Editorial bias in the Lancet

 

Ellen M. Goudsmit PhD C.Psychol.

Bart Stouten PhD

Sandra Howes

Contact details: ellengoudsmit@hotmail.com

 

Abstract

A literature search identified all papers published on chronic fatigue syndrome (CFS) and myalgic encephalomyelitis (ME) in the Lancet between 1995 and 2000. Analysis of the findings revealed a bias towards the views of one school of thought and a lack of papers on the immunological or virological aspects of CFS. The lack of balance contrasts with the mainstream American journals, which generally covered a much wider range of subjects and views. We examine the arguments against covert editorial policies, and illustrate how the present situation in the UK is undermining the medical profession’s understanding of both CFS and ME.

Introduction

Since CFS is not adequately covered in the standard British medical texts, the majority of physicians in this country have to get most of their information on this illness from reputable medical journals. It is therefore essential that the quality of that information is of the highest standard; that it reflects a range of expert views, and that it is accurate and balanced. Following an earlier analysis of papers in the British Medical Journal (BMJ) 1, we conducted a similar search of the other ‘high impact’ general medical journal published in Britain, the Lancet.

Method and results

A search of MEDLINE for publications on “chronic fatigue syndrome” and “myalgic encephalomyelitis” between January 1995 and August 2000 identified 38 articles and letters. They included four papers written by mental health professionals from King’s College Medical School and their colleagues 2-5. Aside from one news item and a short report describing four patients with CFS and neurally mediated hypotension 6, all details of the research on the immunological and neurological aspects of the illness and alternative explanations were relegated to the letters page. Two editorials challenged specific issues: the domination of psychiatrists in the Royal Colleges’ Working Party and the flaws in a study on hydrocortisone 7-8.

To assess if the original papers represented ‘best evidence’, two medical colleagues assessed the quality of the publications. One was a trial of hydrocortisone on patients with normal cortisol concentrations, who arguably did not require and would not benefit from supplementation 2, while the second was a longitudinal study of “post-infectious fatigue” where only six cases actually became ill following infection 3. Both these studies emanated from the team at King’s College, London. The third - from Dutch proponents of the CBT model - reported the response of a heterogeneous sample to fluoxetine 4. Lastly, a review hypothesized that CFS was indistinguishable from conditions such as tension headaches and globus syndrome, which not only misrepresented the illness but also trivialised it 5. Arguments favouring the recategorisation of CFS as a functional somatic syndrome included the preponderance amongst women and the prevalence of child abuse, although the latter has not been documented in controlled research on CFS and certainly not outside America.

The number of non-psychiatric and psychiatric papers was compared using the two-tailed Fisher’s exact test as described elsewhere 1. The results are shown in Table 1. The short report on neurally mediated hypotension and the study on hydrocortisone were the only papers on biomedical aspects during this time 2, 6. The editorial on hydrocortisone 8 was categorised as ‘borderline psychiatric’ rather than ‘biomedical’ as it linked altered functioning of the hypothalamic-pituitary-adrenal axis to sleep deprivation, depression, obesity and starvation, but omitted to mention the literature on low cortisol in patients with chronic medical disorders. The second editorial was also categorised under ‘psychiatric’ as it focused to a significant degree on the role of psychiatrists in the Working Party set up by the Royal Colleges 7. While the latter did not promote a psychiatric aetiology, it did not focus on biomedical research. Using a broad definition of ‘psychiatric’, the analysis revealed that the Lancet published a smaller proportion on papers on non-psychiatric subjects than the four comparable American journals.

Journal(s)

Articles

Lancet vs Journal(s)

 

Psychiatric

Non-psychiatric#

 

 

N

(%)

N

(%)

P

Lancet

5

(71)

2

(29)

n.a

Am J Med, Ann Int Med, Arch Int Med, JAMA

7

(16)

38

(84)

.005

Am J Med

6

(18)

27

(82)

.011

Ann Int Med

1

(50)

1

(50)

1.000

Arch Int Med

0

(0)

6

(100)

.021

JAMA

0

(0)

4

(100)

.061

 # Broadly-defined.  For further information about  the American journals, see appendix.

Table 1. Psychiatric and non-psychiatric articles published in Lancet compared to Am J Med, Ann Int Med, Arch Int Med and JAMA between 1 Jan 1995 - 31 Aug 2000.

