JRSM >
Volume 96, Number 3 >
Pp. 155-156
Erw Lon, Pentre Galar, Crymych, Pembrokeshire SA41 3QP, UK
Professor Lee and colleagues (October 2002 JRSM1) presented their findings in 1000 Gulf War veterans who came to their programme for medical advice, and combined the data with those from two previous surveys, each of 1000 veterans, by Coker et al.2 and their own group3. I question the methodology and statistical models adopted by Lee et al. in their two papers1,3.
First, why did they not adhere to the reviewing criteria adopted in the first paper by Coker? Introduction of the term ‘wellness’ serves to confuse. There is no differentiation between individuals who were well when they attended and those who considered that they were unwell. Second, would it not have been better to adhere to Chapter 18 diagnosis (International Statistical Classification of Diseases) throughout the series, with appropriate comment? To change the criteria between papers prevents comparison of like with like. Third, whilst chronic fatigue syndrome (CFS) is included in the initial paper—87 diagnosed with the syndrome—it has been abandoned by Lee et al., despite the predominance of relevant symptoms. ‘Multiple unexplained physical symptoms’ and somatic disorders are included, which have no accepted criteria.
The major concern is the inclusion of 1057 individuals who were well by any definition and attended for reassurance, information, etc. This fundamentally distorts the statistics: overall, the percentages should be taken not from 3000 but from the 1943 individuals who attended because they felt ill. If the Chapter 21 individuals are extracted from Table 2, this leaves 837 ill in the first 1000, 494 in the second 1000 and 612 in the third 1000. Affective symptoms are then, respectively, 59% (from 50%), 98% (from 49%) and 62% (from 38%); fatigue 50% (from 42%), 91% (from 45%) and 42% (from 26%). This is reflected throughout. Table 3 then becomes uninterpretable. Table 5 should be based on 1943 individuals and is effectively distorted by 50%—thus, post-traumatic stress disorder (PTSD) becomes 19% (from 12%), depression 9% (from 6%) and so on.
Even if we allow for the self-selection in the veterans seen by Lee et al. there is a major discrepancy regarding PTSD between this cohort and the series reported a few weeks later by Wessely's group in the BMJ4. In Lee's paper the prevalence is said to be 12% but should be 19%, as I have shown; in Wessely's paper it is no more than 3%. Could the explanation be that Lee et al. dropped CFS from the reviewing criteria used by Coker et al.? Should many of their PTSD diagnoses have been CFS?
Because of these aberrations, the two papers by Lee et al. contribute little to the understanding of illnesses affecting Gulf War veterans.
REFERENCES
Gulf Veterans' Medical Assessment Programme, Baird Health Centre, St
Thomas' Hospital, London SE1 7EH
1 Gulf Veterans' Illnesses Unit, Ministry of Defence, St Christopher House,
London SE1 0TD, UK
As is made clear in our paper, the intent was to review the diagnoses of all 3000 who attended and to make some assessment of wellbeing, as detailed by Lee et al1. Those attending the Gulf Veterans' Medical Assessment Programme (GVMAP) are referred by their medical attendants on the basis that they have health concerns relating to their service in the Gulf. It is therefore appropriate to include all 3000 referred, as the population in which the prevalence of diseases found is expressed rather than restricting the population denominator to just those found to be ill. When one discusses the prevalence of disease in a population, one usually includes the whole population as the denominator rather than just the unwell portion. This addresses Wilson-Ing's concern that we have fundamentally distorted the statistics.
Wilson-Ing further suggests that we have changed diagnostic criteria because of the reduction of Chapter 18 diagnoses. The first point to make is that the bridging 100 exercise in our first paper1 was designed to compare the diagnostic criteria between that paper and that of Coker et al2. Secondly, it is hardly surprising that the diagnostic pattern is different, given both the time that has elapsed and the considerable research effort that has gone into Gulf veterans' illnesses without finding any convincing evidence of unique Gulf-related illness. To suggest that the diagnosis of chronic fatigue syndrome (CFS) has been abandoned by us is disingenuous, since Wilson-Ing is not in a position to make clinical diagnoses on our series. We, who have seen the patients, are satisfied that the symptoms presented do not amount to chronic fatigue syndrome3,4. We have certainly not included medically unexplained physical symptoms or somatic disorders in our diagnoses, but have discussed these aspects in the discussion section of the paper. We would suggest that the references therein provide the background and criteria for our discussion.
In his comments about PTSD, Wilson-Ing is falling into the trap of not comparing like with like. Since PTSD is associated with exposure to trauma, it is likely that we would see more of it, since we were reporting upon the diagnoses of a group of military veterans who had health concerns as a result of their service in a conflict, than Ismail would find in a study examining the mental health of a sample of the whole Gulf veteran population which includes those with no health concerns5. The analogy is to claim that there is a higher prevalence of lung cancer among patients attending a chest clinic than amongst the general population. The suggestion that the alleged excess of PTSD may be misdiagnosed CFS rather misses the point, explicitly stated in our paper, that the suggested diagnoses of PTSD were confirmed by consultant psychiatrists.
REFERENCES
CiteULike 
Complore 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
