Leigh Neal has suggested that the increased incidence of flashbacks that we detected for Gulf War veterans is not a genuine observation but simply the result of contemporary overreporting. This effect he attributes to our ‘compensation culture’ and malingering. While we fully agree that claimants with PTSD may on occasion elaborate psychological symptoms for financial reasons, this factor is hardly novel (Reference WesselyWessely, 2003). There was, for example, an epidemic of war pension claims for shell shock and neurasthenia in the aftermath of the First World War. By March 1921, it was estimated that of the 1.3 million awards, 65 000 were for functional nervous disorders (Reference Jones, Palmer and WesselyJones et al, 2002). So concerned was the Ministry of Pensions that applications were being falsified or exaggerated that they appointed Sir John Collie, an expert in rooting out fraud, to chair their ‘special medical board for neurasthenia and functional nerve disease’. In 1917, Collie had included a chapter on the military in his textbook, in which he observed that ‘the thin line which divides genuine functional nerve disease and shamming is exceedingly difficult to define’ (Reference CollieCollie, 1917: p. 375). In fact, concerns about spurious or exaggerated claims for functional disorders pre-dated this conflict and followed the passing of the Workmen's Compensation Acts of 1897 and 1906. In the 6 years following the 1906 Act, the sums paid in accident compensation rose by 63.5% – despite the fact that the number of people in employment remained the same (Reference TrimbleTrimble, 1981). The research in the 1880s by Herbert Page to establish that most cases of railway spine were without organic basis was driven by the large settlements being paid by railway companies to passengers who had exaggerated or falsified symptoms following accidents. Indeed, the term Rentenkampfneurosen (pension struggle neurosis) had been coined following Bismarck's accident insurance legislation of 1884 and reflected widespread concerns that workers and passengers were defrauding companies through dubious medical claims (Reference LernerLerner, 2001).

Other than agreeing that these things can and do happen, it is always risky to make statements about the incidence of malingering, as clinicians have no particular expertise in its measurement. Dr Neal has no more information than we have, or anyone else for that matter, on the true rates of malingering, let alone whether or not it is increasing. What the above does show is that concern about the phenomenon is certainly not new.

Menachem Ben-Ezra rightly points out that the flashback is a comparatively rare symptom among PTSD sufferers. He argues that other symptoms, such as nightmares, sleep disturbance and elevated anxiety, are common and enduring features, and, therefore, not culture-bound. While we agree that these symptoms were widely reported in the past, their existence per se does not justify the creation of a new and very specific disorder. The complex diagnostic criteria for PTSD in DSM–IV (American Psychiatric Association, 1994) comprise six sub-groups, which extend over three pages. Anxiety, sleep disturbance and nightmares are not disorders in themselves, as most people suffer from them at some time. It is only when they become severe or arise inappropriately that psychiatrists elevate them to psychiatric disorders. With the exception of hallucinogen persisting perception disorder, flashbacks are almost unique to PTSD. As a result, we chose this symptom as a way of trying to evaluate the incidence of this modern diagnosis. It should not be forgotten that PTSD did not enter DSM–III (American Psychiatric Association, 1980) as a result of a series of rigorous epidemiological investigations but in the context of an anti-war movement, which sought to demonstrate that servicemen suffered long-term effects from combat. Only after it had been formally recognised by the American Psychiatric Association was PTSD then subject to intense scientific analysis (Reference YoungYoung, 1995).

Dr Burges Watson has identified not only the growing significance attached to the flashback but also the disparity between the way that flashbacks are described as part of the diagnostic criteria for PTSD and in the DSM–IV glossary. In the former, they are included within ‘acting or feeling as if the traumatic event were recurring (includes a sense of reliving the experience, illusions, hallucinations, and dissociative flashback episodes...)’, while the latter contains a brief definition: ‘a recurrence of a memory, feeling, or perceptual experience from the past’ (American Psychiatric Association, 1994: pp. 428, 766). Dr Burges Watson infers from this that the flashback is a new term for an old phenomenon; what in the past would have been described as a vivid memory of conflict is today called a flashback. The objection to this hypothesis is that we discovered both phenomena in medical records from the First and Second World Wars. We were careful to adopt a rigorous definition of flashback (which included the sense of reliving the traumatic episode) to distinguish it from eidetic memories.

In answer to Dr Hambidge, we were unable to include veterans of the Falklands War because ministerial permission was not granted to study recent war pension files of service personnel still living, and because the Medical Assessment Programme is limited to veterans of the Persian Gulf War. As regards the collection of data, three research assistants recorded symptoms on a standardised form by copying verbatim from medical notes. These were then reviewed in detail by the lead investigator, who re-examined the files to ensure accuracy and consistency of interpretation. War pension files with missing information were excluded from the study. In general, the case notes were comprehensive, often detailing a serviceman's history from enlistment until death. As these are a continuous series of records, there is no reason to suppose that deficiencies in reporting were confined to modern assessors rather than being spread randomly throughout the archive.