Of the five ‘non-biomedical’ papers, three (43%) were categorised as pro-psychiatric, because they discussed psychiatric factors or treatments, or identified psychological variables as predictors in studies which did not include immunological or neurological factors 3-5. An analysis of these papers shows that the Lancet accepted a greater proportion of pro-psychiatric articles than the American journals, although the differences failed to reach significance (see Table 2).

Journal(s)

Articles

Lancet vs Journal(s)

 

Pro-psychiatric

Other

 

 

N

(%)

N

(%)

P

Lancet

3

(43)

4

(57)

n.a

Am J Med, Ann Int Med, Arch Int Med, JAMA

5

(11)

40

(89)

.064

Am J Med

4

(12)

29

(88)

.088

Ann Int Med

1

(50)

1

(50)

1.000

Arch Int Med

0

(0)

6

(100)

.192

JAMA

0

(0)

4

(100)

.236

Table 2. Pro-psychiatric and other articles published in Lancet compared to Am J Med, Ann Int Med, Arch Int Med and JAMA between 1 Jan 1995 - 31 Aug 2000.

The apparent preference for papers focused on psychological issues has increased since the early nineties, as shown by the analysis of publications from 1990-1995.

In this time period, we found 34 items consisting of 4 studies, one viewpoint article (discussing the psychiatric aspects of the illness) and 20 letters and short reports. The four studies included two epidemiological surveys (from King’s College, London), one report on magnesium and one article discussing the evidence of immune activation found in strictly-defined CFS. The letters addressed additional issues such as the possible role of HHV-6 reactivation, Gilbert’s syndrome as well as psychological issues and neurally mediated hypotension. Altogether, the publications covered a wide range of opinions but in terms of original research, there appears to have been a shift in editorial policy sometime during or after 1992.

Editorial bias in recent years

Since this analysis and subsequent correspondence with the editor, the situation has not changed. The Lancet carried a purely descriptive account summarizing the report from a Working Group to the Chief Medical Officer 9 but there was no original research on neurological, virological or immunological aspects of the illness. However, a paper on the efficacy on CBT was published in 2001 10.

That same year, the editor also chose to publish a letter from Mouterde 11 which not only claimed that ME “is unknown in Europe outside the UK”, but also suggested that it might be a result of cultural factors, rather than disease. Aside from the fact that the Lancet had published a report on patients from Iceland a few years earlier 12, a search on MEDLINE would have revealed a number of recent articles from Belgium, the Netherlands, Italy and Spain. Three letters noting the existence of ME outside the UK were published, but there was a notable delay 13-15.

Papers on non-psychiatric subjects which were rejected by the Lancet during this period but subsequently published elsewhere include a study showing an abnormal peripheral cholinergic vascular response in CFS 16, a report of increased resting energy expenditure 17, and research letters with findings relating to the activity of two antiviral pathways 18 and metabolic changes in the brain 19. The latter was also rejected by the BMJ.

Discussion

The lack of balance in the coverage of CFS in the Lancet and BMJ disempowers doctors who rely on these journals to provide them with high quality medical information. While both publications have focused their attention firmly on the psychiatric behavioural aspects of the illness, experts around the world have been discussing the limitations of the current case definitions and the issue of subgroups among CFS patients. Many readers of the British journals will not be aware of this. Indeed, many will not be aware of the evidence undermining the psychiatric explanations being promoted in the journals.

