between disorders and variation of normal human functioning.


Dimensional approaches in a categorical system


Current classification systems of mental disorders are based on a polythetic categorical approach. In these classification systems, a list of characteristic symptoms is provided for each diagnosis. The presence of a, usually predefined, number of symptoms from this list is sufficient to assign the respective categorical diagnosis. ¹⁴ Categorical diagnoses are required to justify treatment in most countries, to communicate efficiently about mental disorders, and to collect epidemiological data. Additionally, a categorical diagnosis may aid in the decision whether to treat or not to treat a patient. ¹⁵ However, categorical classification of mental disorders 
is associated with various limitations including large 
within-category heterogeneity, comorbidity, and difficulties in representing subthreshold symptomatology. ¹⁶ 


Table I. Disorder groupings in the ICD-11 Mental, Behavioural or Neurodevelopmental Disorders chapter and in the ICD-10 Mental and Behavioural Disorders chapter (and relevant disorder groupings from other ICD-11 chapters).

Table II. Overview of new diagnostic categories in the Mental, Behavioural or Neurodevelopmental Disorders chapter in ICD-11. PTSD, post-traumatic stress disorder.

In a dimensional approach, the severity of a symptom or the degree of disturbance of a specific psychological function is rated on a quantitative dimension. There is a growing understanding that psychopathology is continuously graded in severity. ^17,18 Dimensional approaches represent the severity of specific symptoms and psychological dysfunctions, including subthreshold symptomatology. A disadvantage of dimensional classification (eg, in the form of diagnostic profiles), however, is its increased complexity and, therefore, reduced clinical utility compared with categorical classification.


For ICD-11 , the categorical approach of ICD-10 was largely maintained. Yet, dimensional expansions regarding severity, course, and specific symptoms were added for some diagnoses. These dimensional expansions of categorical diagnoses mirror clinical practice, in which dimensional information (eg, severity of illness) is regularly taken into consideration for selecting treatments. ¹⁹ A large shift towards dimensionality concerned personality disorders. ²⁰ The division of personality disorders into discrete categories in ICD-10 is not empirically based. ²¹ Among other problems, a large proportion of patients simultaneously fulfilled the criteria for multiple personality disorders. ^22,23 Against this background, the different personality disorders in ICD-10 were replaced with a single personality disorder diagnosis in ICD-11 which is characterized by problems in functioning of aspects of the self (eg, identity) and/or interpersonal dysfunction (eg, managing conflict in relationships). The ICD-11 personality disorder diagnosis is further differentiated according to severity into mild, moderate, and severe. The diagnosis may optionally be specified by the presence of one or multiple maladaptive personality traits: Negative affectivity, detachment, dissociality, disinhibition, anankastia and Borderline pattern. Whereas a different, more complex, dimensional approach to personality disorders was deemed as not feasible in the development of DSM-5 , ^24,25 there was a strong focus on clinical utility and simplicity in the revision of the personality disorders grouping in ICD-11 .


Another shift towards dimensionality concerned depressive episodes. In ICD-11 , depressive episodes in depressive or bipolar disorders may be described in detail by using qualifiers indicating the presence of specific symptoms: the melancholic features qualifier, the anxiety symptoms qualifier; the panic attacks qualifiers, and the seasonal pattern qualifier. Additionally, depressive episodes can be described according to severity (mild, moderate, or severe) and remission status (in partial or in full remission). For moderate and severe depressive episodes, the presence of psychotic symptoms may also be indicated.


Also for the Schizophrenia or Other Primary Psychotic Disorders grouping in ICD-11 , dimensional symptom specifiers and course specifiers were added. ^1,26 Symptom specifiers describe the current severity of symptoms in six domains: positive symptoms, negative symptoms, depressive symptoms, manic symptoms, psychomotor symptoms, and cognitive symptoms. The severity of each of these symptoms is rated on a 4-point scale ranging from “not present” to “present and severe.” These symptom qualifiers may be used for any diagnosis from the Schizophrenia or Other Primary Psychotic Disorders grouping. Thus, mental health professionals may compliment categorical diagnoses from this disorder grouping by a profile of specific symptoms that conveys additional information regarding symptomatology. The course qualifiers for the Schizophrenia or Other Primary Psychotic Disorders grouping contain two components, allowing characterization of the longitudinal course. The first component (episodicity) differentiates between first episode, multiple episodes or continuous course. The second component concerns the cross-sectional evaluation of the acuity of the symptoms and allows differentiating the current clinical status: currently symptomatic, partial remission, full remission.


