LEARNING OBJECTIVES
After reading this article you will be able to:
-
understand how the DSM and ICD can be misused by expert psychiatric witnesses
-
know how misunderstandings about, and misuse of, the ICD and DSM by psychiatrists have created difficulties for courts and tribunals
-
know how to use and describe the DSM and ICD so as to avoid misunderstanding in courts and tribunals and miscarriages of justice.
‘Statistics will be most valuable if they do not attempt to solve all the problems of administration and psychiatry and sociology under one confused effort of a one-word diagnosis marking the individual.’
(Adolf Meyer on his resignation from the American Psychiatric Association’s Statistical Manual Committee, cited in Grob Reference Grob1985)
‘It is the diagnosing health professional who is responsible for developing a diagnostic formulation appropriate for an individual patient, considering the patient’s individual, social and cultural context.’
(World Health Organization 2024: p. 3)
This article considers further the use of the International Classification of Diseases (ICD) and the Diagnostic and Statistical Manual of Mental Disorders (DSM) in medico-legal reporting. The nature and purposes of the ICD and DSM classifications, their use in legal proceedings, their limitations and how they can be exposed in the testing of expert psychiatric evidence have been described in a companion article in this issue (Rix Reference Rix2025a). A complementary article for the judiciary has been published (Rix Reference Rix2025b).
Users of the World Health Organization’s clinical version of ICD-11, Clinical Descriptions and Diagnostic Requirements for ICD-11 Mental, Behavioural and Neurodevelopmental Disorders (World Health Organization 2024: p. 3), are advised that it represents general guidelines rather than strict requirements and that they should use their own clinical judgement. Users of the American Psychiatric Association’s DSM-5 (American Psychiatric Association 2013: p. 21) are advised that it is a practical, functional and flexible guide for organising information and that diagnostic criteria are offered as guidelines for making diagnoses, and their use should be informed by clinical judgement. DSM-5 (p. 25) also carries a warning that when used for forensic purposes there is a risk that diagnostic information will be misused or misunderstood.
This article describes such misuse and misunderstanding of ICD-11, DSM-5 and their predecessors by expert witnesses, along with misunderstandings of the nature, purpose and status of these classifications. It sets out suggestions as to the information about the ICD and DSM that may assist courts and tribunals and help to avoid misunderstandings and misuse. It concludes with advice to experts on the use of the DSM and ICD.
Psychiatric witnesses’ perceptions and use of the ICD and DSM
Perceptions and misperceptions
Box 1 sets out some of what can be regarded as questionable or potentially misleading perceptions of the ICD and DSM. Box 2 sets out what are probably correct perceptions.
Questionable or potentially misleading perceptions of the DSM and ICD by expert witnesses
‘DSM-IV […] was the considered the bench-mark standard in medical research and medico legal reporting’
(Blackledge v London General Transport Services Ltd [2001])
‘ICD 10 was not the manual of choice for listing accepted psychiatric diagnosis in the UK’
(Kumar v General Medical Council [2012])
‘I’ll tell you one thing that won’t be in DSM-5 and that’s schizoaffective psychosis – that’s going’
(R v Edgington, Central Criminal Court, unreported, 2013)
‘The DSM classification had been legally recognised in Ireland’
(DJ v The Minister for Health [2017])
‘The ICD tended to be used by psychiatrists to diagnose mental illness, whereas the DSM tended to be used by psychologists to diagnose mental health difficulties’
(Graham v Mer Majesty’s Advocate [2018])
‘It [the ICD] has become the international standard diagnostic classification for most epidemiological purposes’
(HSB v Secretary of State for the Home Department [2021])
Correct perceptions of the DSM and ICD by expert witnesses
‘The purpose of [ICD-10 and DSM-IV] is to aid communication between clinicians’
(McDonald or Cross v Highlands & Islands Enterprise [2000])
‘ICD-10 is the classification recognised by the government within the NHS’
(Blackledge v London General Transport Services Ltd [2001])
‘ICD-10 is not a textbook of psychiatry but is designed to offer a world-wide guide to psychiatric symptoms for use in judging and comparing morbidity in different countries and to ensure consistency in operational definitions used in research programmes’
(Brennan v Bedford Borough Council [2002])
‘It was a Scottish Government Requirement that all diagnoses should be made using ICD-10, which was prepared by the World Health Organisation, rather than the American Psychiatric Association’s DSM-V’
(Graham v Her Majesty’s Advocate [2018])
‘[The expert] was at pains to say that the diagnostic “boxes” describing CPTSD, PTSD and various personality disorders were not discrete and that there was overlap [and that] one could not take a legalistic view of the descriptions given in the documents’
(London Borough of Haringey v FZO [2020])
‘The application of criteria such as these cannot involve a checklist but requires a clinical judgment in the light of all the available evidence’
(Modi v Government of India [2022])
Correcting the misunderstandings
How the DSM and ICD are used
That the DSM is in general medical use, even in the USA, is called somewhat into question by the fact that practising clinicians seldom adhere strictly to the specific diagnostic criteria (Clark Reference Clark, Cuthbert and Lewis-Fernández2017). This includes clinicians whose diagnoses in patients’ records are utilised for research. Its general use in the USA may be the truth, but the whole truth is that only seldom is there strict adherence to the DSM’s specific diagnostic criteria. Experts who apply the criteria strictly in court cases and tribunals should not imply that this is how the criteria are applied in clinical practice or even in some of the research for which it is sometimes asserted that the DSM is superior to, or more widely used than, the ICD.
In HSB v Secretary of State for the Home Department [2021], the expert said that ICD-10 (World Health Organization 1992) had become the international standard diagnostic classification for most epidemiological purposes. It is true that a substantial majority of psychiatrists outside the USA mostly use the ICD in daily clinical practice (Reed Reference Reed, Correira Mendonça and Esparza2011), but the DSM is used more often in research around the world (Clark Reference Clark, Cuthbert and Lewis-Fernández2017).