The most popular of these is the cognitive-behavioural model developed by Professor Simon Wessely (King's College, London) and Professor Michael Sharpe (now at Edinburgh). This assumes that whilst CFS/ME may be triggered by an external agent such as a pathogen, it is primarily perpetuated by psychosocial factors and/or physical inactivity. The proponents of this theory have claimed that the resulting deconditioning induces fatigue at increasingly lower levels of exertion which is then wrongly attributed to ‘physical’ disease. Patients become helpless, hopeless and depressed, which further exacerbates the fatigue and worsens the depression.

The cognitive-behavioural model assumes that all patients share the same maladaptive beliefs and behaviours. However, the literature shows that most patients use multiple coping strategies, that there is no universal response to CFS and that the model only applies to a relatively small proportion of this patient population 20, 21.

As the following illustrates, the failure to consider issues such as definition and subgroups, combined with the selective discussion of evidence, has given a totally misleading view of the illness and its treatment.

a) The influence of definitions and the assumption of homogeneity. The Oxford guidelines used by British psychiatrists select a heterogeneous population, including a significant proportion of people suffering from psychological problems 22. Meanwhile, most of the evidence of immune deregulation and neurological abnormalities has been found in more strictly-defined samples 23, 24. The issue of definition is important because it means that research on one population may not be applicable to another. We simply cannot continue to assume, as the proponents of the cognitive-behavioural model do, that the documented differences between these groups are of no clinical significance.

b. The influence of definitions on treatment. According to the team from King’s College Hospital and their colleagues, a rehabilitation programme involving CBT and graded exercise (GET) can help every patient with CFS. However, all the successful trials have been conducted on heterogeneous samples, selected using broad, and sometimes incomplete criteria 22. Moreover, none of these studies assessed the effect on somatic symptoms, or the benefits in the subsets with evidence of immune activation or muscle disease. Suffice it to say, the trials on more strictly-defined samples have been far less impressive 25.

c. ‘Lumping’ and the loss of information. Although the latest guidelines require researchers to compare and contrast subgroups, few papers in the British journals have done so. The failure to differentiate subgroups does not just mean that we are losing important information on the immunological and neurological aspects of the illness. The emphasis on inactivity in patients selected with the Oxford and 1994 CDC criteria also means that we are ignoring the many people whose fatigue may be due to other factors. Accordingly, we do not know if these subgroups respond as well to graded activity as those suffering from phobic avoidance.

d. The influence of bias and the perpetuation of outdated ideas. Recent research has indicated that the degree of inactivity in this population is limited 21, 26 and that the deconditioning in CFS is no worse than that documented in other chronically ill populations 27. Moreover, an elegant study which assessed deconditioning and chronic fatigue failed to support the notion that the former plays a major role in perpetuating the latter 28. In terms of treatment, a controlled trial (from the team at King’s College Hospital) was not able to show that CBT was superior to counselling 29.

For some reason, none of these findings have been reported in the Lancet, or BMJ. Furthermore, doctors will not be aware that the one trial which included an objective measure of activity levels revealed only negligible increases following treatment 30. The data were omitted from the paper in the Lancet 10, but came to light in a report on the trial written for the Dutch minister for Health (see footnote). They support the view that CBT may have helped patients to pace themselves better and that it is premature to attribute the documented improvements to graded exercise.

To summarise, the two main journals in the UK seem to have an uncritical approach to the cognitive-behavioural model; they often accept research which only partly adheres to the internationally accepted criteria, and publish views based more on assumptions and speculation than data from well-defined groups and objective measures. Hence many accounts continue to emphasize the role of inactivity, without any evidence that this is a factor perpetuating the illness. And treatments are recommended without evidence that it helps those with symptoms other than fatigue and emotional distress. Finally, those wishing to express alternative views are, at the time of writing, still relegated to the correspondence section.

The recent guidelines on good publication practice by the Committee on Publication Practice (COPE 2003) state that “editors’ decisions to accept or reject a paper for publication should be based only on the paper’s importance, originality, and clarity, and the study’s relevance to the remit of the journal” 31 We should like the editors of all mainstream British journals to abide by these guidelines in relation to CFS and to respond appropriately where there are concerns.

