Heterogeneity of presentation and the absence of diagnostic biological or other tests make the classification of most mental disorders challenging. Mental health professionals currently rely on the Tenth Edition of the International Classification of Diseases (ICD10; World Health Organization, 1992) or the Fourth Edition of the Diagnostic and Statistical Manual of Mental Disorders (DSMIV; American Psychiatric Association, 2000) to arrive at a diagnosis. As of May 2013 this will change; DSMIV will be replaced by DSM5 and, in 2015, ICD10 will be replaced by ICD11.

The transition to new versions of ICD and DSM has been seen as an opportunity to bring them closer together (e.g., Kupfer et al., Reference Kupfer, Regier and Kuhl2008; First, Reference First2009; Jablensky, Reference Jablensky2009) or even to make them identical to eliminate the confusion caused by different definitions of mental disorders (Frances, Reference Frances2009). The establishment of an ICD–DSM Harmonisation Coordination Group to facilitate this process was cause for optimism that convergence would occur (Jablensky, Reference Jablensky2009). With the differences in ICD10 and DSMIV classifications of post traumatic stress disorder (PTSD) seen as ‘definitional differences without an apparent conceptual basis’ (First, Reference First2009) the path seemed clear for ICD11 and DSM5 definitions of PTSD to be close if not the same.

If the most recent proposals (American Psychiatric Association, 2013; Cloitre, Reference Cloitre2012; Brewin, Reference Brewin2012) for the criteria are adopted, rather than being reduced, the differences between ICD11 and DSM5 will be far greater than between ICD10 and DSMIV. The proposed number of symptom criteria for DSM5 PTSD represents a rise from 17 to 20 compared with DSMIV, while the proposed ICD11 PTSD symptom criteria comprise only six items and introduce a new parallel diagnosis of complex PTSD. Despite considerable discussion about the creation of a new diagnostic category to cover complex presentations of PTSD in DSM5, this is not proposed.

The sole welcome convergence between the classification systems is in the definition of the traumatic event required for PTSD. In the proposed DSM5 criteria (American Psychiatric Association, 2013), the DSMIV A2 criterion that individuals had to experience horror, helplessness or fear at the time of the event is to be removed and the nature of the event tightened to make it more specific. The DSM re-experiencing phenomena will be relatively unchanged. This is also true for the current avoidance criteria, but these will no longer be grouped with the other DSMIV C criteria and will become the DSM5 C criteria on their own.

The biggest proposed change to DSM5 PTSD is the introduction of a new symptom cluster based on confirmatory factor analysis (Friedman et al., Reference Friedman, Resick, Bryant and Brewin2011). This moves PTSD away from being a fear-based disorder (unlike the proposal for ICD11) and introduces a new D criteria cluster defined as ‘negative alterations in cognitions and mood that are associated with the traumatic event’. The D criteria include new items such as ‘persistent and exaggerated negative expectations about oneself, others, or the world’ and ‘persistent distorted blame of self or others about the cause or consequences of the traumatic events’. The proposed hyperarousal criteria for DSM5 are similar to DSMIV but now also include ‘reckless or self-destructive behavior’.

The proposed changes to the criteria for ICD11 are more striking, especially the proposed inclusion of the complex PTSD diagnosis. The criteria for PTSD are to be reduced to just six items – two re-experiencing (nightmares and flashbacks), two avoidance (of thoughts and feelings, and of places, people and activities) and two hyperarousal (increased startle and hypervigilance). Simplification, if valid, is to be welcomed, but by only including flashbacks and nightmares as eligible re-experiencing phenomena, it appears that a number of individuals currently diagnosed with PTSD as a result of distressing vivid images without dissociation will be excluded from the ICD diagnosis. It will be interesting to see if field testing confirms this.

The proposed ICD11 criteria for complex PTSD comprise the PTSD criteria plus affect dysregulation (heightened emotional reactivity, violent outbursts, impulsive or reckless behaviours and dissociation); defeated/diminished self (feeling diminished, defeated, worthless, shame, guilt, despair); and disturbed relationships (difficulties in feeling close, little interest in relationships or social engagement). Among clinicians, the inclusion of complex PTSD will probably be the most popular change to either classification system. The fact that complex PTSD is not to be classified as a developmental disorder, acknowledging that some individuals may develop complex ‘characterological’ symptoms after a traumatic event(s) experienced in adult life, is also likely to be welcomed.

Why the divergence between ICD and DSM? Despite the calls for harmonization, there are differences in what the two systems are trying to achieve. Kendell (Reference Kendell1991) described ICD as a ‘comprehensive classification of all’, whereas DSM ‘is designed to meet the needs of one or perhaps two professions in a single country’. Interestingly, and contrary to what many of us might have thought, ICD and not DSM is the official classification system in the USA (Reed, Reference Reed2010). For PTSD, however, and perhaps more so than for any other disorder, DSM has been widely used internationally as the gold standard for diagnosis, not least because the only way to diagnose it between 1980 and 1992 was by using DSM (it was included in DSM-III (American Psychiatric Association, 1980) but only became an ICD diagnosis in ICD-10 (World Health Organization, 1992)).

The WHO is keen to use ICD11 to improve the clinical utility of diagnosis given concerns about problems with this in earlier versions of ICD and DSM (Reed, Reference Reed2010). The complexity of current diagnostic systems has been argued as a major problem, with over-specification leading to reduced clinical utility. It is hoped that enhanced clinical utility (see Box 1) will lead to improved care and reduced burden of disease.

The simplified proposed criteria for ICD11 do appear to have the potential to improve clinical utility, but it is more difficult to argue this for the increased diagnostic criteria proposed for DSM5. A key issue for everyone concerned with PTSD (e.g., clinicians, patients, families, clinicians, administrators, researchers, lawyers) is what will be the impact of major differences between ICD11 and DSM5. It seems highly likely that a significant number of individuals will satisfy the criteria for DSM5 PTSD but not ICD11 PTSD and vice-versa. There is a major risk that the systems will be used interchangeably depending on whether the presence or absence of a diagnosis of PTSD is desired. The potential for this to cause confusion seems to be high and will need to be carefully managed if clinical (and other) utility is not to be reduced rather than improved.

Despite the same evidence being available to the DSM5 and ICD11 development groups, it appears that for PTSD First's (2009) argument that ‘for the most part DSM5 and ICD11 harmonisation will have to rely primarily on a ‘negotiated expert consensus’ process in which representatives from the DSM5 and ICD11 work groups work together to hammer out differences between the two systems' although true has not been followed.

To achieve global clinical utility, we need a single, simple classification system that is valid and reliable. Our current knowledge is insufficient to allow us to achieve this and history shows that, despite some limitations, it is not be necessary to have two identical systems. Jablensky (Reference Jablensky2009) argued that ‘… having two parallel classifications with some explicitly stated conceptual differences between them does help to highlight the provisional nature of many nosological concepts and their arbitrary definitions’. It is vital that we use the likely level of divergence to undertake research that allows ICD12 and DSM6 to become closer. It can, of course, be argued that it would have been better to have undertaken the necessary research, including field testing of the proposed ICD11 and DSM5 criteria against each other, to inform ICD11 and DSM5 rather than subsequent iterations.

On the surface, it seems that a major opportunity has been missed to harmonize the ICD and DSM classification systems for PTSD. Maybe a more formal agreement between WHO and APA well ahead of the development of ICD12 and DSM6, along with significantly greater common membership of work groups, will result in harmonization. It is to be hoped that the inevitable major differences between the ICD11 and DSM5 definitions of PTSD drive the field forward, but it seems reasonable to be somewhat nervous.