|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
Journal of Clinical Oncology, Vol 25, No 30 (October 20), 2007: pp. 4853-4854 © 2007 American Society of Clinical Oncology. DOI: 10.1200/JCO.2007.13.4569
In ReplyReference Center on Quality of Life in Oncology, Department of General and Thoracic Surgery, University Hospitals of Schleswig-Holstein, Keil, Germany
Surgical Hospital, Department of Visceral and Transplationa Surgery, University Hospital of Ulm, Ulm, Germany
School of Physical and Occupational Therapy, Department of Epidemiology and Biostatistics, Faculty of Medicine, McGill University, Montreal, Quebec, Canada First, we would like to thank Drs Coyne and Palmer for their interest in our study.1 Answers to these critical remarks will address their letter on two levels: those remarks that look at methodological issues of the results, and the conclusions that can and cannot be drawn from those results; and those remarks that focus on attitudes and claims that Coyne and Palmer suspect go along with reporting such results. Let us start with the latter. The way Coyne and Palmer state: "Encouraging cancer patients to seek psychotherapy to extend their lives is disrespectful of patients' autonomy, time, and resources" suggests that we (and other researchers in the field) tell patients that psychotherapy (correct term in this context is psychotherapeutic/psycho-oncological support) will extend their survival time. We do not know of any colleague in the field who would ever do that. We, in our daily work with cancer patients on our surgical wards, certainly would never even infer this idea. We leave that to gurus and other quacks. In fact, we do not even make claims, when in discussions with a patient, that psycho-oncological support will enhance his or her quality of life (QOL) although current scientific evidence in the literature would support such a statement. Rather, we tell patients that at our hospital psycho-oncological support is a normal part of the interdisciplinary/interprofessional care, and we certainly agree with Coyne and Palmer that improved care is an attainable goal. Our daily clinical work is surely devoted to that. Communication of trial results to the scientific community, however, is a completely different ballgame. Here the rules are very clear. You present the results; the scientific community questions these results based on methodological considerations. If we start to base our reviews on other standards, we would soon be back to the days when scientific rigor and evidence were not in vogue. In any case, let us return to those points in our trial that the scientific community—in this case Coyne and Palmer—have the right to discuss critically. Although the weak points in our trial have been addressed in the editorial by Andrykowski,2 we would like to specifically return to those raised by Coyne and Palmer. It is absolutely correct, that a survival difference was not our primary end point. When planning the study (to keep the timeframe in mind, back in 1989), we selected health-related QOL as the primary end point. Accordingly, we based our sample size calculation on health-related QOL, which of course requires a bigger sample than would be necessary to detect survival differences. When we analyzed the data, we had to start by describing the sample and looked first at survival. To our surprise, we found survival differences. This was unexpected, and frankly complicated the analysis. We had anticipated that we would find equal survival, and then were prepared to compare QOL results. Thus, the analysis became much more complicated, and in fact, working with a biostatistician from the University of Toronto and consulting other biostatistical experts at McGill University and in Germany, we could not find a scientifically sound way of combining survival with QOL outcomes. To make the groups comparable in terms of QOL, you would have to input a value for those patients who were not alive at each time point. But the scientific community has, to our best knowledge, not yet agreed on the correct way to do such imputations. The most prominent expert in the field, Dianne Fairclough,3 made suggestions about how to tackle this problem, but just to explain the methodology takes more space than is reasonable in a 3,000-word article. Our own, not very sophisticated approach, consisted of imputing a low (ie, the fifth, respectively, the 95th percentile of the range of QOL values), but more than 0 QOL value for each time point when the patient was not alive. With this rather mechanical approach, we found a significantly better QOL (and mental adjustment) in the treatment arm. But, since we think that this is neither a scientifically solid nor a philosophically satisfying way of handling the problem of QOL of patients not alive, we only incorporated these results in our report to the (German) Ministry of Research. By the way, anybody who can offer a solution to this problem is welcome to correspond with the authors. All things considered, we, therefore, elected to try and present the survival results by themselves. Coyne and Palmer complain that "...it appears that key persons were not blinded to treatment assignment, including the oncology treatment team and the patients... ." How could a patient in a randomized trial which has, as a key ethical element, informed consent for all patients be blinded to psycho-oncological treatment? Furthermore, cooperation with all ward personal was an integral part of the trial. Coyne and Palmer continue to claim that, postoperatively, the treatment group received more additional treatments which by itself may explain the survival differences. In fact, all these variables were included in the multivariate analysis (backward regression) and did not enter the overall model, otherwise we would have reported that information. And as for the crossovers (experimental group to control group, 10; control group to experimental group, 34), the analysis preserved the original randomization (intention to treat) and this is a bias against positive findings. In the last part of their remarks, Coyne and Palmer repeat some comments they made in their publication in Psychological Bulletin4 on other studies in the field and on the field itself, ultimately suggesting that this kind of research should not be done. Since there is a detailed response to this review on the way (Spiegel et al, submitted for publication), we would rather refer to that instead of being repetitive. In summary, we published the results of a study.1 As with any trial, it has its strengths and its weaknesses. We think that the weaknesses and limitations are presented, not only in the article itself, but very comprehensively in the accompanying editorial.2 We agree with most of them and accept all of them. In our opinion, Coyne and Palmer add nothing substantial to those points already addressed by Andrykowski. We agree with them that improved overall care should be a continuous goal in oncology. In fact, we think that this trial that studied the effects of psychotherapeutic/psycho-oncological support for patients with gastrointestinal cancer during their inpatient surgical treatment may serve as an example for improved care within a university hospital. We strongly disagree and reject Coyne and Palmer's allegation that this kind of research is associated with making false promises to patients in daily clinical care. AUTHORS' DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST The author(s) indicated no potential conflicts of interest. REFERENCES
1. Küchler TH, Bestmann B, Rappat S, et al: Impact of psychotherapeutic support on patients with gastrointestinal cancer undergoing surgery: 10-year survival results of a randomized trial. J Clin Oncol 25:2702-2708, 2007 2. Andrykowski MA: Survival benefits associated with provision of psychotherapeutic support to patients with gastrointestinal cancer: Lots of bang for a few bucks? J Clin Oncol 25:2644-2645, 2007 3. Fairclough D. Design and Analysis of Quality Of Life Studies in Clinical Trials (Interdisciplinary Statistics). Boca Raton, FL, Chapmann & Hall/CRC, 2002 4. Coyne JC, Stefanek M, Palmer SC: Psychotherapy and survival in cancer: The conflict between hope and evidence. Psychol Bull 133:367-394, 2007[CrossRef][Medline] Related Correspondence
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2007 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