Two immigration and asylum cases suggest some misunderstanding either by experts or the tribunal as to the use of diagnostic classifications. In Secretary of State for the Home Department v AG [2021], there is a reference to how the expert ‘undertook a diagnostic assessment of PTSD [post-traumatic stress disorder] in accordance with the recognised checklist “DSM-5” and assessed the symptoms against the criteria’. DSM-5 is not, and should not be used as, a ‘checklist’. In SA v Secretary of State for the Home Department [2021], the expert ‘administered the relevant diagnostic tests in accordance with DSM-V’, but DSM-5 does not require tests to be performed to make a diagnosis.
ICD’s R and Z codes and DSM’s V codes
Probably the most critical and informed consideration of the ICD and DSM in criminal proceedings is in R v Dowds [2012] (Box 3), where Lord Justice Hughes began by quoting from the introductions to the ICD and DSM, including the cautionary section in the latter. He then applied the concept of the ‘imperfect fit’ between questions of ultimate concern to the law and the information contained in a clinical diagnosis to the case where the issue was whether, on an indictment for murder, acute voluntary intoxication was capable of giving rise to the partial defence of diminished responsibility on the basis that it was a ‘recognised medical condition’.
The imperfect fit between psychiatric and legal concepts
Stephen Dowds killed his partner by inflicting approximately 60 knife wounds on her. There was every likelihood that the killing had occurred when he was to an extent intoxicated. At the outset of his trial the judge ruled that as a matter of law simple voluntary and temporary drunkenness was not capable of founding the partial defence of diminished responsibility manslaughter. In consequence, diminished responsibility was not raised before the jury.
Dowds’ appeal challenged that ruling on the basis that ICD-10 contains the condition of ‘acute intoxication’ (at F10.0), defined as ‘a condition that follows the administration of a psychoactive substance resulting in disturbances in level of consciousness, cognition, perception, affect or behaviour, or other psychophysiological functions or responses’ (similar to DSM-IV’s ‘alcohol intoxication’) and that acute intoxication is therefore a ‘recognised medical condition’ under the Homicide Act 1957, section 2 (as amended).
Lord Justice Hughes said that the deceptively simple argument for the appellant bypassed the very clear general rule of law preventing a defendant from relying on his voluntary intoxication that is well entrenched and formed the unspoken backdrop for the amendment to the Homicide Act. There had been no hint of any dissatisfaction with that rule of law. If Parliament had meant to alter it, or to depart from it, it would undoubtedly have made its intention explicit. Such an intention could not be inferred from the adoption in the new formulation of the expression ‘recognised medical condition’ because the origins of that were clearly explained by the Law Commission. They explicitly did not include writing the terms of ICD-10 and/or DSM-IV into the legislation, for which purpose, he said, those terms are demonstrably unsuited. Referring to the warning in DSM-IV about the risk of an ‘imperfect fit’ he said:
‘The particular “imperfect fit” there under consideration is the divergence between the level of impairment which may bring a patient within a DSM-IV classification and the level necessary to have legal impact. But exactly the same considerations apply when the question is whether the doctors’ classification system addresses the legal issue in any particular case. There will inevitably be considerations of legal policy which are irrelevant to the business of medical description, classification, and statistical analysis.
The “imperfect fit” to which the authors of DSM-IV refer is nowhere more clearly demonstrated than in the breadth and kind of conditions included in both ICD-10 and DSM-IV. ICD-10 includes, for example, “unhappiness” (R45.2), “irritability and anger” (R45.4), “suspiciousness and marked evasiveness” (R46.5), “pyromania” (F63.1), “paedophilia” (F65.4), “sado-masochism” (F65.5) and “kleptomania” (F63.2). DSM-IV includes similar conditions and also such as “exhibitionism” (569), “sexual sadism” (573) and “intermittent explosive disorder” (663/667) […] Not all of these are treated by the classification systems as mental disorders, but all are, doubtless, “recognised medical conditions” in the sense that they are perfectly sensibly included in guides for description of patients by doctors. It follows that a great many conditions thus included for medical purposes raise important additional legal questions when one is seeking to invoke them in a forensic context. “Intermittent explosive disorder”, for example, may well be a medically useful description of something which underlies the vast majority of violent offending, but any suggestion that it could give rise to a defence, whether because it amounted to an impairment of mental functioning or otherwise, would, to say the least, demand extremely careful attention. In other words, the medical classification begs the question whether the condition is simply a description of (often criminal) behaviour, or is capable of forming a defence to an allegation of such.’
The appeal was dismissed.
(R v Dowds [2012])
Many psychiatrists would take issue with his statement that ‘unhappiness’, ‘irritability and anger’ and ‘suspiciousness and marked evasiveness’ are ‘doubtless, “recognised medical conditions”’ even if they are ‘perfectly sensibly included in guides for the description of patients by doctors’, but his statement is hardly surprising. These R codes are listed in the Annex ‘Other conditions from ICD-10 often associated with mental or behavioural disorders’, but that does not make them ‘doubtless, “recognised medical conditions”’. They may often, but not always, be associated with mental or behavioural disorders; they are not in themselves mental or behavioural disorders. The R codes are for ‘Symptoms, signs and abnormal clinical and laboratory findings, not elsewhere classified’. Indeed, that they are used when it has not been possible to make a diagnosis of a (recognised) medical condition surely means that there is no recognised medical condition. But the importance of this judgment is not so much that it is a further example of how psychiatric experts can confuse the courts with their reliance on the ICD or DSM (which uses V codes for similar purposes to the ICD’s use of R and Z codes), but the fact that it calls into question how diagnosis assists as to questions of ultimate concern to the law. It also makes clear that medical description, classification and statistical analysis can be irrelevant to considerations of legal policy.
Misuse of the Z codes from ICD-10’s Annex occurred in R (on the application of Michaelides) v Police Medical Appeal Board [2019] (Box 4). Notwithstanding the potentially misleading use of the term ‘condition’ in the title of the Annex to ICD-10, the subheading for the Z codes is ‘Factors influencing health status and contact with health services’ and Z56 is for ‘Problems related to employment and unemployment’. Such ‘factors’ or ‘problems’ are not ‘diagnoses’, arguably not conditions and certainly not disorders.
