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.

Correcting the misunderstandings

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.

TABLE 1 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’.