Biased editorial policies are not benign. The selective discussion of the research means that many doctors may have offered their patients an explanation and treatment options which are neither appropriate - nor helpful. Editorial freedom is important but should not undermine the scientific process. Moreover, it should not do harm.

Conclusion

Like the BMJ, the Lancet appears to have denied its readers the opportunity to make informed decisions based on the existing scientific literature. Instead, it has published papers that present a limited view of the illness, the research, and the experiences of physicians who care for these patients.

Readers expect the Lancet to provide reasonably impartial and authoritative information on all diseases. However in terms of CFS, it has failed to live up to its reputation.

References

1. Goudsmit E and Stouten B. Editorial bias in the British Medical Journal. JCFS 2004; 12, 4; 47-59 (available online from November 2005).

2. Cleare AJ, Heap E, Malhi GS, Wessely S, O'Keane V, Miell J. Low-dose hydrocortisone in chronic fatigue syndrome: a randomised crossover trial. Lancet 1999; 353: 455-458.

3. Wessely S, Chalder T, Hirsch S, Pawlikowska T, Wallace P, Wright DJ. Postinfectious fatigue: prospective cohort study in primary care. Lancet 1995; 345: 1333-1338.

4. Vercoulen JH, Swanink CM, Zitman FG, et al. Randomised, double-blind, placebo-controlled study of fluoxetine in chronic fatigue syndrome. Lancet 1996; 347: 858-861.

5. Wessely S, Nimnuan C, Sharpe M. Functional somatic syndromes: one or many? Lancet 1999; 354: 936-939.

6. Rowe PC, Bou-Holaigah I., Kan JS, Calkins H. Is neurally mediated hypotension an unrecognised cause of chronic fatigue? Lancet 1995: 345, 8950: 623-624.

7. Anonymous. Frustrating survey of chronic fatigue. Lancet 1996; 348: 971.

8. Jeffcoate WJ. Chronic fatigue syndrome and functional hypoadrenia- fighting vainly the old ennui. Lancet 1999; 353: 424-425.

9. Clark C, Buchwald D, Macintyre A, Sharpe M, Wessely S. Chronic fatigue syndrome: a step towards agreement. Lancet 2001; 359: 97-98.

10. Prins JB, Bleijenberg G, Bazelmans E, et al. Cognitive-behaviour therapy for chronic fatigue syndrome: a multicentre randomised controlled trial. Lancet 2001; 357: 841-847.

11. Mouterde O. Myalgic encephalomyelitis in children. Lancet 2001; 357: 562.

12. Hyde B and Bergmann S. Akureyri Disease (Myalgic Encephalomyelitis), forty years later. Lancet 1988; 2: 1191-1192.

13. Speight N, Franklin A. Does myalgic encephalomyelitis exist? Lancet 2001; 357: 1890.

14 Grossman ER. Does myalgic encephalomyelitis exist? Lancet 2001; 357: 1889-1890.

15. Van Houdenhove B. Does myalgic encephalomyelitis exist? Lancet 2001; 357: 1889.

16. Spence V, Khan F, Belch JFF. Enhanced sensitivity of the peripheral cholinergic vascular response in patients with chronic fatigue syndrome. Am J Med 2000; 108: 736-739.

17. Watson WS, McMillan DC, Chaudhuri A, Behan PO. Increased resting energy expenditure in the chronic fatigue syndrome. JCFS 1998; 4, 4: 3-14.

18. Gow JW, Simpson K, Behan PO, Chaudhuri A, McKay IC, Behan WM. Antiviral pathway activation in patients with chronic fatigue syndrome and acute infection. J Infect Dis 2001; 33; 12: 2080-2081.

19. Chaudhuri A, Condon BR, Gow JW, Brennan D, Hadle DM. Proton magnetic resonance spectroscopy of basal ganglia in chronic fatigue syndrome. Neuroreport 2003; 14: 2: 225-228.

20. Saltzstein BJ, Wyshak G, Hubbuch JT, Perry JC. A naturalistic study of the chronic fatigue syndrome among women in primary care. Gen Hosp Psychiatry 1998; 20, 5: 307-316.