The misuse of the ICD’s Z codes
Mr Michaelides was a retired police officer. His application for a police injury pension was refused. Entitlement is dependent on proof that the applicant has ceased to be a member of the police force and is permanently disabled as a result of an injury received in the execution of their duty.
He sought a judicial review of the lawfulness of a decision taken by the Police Medical Appeal Board (PMAB), which refused his application. When a force medical officer assessed him as unfit to work, he was referred to a consultant psychiatrist. The psychiatrist’s report gave the ‘Diagnosis’ [sic] as:
‘1. F32.1 of ICD-10 Moderate Depressive Episode with Somatic Syndrome
2. Z56.0 of ICD-10 Problems with employment (perceived racial discrimination)’.
These ‘diagnoses’ were adopted by an occupational physician as the basis for his opinion that the claimant was ‘medically unfit for performing the ordinary duties of a member of the police force … in respect of the following condition(s) Depressive Episode (ICD-10 F32 1) Problems with employment (ICD-10 Z56.0)’. He stated that the depressive episode and problems with employment were conditions [sic] that were likely to be permanent. The claimant was retired and applied to the Chief Constable for an injury award. His application was refused and hence his appeal to the PMAB.
As to the Z56.0 ‘diagnosis’, the PMAB said that this was not a medical condition but a descriptive factor relating to the circumstances of a medical condition, and therefore could not lead to permanent disablement. It said that the diagnosis was not binding.
At the judicial review, counsel for the force solicitor submitted that, as to the PMAB’s observations on code Z56.0, a reading of ICD-10 showed it was correct in stating that it does not amount to a medical condition. The court accepted this, stating that it was open to the PMAB to treat the reference to Z56.0 as it did. It is not a diagnosis of a medical condition and there was no requirement for it to treat it as such. The court said that it is to be remembered that it is a code which cross-references to the ICD-10.
(R (on the application of Michaelides) v Police Medical Appeal Board [2019])
Some consequences of misunderstandings and misuse
Post-traumatic stress disorder problems
Over-reliance on PTSD’s diagnostic criteria has created particular problems for the courts. One problem relates to the nature and severity of the causal trauma or the ‘gateway’ requirement in the ICD and DSM. ICD-11 allows for a broader range of traumatic events than DSM-5, but for ICD-11 the trauma has to be directly experienced.
Table 1 summarises how courts have had to resolve issues when the ICD and DSM have been pitted against each other, when the issue has been DSM-IV and DSM-5 or DSM-5 and ICD-11 differences as to the means by which the trauma is perceived, and when the nature and severity of the trauma have been in issue.
When the diagnosis is not in the classification
The London Borough of Haringey appealed against a High Court award of compensation to FZO, a man who had been sexually abused in the council’s care.
At the trial, there were two difficulties. First, there was the fact that, at that time, ICD-11 was not yet in use in the World Health Organization member states. ICD-11 was published in 2018, adopted in 2019 and only became effective in member states in 2022. Second, there was an issue as to whether the respondent, who had been indecently assaulted and buggered on multiple occasions between the ages of 13 and 15, was suffering from CPTSD as diagnosed by his expert witness. The appellant’s expert noted that ICD-11 remained ‘unpublished’ [sic]. His view was that CPTSD did not appear in standard diagnostic texts, and it was an inapposite diagnosis because the abuse was not perceived as ‘traumatic’ by the respondent, either at the time or for years afterwards.
Regarding diagnosis, the judge had preferred the opinion of FZO’s expert that he suffered from CPTSD. She found that what occurred fitted the wording of the descriptors in ICD-11, which in one of its forms was ‘designed to cover precisely the circumstances where the events evolve over time rather than a sudden dramatic trauma which is adequately and properly covered by PTSD’. She rejected a submission for the Haringey Council that the abuse had to be violent to qualify.
In dismissing the Council’s appeal, Lord Justice McCombe considered that the court was entitled to prefer the evidence of FZO’s expert to that of the Council’s expert. Lord Justice Simon Davies and Lady Justice Nicola Davies agreed, the latter adding:
‘At the time of the trial, Complex PTSD was not included within the international classification of diseases published by the World Health Organisation. [The expert for the respondent] stated that Complex PTSD is to be accepted as a diagnostic category within the publication in 2020. In giving evidence [the expert for the respondent] did not rely upon any articles in the medical literature which reference Complex PTSD, its symptoms nor the proposed classification. […]
Further, in preferring the evidence of [the expert for the respondent], the judge was accepting a diagnosis which has yet to be formally included within the World Health Organisation international classification. If that is a course which the court is minded to take, then real rigour is required in assessing the relevant evidence in respect of such a diagnosis.’
(London Borough of Haringey v FZO [2020])
Resolution by the courts of issues arising from the diagnostic criteria for post-traumatic stress disorder (PTSD)
These issues are unsurprising. A Dutch population survey (Mol Reference Mol, Arntz and Metsemakers2005) found that many people experience symptoms of PTSD in response to adverse events, such as divorce, and adverse circumstances, such as unemployment, but without triggers that satisfy the gateway criteria in the ICD or DSM. Another study (Larsen Reference Larsen and Pacella2016) found that PTSD symptoms were equally common following events defined as traumatic versus non-traumatic according to DSM-IV. Some experts use the term ‘subthreshold post-traumatic stress disorder’, which is technically correct if by threshold they mean the gateway requirement, but potentially misleading as it may be taken to imply that the condition is less severe than that of someone with PTSD who has satisfied the gateway criterion.
The risk of injustice resulting from over-reliance on the DSM and ICD also arises from the fact that, as suggested by studies of internally displaced Ukrainians (Shevlin Reference Shevlin, Hyland and Vallières2018) and US military veterans (Wisco Reference Wisco, Marx and Miller2017), fewer people are diagnosed with PTSD using ICD-11 than DSM-5, and also from the fact that the ICD and DSM do not identify exactly the same groups (Stein Reference Stein, McLaughlin and Koenen2014).
