Letters

Psychological coping and cancer

Study results should not have been dismissed

Search strategy used is inadequate

Authors' reply

Study results should not have been dismissed

EDITORThe review by Petticrew et al is flawed.¹

Firstly, they do not compare like with like and ignore differing methods. Different instruments assess coping styles differently and may not be comparable. Thus conclusions cannot be drawn convincingly. Out of 28 studies, 13 had less than three years' follow up (shortest eight weeks). However, in relation to early breast cancer less than five years of follow up produces inconclusive results.

Secondly, our study was relegated to the realms of irrelevance by saying that the recent large UK study (n=578), while of higher quality, reported mixed findings: helplessness or hopelessness predicted recurrence when patients with high and low scores were compared but not when it was the predominant coping style.² The main importance of our clearly stated findings concerning helplessness and hopelessness was missed. Mixed findings were not reported.

Although predominant coping style was arbitrarily defined, helplessness or hopelessness is robust across many studies. A high score on this subscale is a valid measure, reflecting a response amenable to psychological intervention. This finding was crucially important for encouraging adequate provision of psychological care in oncology, as was the small, cautiously interpreted tendency to poorer outcome found in severely depressed patients. Patient responsibility (or blame) is removed, and the onus shifts to the medical profession to identify and provide help for those at risk.

Dismissal of our studywhich used rigorous methods, controlled for known prognostic variables, was sufficiently large, had a long follow up, and found a survival effectdoes the research no service. Better to conclude that many poor studies fail to show any effect of coping response on survival, but high quality studies provide evidence suggesting that further good research is needed. Meanwhile the adverse effect on outcome of a helpless or hopeless coping response clearly argues for the provision of better psychological resources.

Maggie Watson, consultant clinical psychologist. 
Janis Davidson-Homewood, research assistant. 
Royal Marsden Hospital NHS Trust, Sutton, Surrey SM2 5PT

Jo Haviland, statistician. 
Judith Bliss, team leader. 
Institute of Cancer Research, Clinical Trials and Statistics Unit, Sutton, Surrey SM2 5NG

Search strategy used is inadequate

EDITORThe systematic review by Petticrew et al on the influence of psychological coping on survival in and recurrence of cancer seems comprehensive, but the search strategy used is inadequate.¹ Searches that rely on keywords in medical databases will miss many important articles.² More can be retrieved by citation searching, but often no alternative exists to hand searching journals.

The end result is that Petticrew et al have identified 26 prospective studies, but we are aware of 12 additional prospective studies in this field (an extra 1425 patients) (references available from us on request). Of course, these studies vary in quality and use different markers of coping (predominantly locus of control measures), but they nevertheless require appraisal in a truly systematic review. The authors have excluded at least 12 non-prospective studies, but retrospective studies can add valuable information to the debate.

Furthermore, we cannot understand why Petticrew et al consider an early assessment of coping style to be a measure of quality, unless it is a proxy for length of follow up. If they had adopted more stringent methodological criteria, as they say, it might have been possible for studies with false negative results to be excluded as much as those with false positive results. Petticrew et al note that only four studies identified over 200 patients, but in fact seven studies have done this. The missing ones are by Funch and Marshall, Holland et al, and Stavraky.^3-5

It is inappropriate either to force superficial positive thinking or to blame patients for not thinking positively. Methodological rigour is required if research is to contribute to this debate.

Alex Mitchell, lecturer in psychiatry. 
a.j.mitchell{at}leeds.ac.uk

Manoj Kumar, consultant in psycho-oncology and liaison psychiatry. 
University of Leeds, St James's University Hospital, Leeds LS9 7TF

Authors' reply

EDITORWatson et al point out that coping can be measured in different ways. This is why our review included studies using any measure; overall, we found the evidence unconvincing. As regards the importance of helplessness or hopelessness, these studies give little convincing evidence that it significantly influences survival or recurrence. We agree, however, that further good research is needed.

Mitchell and Kumar cannot understand why we consider an early assessment of coping style to be a measure of quality. Yet the early measurement of coping style is a basic criterion for causality: if a factor is postulated to be causal, logically it has to operate before the outcome occurs. Measuring coping style close to diagnosis seems the most appropriate way to assess this factorthe longer after diagnosis, the greater the possibility that coping style may be a result of disease progression, or treatment, rather than an effect. This is also why we included only prospective studies.

Mitchell and Kumar also suggest that we missed 12 studies, but we had in fact seen and excluded most of the studies on their list. These investigated factors such as social support and general wellbeing, while our review addressed coping styles. Two studies initially seemed to be relevant, but one reported insufficient data for inclusion and the other investigated rationality and anti-emotionality and so does not seem to be a study of coping. They also imply that our search relied solely on keywords, which it did not. Mitchell and Kumar's claim that we missed a significant portion of the literature is quite simply wrong.

Mark Petticrew, associate director. 
MRC Social and Public Health Sciences Unit, Glasgow G12 8RZ

Ruth Bell, lecturer. 
Department of Epidemiology and Public Health, Medical School, University of Newcastle on Tyne, Newcastle on Tyne NE2 4HH

Duncan Hunter, assistant professor. 
Community Health and Epidemiology, Abramsky Hall, Queen's University, Kingston, ON, Canada K7L 3N6