A review of the ICD-11 evaluative field studies


A series of field studies evaluated how well the ICD-11 CDDG function when applied by health professionals. These evaluative field studies were conducted either with real patients (ie, ecological field studies) or online with prototypical patient descriptions (ie, online vignette-based field studies). ²⁷ A large ecological field study of the ICD-11 MBND chapter examined the reliability and clinical utility of 16 ICD-11 diagnoses in a sample of 339 clinicians from 13 countries. ^28,29 When the ICD-11 diagnostic guidelines were applied to 1806 patients, interrater reliability was excellent for some diagnoses (eg, for social anxiety disorder), but improvable for others (eg, for dysthymic disorder). On average, the reliability of the ICD-11 CDDG was higher compared with previously reported estimates of the ICD-10 CDDG. ²⁸ Additionally, clinicians’ evaluations of clinical utility were positive: A large majority of clinicians (82.5% to 83.9%) perceived the ICD-11 CDDG as quite or extremely easy to use, accurate, clear, and understandable. However, utility ratings varied between countries. ²⁹ In a different ecological field study with 23 practitioners from Mexico, interrater reliability was high for psychotic disorders, moderate for stress-related and mood disorders, and small for anxiety and fear-related disorders. ³⁰ 


A comprehensive online vignette-based field study investigated the diagnostic accuracy and clinical utility of the ICD-11 CDDG compared with the ICD-10 CDDG in a sample of 928 clinicians from all WHO regions. ³¹ Diagnostic accuracy, time required to come to a diagnosis, and perceived clinical utility (ie, ease of use, goodness of fit, clarity) were more favorable for ICD-11 compared with ICD-10 . However, advantages of the ICD-11 over the ICD-10 were largely limited to new diagnostic categories in ICD-11 . After excluding all vignettes that pertained to new diagnostic categories in ICD-11 , there was no significant difference in diagnostic accuracy, goodness of fit, clarity, or time required for diagnosis, but the perceived ease of use was significantly higher for ICD-11 compared with ICD-10 . For feeding and eating disorders, a vignette-based online field study with 2288 practitioners found higher diagnostic accuracy and perceived clinical utility of ICD-11 compared with ICD-10 . ³² Also for Schizoaffective Disorder, a vignette-based online field study with 873 practitioners showed small improvements in diagnostic accuracy using ICD-11 compared with ICD-10 . ³³ A different online vignette-based field study with 1738 practitioners from 76 countries revealed a higher diagnostic accuracy of practitioners diagnosing based on ICD-11 compared with practitioners diagnosing based on ICD-10 for disorders specifically associated with stress. ³⁴ Additionally, in a web-based field study, a sample of 163 mental health professionals rated the ICD-11 classification of personality disorders (including three levels of severity and trait qualifiers) as more useful regarding its utility for treatment planning, communicating with patients, comprehensiveness, and ease of use compared with the ICD-10 classification of personality disorders. ³⁵ 


In sum, the results from the evaluative field studies paint a positive picture of the ICD-11 MBND chapter. However, there are different limitations of evaluative field studies that make overly enthusiastic appraisals of ICD-11 premature. First, the samples could be biased in such a way that practitioners who are positive towards ICD-11 are more likely to participate in ICD-11 field studies. This could be particularly the case for online field studies for which participants had to register on their own initiative. Second, individuals’ knowledge that they participate in a field study modifies their behavior. ³⁶ Thus, behavior in ICD-11 evaluative field studies might not adequately reflect diagnostic decision making in routine care. Third, there is some concern over the artificiality of vignette studies. Because vignettes describe prototypic cases, they might not accurately reflect the complexity of real-life situations. ³⁷ In summary, whereas the field studies give first indications regarding diagnostic accuracy and clinical utility, further ecological field studies are needed to reveal how well the ICD-11 works when applied by clinical practitioners under regular conditions.