When the disorder is not in a classification (or so it is thought)
In Shorter v Surrey and Sussex Healthcare NHS Trust [2015] the expert for the claimant acknowledged that ‘abnormal grief’ was not in DSM-IV or ICD-10, but ‘persistent complex bereavement disorder’ was in DSM-5 as a ‘condition for further study’. He said that it was an expression that would be recognised by any psychiatrist. The defendant’s expert said that he did not use the term ‘abnormal grief’ because there was no universally accepted definition of it as a separate psychiatric disorder. The court resolved the issue by accepting a diagnosis of major depressive disorder.
Complex PTSD (CPTSD) has the potential to create particular difficulties: it is in ICD-11 and not in DSM-5. But it was ICD-11’s status in itself, and specifically the status of CPTSD, that were issues in London Borough of Haringey v FZO [2020], an appeal by a borough council against the award of compensation to FZO, who had been sexually abused in the council’s care (Box 5). Unlike COVID-19, which did not exist in 2018, CPTSD did. FZO was suffering from the condition. The fact that the World Health Organization did not officially recognise the diagnosis until May 2018 does not mean that FZO only began to suffer from it in May 2018. The fact that it is not in DSM-5 does not mean that his condition does not exist. It is his condition that matters and not the label. But what this case also illustrates is that if a diagnosis is not in one or another classification, the expert needs to provide the court with appropriate evidence about it from the medical literature.
A similar battle, or perhaps skirmish is a better word, occurred in McClarnon v The Sisters of Nazareth [2024]. At the date of the plaintiff’s expert’s report in November 2015, McClarnon was diagnosed with ICD-10 PTSD and alcohol dependence syndrome. By the time of the trial, ICD-11 had been, in the expert’s words, ‘published’, and he said that the same symptoms would have led him to diagnoses of CPTSD and alcohol use disorder (severe). When the defendant’s expert began to give evidence he was specifically asked about CPTSD. He said it was not accepted in DSM-5 and began to say that ICD-11 was ‘not sufficiently persuasive to…’ when counsel for the plaintiff objected that there had been no challenge to the diagnosis made by the plaintiff’s expert, which counsel for the defendant accepted, so the court did not hear why, in the opinion of the defendant’s expert, he did not find ICD-11 sufficiently persuasive.
Over the years some diagnoses have come and some have gone. In a murder trial, the case of R v Nicola Edgington, Central Criminal Court, unreported 2013, which took place in January and February 2013, the expert’s predicted demise of schizoaffective disorder was mistaken. The issue was the partial defence of diminished responsibility manslaughter. It was the evidence of the expert called by the prosecution that ‘I’ll tell you one thing that won’t be in DSM-5 and that’s schizoaffective psychosis – that’s going’. DSM-5 was published 3 months later, on 18 May 2013. It is true to say that ‘schizoaffective psychosis’ is not in DSM-5, but it was not in DSM-IV. However, schizoaffective disorder is in DSM-5 and, furthermore, the criteria are unchanged from those for schizoaffective disorder in DSM-IV. Obviously, it is not known whether the rejection of this diagnosis by the prosecution expert influenced the jury, but the verdict was one of murder and not diminished responsibility manslaughter. If the trial had taken place 3 months later, or DSM-5 published 3 months earlier, the expert might have been challenged as to this seemingly authoritative statement.
Controversial diagnoses
The condition intermittent explosive disorder (IED) picked out by Lord Justice Hughes in Dowds was an issue around the same time in a murder trial that gave rise to a complaint about an expert to the General Medical Council (Kumar v General Medical Council [2012]. The complaint was that the expert expressed the opinion that the defendant was suffering from the ‘medical condition IED’, which was in DSM-IV, but ‘did not explain that IED was not recognised in the International Classification of Diseases, ICD, and was a controversial diagnosis’.
The problem of drug-induced psychosis
Experts frequently disagree over diagnoses of drug-induced psychosis. Where their opinions, agreements and disagreements are informed by reference to ICD and/or DSM criteria it is usually with little or no regard to the research. Where cannabis or cocaine has been used, an expert may say that an accused suffered a cannabis- or cocaine-induced psychosis on the basis of their psychopathology matching that set out in ICD-11 and DSM-5, but not mentioning that there are no psychopathological differences sufficient to make an absolute distinction between a primary (non-drug-induced) psychotic illness and a cannabis- (Baldacchino Reference Baldacchino, Hughes and Kehoe2012) or cocaine-induced (Vergara-Moragues Reference Vergara-Moragues, Mestre-Pintó and Gómez2016) psychosis.
In R v Foy [2020], an appeal against a murder conviction, an expert who had been instructed pre-trial by the defence expressed the opinion that the appellant had a substance-induced psychotic disorder, saying that the ‘clinical pattern is typical of the paranoid psychosis associated with cocaine’ and that his account, along with evidence from the police camera, was ‘so characteristic of cocaine psychosis that it cannot be ignored’ (emphasis added). As this was a murder case and clearly this evidence would not establish a defence of diminished responsibility, his legal team was not in a position to advance the defence of diminished responsibility manslaaughter at trial. Whether or not the expert relied on ICD or DSM criteria is not clear, but in my experience such opinions are often given relying on one or the other of these.
A recent review of 72 studies carried out since 2000 concluded that ‘distinguishing between substance-induced psychosis, primary psychotic illnesses, and psychotic illnesses with comorbid substance use remains a difficult challenge for clinicians’ and ‘in most cases, chronological criteria are not sufficient to prove a direct causal effect between the substance and psychosis’ (Fiorentini Reference Fiorentini, Cantù and Crisanti2021). However, for a diagnosis of cannabis psychosis ICD-11 refers to whether the psychosis persists for more or less than 1 month and DSM-5 suggests that if symptoms last longer than 1 month a diagnosis other than cannabis-induced psychosis should be considered. Likewise, the review referred to Wilson et al (Reference Wilson, Szigeti and Kearney2018), who summarised results from six studies, assessing the differences, particularly concerning psychopathology, between individuals with a diagnosis of substance-induced psychosis and individuals affected by primary psychosis. The findings did not reveal consistent differences.