21. Van der Werf SP, Prins JB, Vercoulen JHMM, van der Meer JWM, Bleijenberg G. Identifying physical activity patterns in chronic fatigue syndrome using actigraphic assessment. J Psychosom Res 2000; 49, 5: 373-379.

22. Friedberg F. A subgroup analysis of cognitive-behavioral treatment studies. JCFS 1999; 5, 3-4: 149-159.

23. Landay AL, Jessop C, Lennette ET, Levy JA. Chronic fatigue syndrome: clinical condition associated with immune activation. Lancet 1991; 338: 707-712.

24. Costa DC, Tannock C, Brostoff J. Brainstem perfusion is impaired in patients with chronic fatigue syndrome. QJM 1995; 88: 767-773.

25. Lloyd AR, Hickie I, Brockman A, et al. Immunologic and psychologic therapy for patients with chronic fatigue syndrome: a double-blind, placebo-controlled trial. Am J Med 1993; 94: 197-203.

26. Sisto SA, Tapp WN, LaManca JJ, et al. Physical activity before and after exercise in women with chronic fatigue syndrome. QJM 1998; 91, 7: 465-473.

27. De Becker P, Roeykens J, Reynders M, McGregor N, De Meirleir K. Exercise capacity in chronic fatigue syndrome. Arch Intern Med 2000; 160: 3270-3277.

28. Bazelmans E, Bleijenberg, G, van der Meer JWM, Folgering, H. Is physical deconditioning a perpetuating factor in chronic fatigue syndrome? A controlled study on maximal exercise performance and relations with fatigue, impairment and physical activity. Psych Med 2001; 31: 107-114.

29. Ridsdale L, Godfrey E, Chalder T, et al. Chronic fatigue in general practice: is counselling as good as cognitive behaviour therapy? A UK randomised trial. Br J Gen Pract 2001; 51: 462, 19-24.

30. Van Essen M, de Winter LJM. Cognitieve gedragstherapie by het vermoeidheidssyndroom (cognitive behaviour therapy for chronic fatigue syndrome). Report from the College voor Zorgverzekeringen. Amstelveen: Holland. June 27th, 2002. Bijlage B. Table 2. p. 57.

31. Committee on Publication Ethics. Guidelines on Good Publication Practice (Section 8.1). www.publicationethics.org.uk. Accessed 19.9.04

32. Institute for Scientific Information. 2000 Journal Citation Reports. Summary List: Medicine, General and Internal. Philadephia, PA: Institute for Scientific Information 2001.

Footnote

Averaged actometer scores for the CBT group and ‘no intervention’ controls at baseline were 67.9 and 64.9 respectively. After treatment, they were 68.8 and 67.2 and at follow-up, they were 72.2 and 68.7. The scores of the second comparison group (support from fellow patients) were 65.8 at baseline, 62.4 after treatment and 63.8 at follow-up. Standard deviations ranged from 20.4 to 27.7.

Appendix

To assess whether there was evidence of a pro-psychiatric bias, we compared the two British journals with similar publications listed in the Journal Citation Reports for 2000 (32). We limited the analysis to the top ten journals in the ‘Medicine, General and Internal’ category, sorted by impact factor. This category includes the Lancet (ranked 3) as well as the BMJ (ranked 7). The highest ranked journal, the New England Journal of Medicine, did not publish any articles on CFS during this period so could not be included. This was also the case for two other journals in the top ten: Medicine and the Proceedings of the Association of American Physicians. The Annual Review of Medicine published a review but no original research.

Abbreviations and JCR rank of American journals

Am J Med American Journal of Medicine (8)

Ann Int Med Annals of Internal Medicine (4)

Arch Int Med Archives of Internal Medicine (6)

JAMA Journal of the American Medical Association (2)

Acknowledgements

This paper was written with help from two physicians who wish to remain anonymous. We also wish to thank Drs. Trudie Doorduin for her comments on earlier drafts.