Critical evaluation and future directions


In the development of the ICD-11 MBND chapter, important steps have been taken to ensure clinical utility, global applicability, and scientific validity. There were also notable steps towards dimensionality regarding symptom severity and time course. Yet, one might argue that changes from ICD-10 to ICD-11 were relatively modest in that both systems are categorical, classifying mental phenomena based on self-
reported or clinically observable symptoms. In this paragraph, we discuss different new approaches to psychiatric classification and nosology that might inform future revisions of the ICD.


New approaches in diagnostic classification


Various different approaches to advance psychiatric nosology have been introduced over the last years. Of these approaches, the National Institute of Mental Health’s Research Domain Criteria (RDoC) project ³⁸ has received the most attention. RDoC is a research framework for the investigation of mental disorders that is not intended for immediate practical clinical use. The aim of RDoC is to provide a biologically informed framework for understanding mental disorders. The RDoC matrix distinguishes six domains of functioning (negative valence systems, positive valence systems, cognitive systems, social processes, arousal and regulatory systems, and sensorimotor systems) with various subconstructs and eight units of analysis: genes, molecules, cells, circuits, physiology, behavior, self-report, and paradigms. Varying degrees of functioning and dysfunctions in general psychological and biological systems may be described within this matrix. However, there is one major limitation: The RDoC matrix is too complex to guide diagnosis in clinical practice.


Neither the structure of the ICD-11 MBND chapter nor the structure of DSM-5 are based on neurobiology. Because of the large degree of biological heterogeneity within diagnostic categories of current classification ³⁹ and difficulties distinguishing some diagnostic categories genetically ⁴⁰ and neurobiologically, ⁴¹ different approaches have been proposed to shift diagnostic boundaries in a way that biologically more homogeneous subgroups are formed. One such approach is “reverse nosology,” which suggests redefining diagnostic categories based on their molecular, cellular, and circuit basis. ⁴² In this approach, patients that display a similar neurobiology (eg, similar brain activation patterns) are grouped in the same diagnostic category, although the self-reported symptoms or observable psychopathology may be fundamentally different. Thus, clinical practitioners would be no longer able to diagnose based on their clinical impression and self-report. Additionally, there would be large difficulties in communicating about a diagnosis because it might contain information regarding neurobiology, but only little information regarding observable psychopathology.


A different group of approaches aims to form biologically more homogeneous subgroups within existing diagnostic categories. For example, the Systems Neuroscience of Psychosis (SyNoPsis) project ⁴³ aims to link clinical manifestations of Schizophrenia onto specific brain systems. SyNoPsis differentiates three behavioral domains of Schizophrenia symptoms that match the function of three higher-order corticobasal brain systems: Language (associative loop), affect (limbic loop), and motor behavior (motor loop). Within the SyNoPsis project, also a psychometric instrument that assesses symptoms from these three behavioral domains has been developed which is used to identify clinically and neurobiologically homogeneous subgroups of schizophrenia patients (Bern Psychopathology Scale ⁴⁴ ). Biologically defining subgroups of patients might then improve care by tailored treatment selection and earlier detection. However, thus far, the connection between neurobiology and psychopathology is not sufficiently understood to establish a diagnostic system on it.


A third approach to psychiatric nosology emphasizes the 
hierarchical structure of psychopathology. For example, the Hierarchical Taxonomy of Psychopathology (HiTOP ⁴⁵ ) suggests that arbitrary boundaries between diagnostic categories limit the reliability and validity of traditional taxonomies. This taxonomy is based on dimensional assessments of psychopathology and differentiates different levels of psychopathology with specific symptoms (eg, appetite loss) at the bottom and broader spectra or super-spectra as broader constellations of syndromes (eg, internalizing and externalizing spectra) at the top of the hierarchy. Factor analytic evidence also suggests the presence of a general psychopathology factor that explains the co-occurrence of symptoms across various disorders. ^46,47 This general psychopathology factor describes individuals’ propensity to develop any form of psychopathology and is related to increased life impairment. ⁴⁶ For clinical practice, however, scores on higher order psychopathology dimensions are difficult to interpret leading to a low clinical utility of hierarchical approaches. Yet, dimensional information regarding specific aspects of psychological dysfunctions might aid in guiding interventions.