Differences between classifications
Cases in which the diagnosis depends on which classification is used have the potential to create difficulties for the courts. Of the 103 disorders appearing in both ICD-11 and DSM-5, there are major differences between 20 (First Reference First, Gaebel and Maj2021). Of these only one, PTSD/CPTSD, appears to have caused significant difficulties, as reflected in reported cases. This does not mean that others may not have caused difficulties in unreported cases. But even minor definitional differences may create difficulties. The potential for difficulties as a result of differences between ICD-11’s ‘substance dependence’ plus ‘harmful pattern of use of substances’ and DSM-5’s ‘substance use disorder’ is suggested by SD v Secretary of State for Work and Pensions (ESA) (Employment and support allowance: Regulation 29) [2016].
In Noble v Owens [2008], the evidence was that the ICD diagnosis was PTSD and the DSM diagnosis was anxiety disorder. The expert was firmly of the view that PTSD was the correct diagnosis. But in the court’s judgment ‘the precise characterisation of Mr Noble’s psychiatric disorder does not signify. What matters are the symptoms of Mr Noble’s condition and the prognosis’.
In DL-H v Devon Partnership NHS Trust v Secretary of State for Justice [2010], the tribunal was initially troubled by the fact that sometimes ICD-10 was used, sometimes DSM-IV. The tribunal observed that they were broadly similar but differed in their detail and gave the example of how one of the criteria for antisocial personality disorder in DSM-IV is that the patient had a conduct disorder before the age of 15, whereas ICD-10 does not list that as a criterion for dissocial personality disorder. Concern was expressed as to reliance on what were perceived as ‘over-prescriptive criteria’.
Conditions that appear in one but not both
There are 7 DSM-5 disorder categories that do not appear in ICD-11 and 19 in ICD-11 that do not appear in DSM-5. The use of diagnoses that do not appear in one or other, or indeed in either, classification sets the scene for a weaponisation of the classification(s) in adversarial proceedings.
In McDonald v Conroy [2024], the plaintiff sought an amendment to her statement of claim to include ‘complex PTSD’ as a particular of injury, based on her expert psychiatric evidence. The defendant resisted on the grounds that CPTSD is not a recognised disorder in DSM-5. The court permitted the amendment. In 2009, CPTSD was in neither the DSM nor the ICD and this became an issue in AN, R (on the application of) v Secretary of State for the Home Department [2009], where the court did not find the debate about the fact that CPTSD did not appear in either classification of much assistance.
A particular problem faced by the courts is in cases where there is psychopathology that does not fall into a defined category, whether under the ICD or DSM. Although C v D [2006] and Lawson v Glaves-Smith [2006] do not refer explicitly to the use of the DSM or the ICD, given the reference to these cases in D v The Bishop’s Conference of Scotland [2022] and the involvement of the same expert in all three, it is relevant to consider all of them in relation to this problem. In C v D, the court accepted the expert’s evidence that C was not suffering from a diagnosed illness but nonetheless ‘has suffered and continues to suffer from mental abnormality as distinct from emotional distress’. In Lawson v Glaves-Smith, where the expert’s opinion was that the claimant had not suffered from PTSD, the court concluded that there was no plausible explanation other than that the claimant suffered disabling trauma and psychological symptoms as a direct result of what happened to her, and whether it was right to attach the label PTSD was perhaps less significant than to assess the symptoms she actually suffered herself (even if they did not necessarily correspond completely to those incurred by other people who have been diagnosed with PTSD). In D v The Bishop’s Conference of Scotland, neither expert diagnosed psychiatric injury. As well as considering C v D and Lawson v Glaves-Smith, the court also drew attention to what Lord Reed said in Rorrison v West Lothian College [1999], to the effect that, although what constitutes a recognised disorder is a matter for expert evidence, the ICD and DSM classifications are not necessarily conclusive. Commenting on the evidence of the claimant’s expert, the court said it was important to recognise that he was using DSM-5 as the sole basis for deciding if a ‘defined or classified psychiatric injury existed’ but considering the evidence in the round, there was other evidence from the expert that pointed clearly towards psychiatric injury at that time, albeit not falling within DSM-5.
A similar issue arose in Dudley Metropolitan Borough Council v Mailley [2022], where the experts agreed that Mrs Mailley’s hoarding behaviour was at a relatively severe level, but they also agreed that a diagnosis of hoarding disorder (whether under DSM-5 or ICD-10) should not be made, albeit for different reasons. Mrs Mailley’s expert was of the view that DSM-5 requires that hoarding should not be due to identifiable mental disorder and here it was likely to be due to her depression. The Council’s expert was unwilling to make a diagnosis of hoarding disorder because, he said, it was not universally accepted and it did not appear in ICD-10. The court was not impressed with either rationale. The fatal flaw in the analysis of Mrs Mailley’s expert was that Mrs Mailley had been, from a lay understanding of the term, ‘hoarding’ items for many years and when not suffering from depression. As to the Council’s expert, the court’s judgment was that he was too influenced by (and not correct about) the lack of formal adoption of hoarding disorder within international classifications. He accepted in examination in chief that he was wrong about this, in that it was ‘added to ICD 10 in 2017’, which must be a reference to the publication of ICD-11.
The importance of applying clinical judgement and not counting ticks
In Hibbert v The Ministry of Defence [2008], although the claimant satisfied Criterion A of DSM-III-R PTSD and had symptoms consistent with Criteria B, C and D, in one expert’s clinical judgement ‘there was not a sufficiently serious range of experiences to amount to PTSD’. In explaining why he was not persuaded that the claimant was suffering from PTSD, the judge said that the other expert, who diagnosed PTSD, was paying regard to the presenting symptoms but disregarding his clinical judgement as to the severity of such symptoms.
Similarly, a comment in Calvert v William Hill Credit Ltd [2008] indicates that the courts will not unquestionably accept a threshold or above-threshold score as evidence of even just probable disorder. The claimant ‘scored five on the DSM-IV diagnostic test […], sufficient to justify his classification as a “probable pathological gambler”’, but the court acknowledged that there may be cases in which such a score on ‘the DSM-IV tests may be attributable in whole or in part to special factors which, upon close analysis, lead to a contrary conclusion’. This illustrates the observation of Clark et al (Reference Clark, Cuthbert and Lewis-Fernández2017) that thresholds are pragmatically necessary for various reasons but diagnostic thresholds can also be ‘highly consequential’ in forensic settings.
Likewise, the references to ‘sufficient ticks’ and a ‘rigid analysis’ in Vernon v Bosley (No 1) [1996] illustrate the difficulties for the court when, in contravention of the guidance about the use of the DSM and ICD and the need to take into account clinical judgement, the criteria are applied too strictly. In Lindsay v Wood [2006] one expert expressed the opinion that ‘if one strictly applies the ICD-10 criteria for an Organic Personality Disorder (the Claimant) does not presently meet the criteria for this condition’, but in the course of the experts’ meeting agreed that the claimant was ‘suffering from a mental disorder within the meaning of the Mental Health Act 1983, i.e., organic personality disorder’. As mentioned above, in DL-H v Devon Partnership NHS Trust v Secretary of State for Justice [2010] the tribunal referred to the risks in using ‘over-prescriptive criteria’.
Discussion
The courts are not assisted by expert psychiatric witnesses who: use the DSM or ICD classifications in preference to more authoritative or up-to-date textbooks or peer-reviewed publications; make misleading claims about the use or status of a classification in order to endow it with an authority in legal proceedings that it does not have and that is contrary to guidance as to its application; fail to make the court aware of the classifications’ questionable scientific validity; use them as checklists; do not appreciate that not every entry in the classifications is a mental disorder; distract the court with disagreements about diagnosis when the court is concerned not with the diagnosis but with the condition, its onset, its effects, its trajectory and/or its prognosis; fail to apply clinical judgement; fail to use the classifications with flexibility; rely on their diagnoses without considering other relevant evidence; disregard research that calls into question the reliability of the classifications in delineating disorders, establishing their severity or identifying the presence of disorders; and failing to point out, where appropriate, that a listed diagnosis is controversial.
The danger of the inappropriate use of the ICD and DSM is that they can give a spurious impression of certainty in delineating disorder from normality and one disorder from another. But psychiatry is ‘not an exact science’, as Baroness Hale, Hon FRCPsych, said in R (B) v Ashworth Hospital Authority [2005] and, as the above analysis indicates, psychiatrists’ reliance on these classifications may be more of a hindrance than a help to the courts. Not all courts understand the classifications as well as those quoted in Box 6.
Judicial verdicts on the application of ICD and DSM
‘DSM-III-R may provide the medical profession with a useful diagnostic tool but PTSD and its DSM-III-R classification should not […] be adopted in personal injury litigation as the yardstick by which the plaintiffs’ success or failure is to be measured.’
(Lord Justice Stuart-Smith in Vernon v Bosley (No 1) [1996])
‘The plaintiff was clearly in difficulty in picking up sufficient ticks within boxes B, C and D of the diagnostic criteria […] But psychiatric illness is too complex and insufficiently concrete to be subjected to such a rigid analysis.’
(Lord Justice Thorpe in Vernon v Bosley (No 1) [1996])
‘For the purpose of the assessment of the pursuer’s claim for damages it is the practical effect on him, rather than the appropriate diagnostic label, that matters.’
(Duthie v MacFish Ltd [2000])
‘I am not persuaded that the difference in their diagnosis and in the precise label to be attached to the pursuer’s condition makes a lot of difference.’
(Hynd v David Reekie & Sons Ltd [2013])
‘Ultimately, it would appear that the unanimous view of the doctors is that the DSM is an invaluable protocol and tool, but it is not a mere checklist […] It is a valuable tool, but in my view, one must weigh heavily the essential and important ingredient of the diagnosis of an experienced medical professional coming to an informed view aided, as I say, by the collective wisdom and guidance to be found in DSM 5 in this case, or indeed in ICD 10 which is occasionally mentioned, but seems to lag somewhat behind DSM-5 in popularity of reference when evidence is given before this Court.’
(W v The Minister for Health and Children [2016])
‘As some of the manuals themselves make clear, care is needed before the classifications are used in a forensic context.’
(R v Wilcocks [2016])
‘I do not think that focusing on labels is of much assistance and tends to detract from the substance of the concern.’
(LKM v NPM [2023])
Psychiatrists who rely on the ICD or DSM may better assist the court if they answer ‘up front’ the questions that lawyers and the courts often, or might, ask about these classifications. Box 7 gives examples of such questions and answers and in a form that could be appended to an expert report.
Frequently asked questions by lawyers about the DSM and ICD
What are the DSM and ICD?
DSM refers to the Diagnostic and Statistical Manual of the American Psychiatric Association. ICD refers to the World Health Organization’s (WHO) International Classification of Diseases.
What are the DSM and ICD for?
They have similar purposes.
The DSM is a manual designed to meet the needs of one, or perhaps two, professions – psychiatrists and psychologists – in a single country, the USA, in order for clinicians and researchers to be able to diagnose and classify mental disorders.
The ICD has been developed to facilitate diagnosis in clinical settings by a wide range of health professionals in countries of very varied sizes, cultures and resources as well as to serve as a statistical classification system.
They both classify disorders not people.
What are the differences between the DSM and ICD?
The main differences are in their approach to diagnosis. The DSM uses an algorithmic approach which simply means counting symptoms and signs, applying limits based on periods of time and applying exclusion criteria. The ICD uses a prototypical approach, which means matching the person’s presentation against a description of the typical presentation of the disorder.