Reconciling the needs of different user groups: 
a stepwise approach


A potential problem of current categorical classification systems is that they aim to serve many purposes for various different groups of users. For example, primary care practitioners need well communicable, comprehensible diagnostic categories. Researchers, on the other hand, often prefer detailed dimensional assessments. ¹⁵ Whereas complex approaches like RDoC are suitable for research contexts, the categorical approach in ICD-11 provides a higher clinical utility.


To ensure that future versions of the ICD meet the needs of different user groups, a stepwise procedure to diagnosis might be appropriate. In this stepwise approach, each diagnostic step describes a patient’s psychopathology with increasing detail. In the first diagnostic step, a patient’s symptoms may be categorized into broad diagnostic categories. Regarding level of detail, this step might be similar to the ICD-10 Primary Care Version for Recognition and Management of Mental Disorders. On this diagnostic level, patients that experience a level of distress requiring specialized treatment and further diagnostics may be identified. In the second diagnostic step, more specific differential diagnosis might be made. For practitioners in specialized mental health facilities and ambulatory care, the ICD-11 CDDG provide the optimal level of detail. The CDDG contain detailed descriptions regarding the core symptoms of disorders, differential diagnosis, and boundaries with normal human functioning.


In specialized treatment settings and for research, additional dimensional assessments are required to more precisely describe psychopathology. Thus, a third diagnostic step might enrich categorical diagnoses with dimensional assessments, combining the advantages of both approaches (see the ICD-11 Schizophrenia or Other Primary Psychotic Disorders grouping). In this diagnostic step, each categorical diagnosis could be complemented with a symptom profile that provides specific information regarding domains of psychological malfunctioning. Based on this stepwise approach, rapid communication will be possible based on diagnostic categories and dimensional assessments will provide more nuanced profiles for contexts in which detailed dimensional information is needed beyond the overall degree of severity to inform treatment (eg, psychotherapy) and for research. ¹⁷ Importantly, the use of nuanced and partly dimensional descriptions of psychopathology is not new to psychiatric treatment: There are various dimensional psychometric scales used in psychiatric hospitals (eg, Beck Depression Inventory-II ⁴⁸ ; Positive and Negative Symptoms Scale ⁴⁹ ) and doctor’s letters frequently communicate differentiated clinical assessments.


Thus far, steps towards enriching diagnostic categories with symptom profiles was limited to some disorder groupings in ICD-11 (eg, Personality disorders and related traits, mood disorders, schizophrenia or other primary psychotic disorders). Yet, there is large potential for enriching further categorical diagnoses with symptom profiles. For example, it has been suggested to assess all symptoms of substance use disorders in DSM-5 on (at least) a 3-point scale. ⁵⁰ 


Summary


The development of the ICD-11 MBND chapter was characterized by a focus on clinical utility, global applicability, and scientific validity. Thus far, mental health professionals’ evaluations of the ICD-11 are relatively positive. Changes from ICD-10 to ICD-11 include the introduction of new diagnoses, the refinement of diagnostic criteria of existing diagnoses, and notable steps in the direction of dimensionality for some diagnoses. However, there was no paradigm shift from ICD-10 to ICD-11 . There are promising new approaches to psychiatric nosology, which, however, have a low clinical utility. We argue in favor of a stepwise approach to diagnosis that retains categorical classification to ensure clinical utility, ⁵¹ but allows more detailed dimensional assessments of psychopathology to inform treatment in specialized settings and research. Expanding the stepwise approach to diagnosis introduced for some diagnostic categories in ICD-11 may help to meet the needs of different user groups of the ICD.


The authors declare that they have no conflict of interest