How are the classifications produced?
They are both the products of committees. As some of the DSM committees meet in secret, its processes cannot be fully described. The classifications reflect who is on the committees and what their interests are. In the case of the DSM, they may be US pharmaceutical companies and patient advocacy groups. In the case of the ICD, they may reflect the interests or concerns of particular WHO member states.
They represent a consensus, but not what would be called a scientific consensus.
So, what is the validity of the classifications?
There is little or no research to validate the distinctions made between what appear to be discrete disorders or to validate the distinction between ‘normality’ and ‘pathology’. The fact is that the ICD’s prototypical descriptions are of limited value because most people with a mental disorder do not conform to these stereotypical descriptions. The fact is that more people with mental disorder have conditions that represent an overlap between several categories of disorder than fall neatly into one category. The fact is that there is no scientific basis for basing a diagnosis on an arbitrary number of symptoms or signs or an arbitrary period of time. The core features of most mental disorders lie on a continuum with normality and for these the dividing line between normality and abnormality is somewhat arbitrary. There is evidence to show that at least for post-traumatic stress disorder (PTSD), ICD and DSM criteria identify different people as having the disorder and they create a dividing line between those with and those without the disorder that is inconsistent with the severity of their psychopathology, thus calling into question the validity of the dividing line.
How are the ICD and DSM used?
They are meant to be used flexibly and with the application of clinical judgement. Even though the DSM, with its more tightly defined and counted criteria, is regarded as having a superiority for research, not only do practising clinicians in the USA seldom adhere strictly to the specific DSM diagnostic criteria, but clinicians’ loosely chosen diagnoses are relied on in research. Expert psychiatric evidence is sometimes presented to the courts with an apparent disregard for the requirements of flexibility and the application of clinical judgement and even occasionally with their explicit disregard.
What is the value of the ICD and DSM in legal proceedings?
Guidance as to the use of the DSM specifically cautions against its ‘forensic’ use. The ICD and DSM are not designed or intended to answer the questions of ultimate concern to the law in cases where psychiatric or psychological conditions are in issue.
Which one is better for medico-legal purposes? I’ve heard psychiatrists submit that the DSM is preferable because it is used in research
The DSM is used in research but so is the ICD. The WHO produces a version Diagnostic Criteria for Research. The disadvantage of using the DSM or ICD’s research version is that the rigorous application of criteria, necessary to recruit relatively homogeneous groups of participants, for example for a treatment trial, leaves too many people unclassified and therefore to be allocated to the ICD categories ‘not otherwise specified’ and the DSM-5 categories ‘not elsewhere defined’. It may not matter to researchers that babies are getting thrown out with the bathwater if it means that conclusions based on those in the bath can be confidently applied to others equally narrowly defined. But the court has to understand an individual defendant, claimant, complainant, witness, etc. whether or not thrown out with the bathwater.
What if the experts disagree as to diagnosis because one relies on DSM and the other on ICD?
It may not matter. Particularly when criteria are applied as flexibly as they are meant to be, and in accordance with clinical judgement, there may be no significant difference as to the nature, severity, treatment and prognosis of the condition and no difference material to the issues before the court. Even differences as to causation, as for example in post-traumatic stress disorder or bereavement, may not have a bearing on the determination of the issues before the court.
What if the diagnosed disorder is in one classification but not the other?
It may not matter much. Just because the condition does not have a diagnostic label does not mean that it does not exist. Complex post-traumatic stress disorder (CPTSD) is the example most familiar to the courts. Let’s start with COVID-19. It did not exist in 2018. CPTSD is different. It was in 2018 that CPTSD first appeared in ICD-11. This does not mean that it did not exist before 2018, as might be said of COVID until 2019. People thereafter diagnosed with CPTSD had not developed the condition overnight in May 2018. It is not in DSM-5. This does not mean that they do not have a disorder. It only means that it is not named and classified with the same specificity as in ICD-11.
But it may matter. The diagnosis may be controversial, such as intermittent explosive disorder or paedophilia. The expert should be expected to explain their reliance on a controversial or unusual diagnosis. It may be a newly delineated disorder, as CPTSD was at the turn of the 21st century. The expert should be expected to explain its origins and delineation by reference to published commentary, authoritative discussion and, preferably, research.
What if the expert does not mention the ICD or DSM?
It does not matter. Psychiatrists can communicate with each other without the ICD or DSM and reference to the ICD or DSM may not be necessary to assist the court. Indeed, reliance on the ICD or DSM may confuse the court, rendering expert evidence of little or even no assistance.
Courts do not need DSM or ICD diagnoses, although some ask for them. Psychiatrists can often, perhaps should always try to, assist the court without reference to the ICD or DSM. They can be more a hindrance than a help.
The courts, like clinicians, are concerned with people, and a clinician’s diagnosis, however based, assists only partly, if indeed at all, the court’s understanding of the person. Diagnostic classifications are about disorders. They have a value in assisting governments and the WHO as to the incidence and prevalence of mental disorders and in identifying for research people who are sufficiently homogeneous in terms of their disorder to advance medical knowledge and treatment with as few confounding variables as possible. Their forensic value is limited.
I’ve heard that unhappiness was listed as a mental disorder in ICD-10. Is that right?
No. It was listed in its Annex under ‘Symptoms, signs and abnormal clinical and laboratory findings, not elsewhere classified’ and coded R45.2. It is no more a mental disorder than ‘Holiday relief care’ (Z75.5), ‘Lack of physical exercise’ (Z72.3), ‘Examination for medicolegal reasons’ (Z04) or ‘High-risk sexual behaviour’ (Z72.5).
_______________________
Note: This box can be reproduced by experts as an appendix to reports, in communications with solicitors or other commissioners of expert psychiatric evidence or in any other way so as to assist courts and tribunals, subject to the inclusion of this acknowledgment of its source: Rix K (2025) DSM and ICD classifications in medico-legal reporting: misperceptions, misunderstandings and misuse. BJPsych Advances. Available from: https://doi.org/10.1192/bja.2025.10143
Box 8 provides guidance to psychiatrists when assisting a court as to diagnosis and contemplating reliance on the ICD or DSM. The first piece of guidance echoes in short form the questions asked by Markon (Reference Markon2013):
‘Curiously, largely absent from these discussions has been the question of whether or not we should have authoritative nosologies at all. That is, do we need a DSM, ICD […] or similar document in any form? More importantly, are such nosologies helpful or harmful to the scientific process? What consequences would there be for abandoning authoritative nosologies?’
Guidance to expert witnesses on the use of DSM and ICD classifications
-
Consider whether a simple, conventional diagnosis will do, without having to use the DSM or ICD.
-
Having regard to the risk of reliance on the DSM or ICD being a hindrance rather than a help, give careful thought to whether your reliance on one or both is going to assist the court.
-
Provide sufficient information about the classification’s status, nature, purpose and limitations, so as to avoid misunderstanding or misuse.
-
Do not try to bolster the status of a classification in order to add weight to your evidence, by making misleading statements about the classification.
-
Remind yourself of the explanations as to how they are to be used – flexibly and with the exercise of clinical judgement.
-
If you rely on numbers of features or specified durations, which is probably not a good idea, be ready to identify the research on which they are based.
-
Just as in ordinary clinical practice most patients do not conform to textbook descriptions, the subject of your report may not conform to the tidy stereotypical descriptions in the ICD and they may satisfy some, but not all, of the criteria of two or three different DSM categories. If many of your patients fall into the ‘not otherwise specified’ or ‘not elsewhere defined’ categories, the same may apply to the subjects of your reports. If so, say so and do not struggle for a match or category that obscures or oversimplifies the complexity of the presentation. It may be the complexity of the person’s condition that the court needs to appreciate.
-
An ICD or DSM diagnosis is not a substitute for a holistic diagnostic formulation appropriate for the person, considering their individual, social and cultural context.
-
An ICD or DSM diagnosis is only part of the psychiatric assessment, so be sure to consider the totality of the available evidence, some of which might cast doubt on the ICD or DSM diagnosis.
-
Be mindful of the range of reasonable opinion and refer to possible alternative, particularly overlapping, diagnoses that might be made by a reasonable colleague.
-
Again, having regard to the range of reasonable opinion, if you have used DSM, and in case another expert uses ICD (or vice versa), pay careful attention to the potential significance of the differences between the present two diagnostic systems, which are major for 20, and of a minor definitional nature for 42, disorders or diagnostic entities. Sort these out in the report and save time at the experts’ discussion. They may have no bearing on the issues in the case.
-
If a disorder or diagnostic entity is in one classification but not the other, investigate carefully why this is. Using a controversial diagnosis carries risks and requires explanation.
-
Do not turn DSM-5 V codes or ICD-11 Z codes into diagnoses. Low income (DSM-5) and problems related to a life management difficulty (ICD-11) are not mental disorders.
-
It probably does not matter when a classification was published, adopted or put into effect. These are dates on which classifications change, not people’s conditions.
For ‘scientific’ read ‘legal’.
Conclusion
Reliance on the ICD and DSM in medico-legal reports carries risks for expert witnesses, whose credibility may be called into question, but more importantly for the delivery of justice. To manage these risks experts have to handle them with caution and care.
MCQs
Select the single best option for each question stem
-
1 About the DSM or ICD:
-
a clinicians in the USA adhere strictly to DSM diagnostic criteria
-
b the ICD is more widely used than the DSM in research around the world
-
c they are the only authoritative classifications of mental disorder
-
d the DSM is legally recognised in Ireland
-
e their purpose is to aid communication between clinicians.
-
-
2 Which of the following is not an example of a difficulty for a court or tribunal when experts rely on the ICD or DSM?
-
a identifying what may be a controversial mental condition
-
b whether for a diagnosis of PTSD there has to be a recollection of the trauma
-
c the magnitude of a traumatic stress that causes PTSD according to DSM diagnostic criteria
-
d the imperfect fit between questions of ultimate concern to the law and information contained in a clinical diagnosis
-
e disagreement as to whether or not the ICD or DSM criteria are met.
-
-
3 Classifications such as the ICD and DSM can be used by the courts:
-
a. to determine whether a person’s psychiatric condition amounts to a psychiatric injury
-
b. to decide what is a recognised medical condition
-
c. in personal injury litigation as the yardstick by which a claimant’s success or failure is to be measured
-
d. to decide whether a person has a disability
-
e. to decide in a criminal case whether they have a defence of insanity.
-
-
4 According to the ICD or DSM, which the following is a form of mental disorder?
-
a high-risk sexual behaviour
-
b lack of physical exercise
-
c holiday relief care
-
d schizoaffective psychosis/disorder
-
e examination for medico-legal purposes.
-
-
5 As regards the DSM and ICD, courts and tribunals should be informed:
-
a as to which is the right classification to use for medico-legal purposes
-
b as to the date on which the classification became effective
-
c that the DSM is for use by psychologists and the ICD is for use by psychiatrists
-
d that reliance on the ICD or DSM may confuse the court, rendering expert evidence of little or even no assistance
-
e that their utility is in classifying people according to the mental disorders with which they have been diagnosed.
-
MCQ answers
-
1 e
-
2 a
-
3 b
-
4 d
-
5 d
Data availability
Data availability is not applicable to this article as no new data were created or analysed in this study.
Acknowledgements
The judgments that are the source of the illustrative cases in Boxes 3, 4 and 5 are © Crown copyright. Access to these has been through the website of the British and Irish Legal Information Institute (BAILII). In accordance with its reproduction and copyright policy we therefore identify BAILII as the source of these and advise readers that these judgments are freely available on the BAILII website: https://www.bailii.org/.
Funding
This work received no specific grant from any funding agency, commercial or not-for-profit sectors.
Declaration of interest
None.
Open access
eLetters
No eLetters have been published for this article.