|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
© 2003 American Society for Clinical Oncology
In Reply:Mayo Clinic, Rochester, MN We are appreciative of Dr Cella’s insightful comments on our article in the December 1, 2002 issue of the Journal of Clinical Oncology.1 Cella poses the important question, "what do global quality-of-life [QOL] questions really measure?" We acknowledge that a better trial design might have been to assess QOL before any treatment results were known to the patient. However, all patients in the QOL analysis had at least stable disease after one cycle of therapy and reported their follow-up QOL scores before possessing subsequent knowledge of treatment effect. The patients with rapid disease progression did not have follow-up QOL scores and were excluded from the analysis, as mentioned in the article.1 Dr Cella is correct in pointing out that positive results of tumor response data may indeed impart a positive effect in patients, which might translate into positive reports of various aspects of QOL. However, a great many other factors may also influence patient QOL scores, and it is really little more than conjecture to speculate what aspects of a patient’s experience have led to a change in reported QOL scores, be they global or otherwise. Indeed, the "do something" rule that Dr Cella proposes to impart improvement in QOL scores could have affected any number of the individual aspects of QOL, not just the global scores. It is unique to QOL research that we question the face validity of a result that indicates a group of patients telling us they are feeling better. Just as it was once thought to be absurd that anyone other than clinicians should rate patient pain, it is an indication of the still-evolving acceptance of QOL assessment in oncology that we still question the credibility of a group of patient’s ability to be able to accurately report an improved overall QOL. Dr Cella’s letter raises a central question in assessing patient QOL: who is the better arbiter of a patient’s overall QOL—a methodologist’s artificial summation of various aspects of QOL into a summated score, or a patient’s inherent gestalt representation of overall QOL obtained by weighing the various aspects internally? Although we firmly believe it is important to measure all aspects of QOL in any clinical trial that may be important to the patient, the disease process, and the treatment intervention and side effects, we see no justification, either in our results or in Dr Cella’s posited "do nothing" explanation, for eschewing an omnibus assessment of patient global QOL. Dr Cella suggests that the singular statistically significant difference in component QOL scores being the emotional scale indicates that patients might have weighed that construct heavily in their overall QOL assessment. It is just as plausible, however, that the numerous borderline statistically significant effects reported in our Table 4 combined to produce a significant effect when patients inherently combined these issues into their global rating. The statistical significance is somewhat secondary, as we stated in our article, in that scores for appetite, nausea, vomiting, constipation, insomnia, and pain all showed differences from baseline to treatment phase of more than 10 points — our clinically meaningful cutoff. The small sample size could well be the most important reason for the lack of statistical significance. More important, however, is the fact that there were a number of component QOL scores that showed clinically significant improvement that the patient could have weighed, on an individual basis, into their report of an improved global QOL. This would argue for the importance of having an overall measure of patient QOL that is patient-based rather than mathematically artificial, which may or may not agree with the inherent weighting that a patient may make in constructing a global assessment. Further, the patient may take other aspects into account that are not specifically measured among all the subcomponents of a QOL instrument. In other words, to a patient, their global QOL assessment may indeed be something entirely different from a mere summation of all the parts. We thank Dr Cella for pointing out the two typographical errors in the article. The clinically significant effect used in our study was indeed one half of a standard deviation (SD), as has become a standard in our research.2 The numerical result for the "Diarrhea" score for non–complete response or non–partial response patients in Table 4 should have read 42, not 32 as originally published. In summary, Dr Cella provides good food for thought and one possible explanation for variability of correlation among component QOL scores and global QOL. There are myriad other possible alternative explanations. Collectively, however, the discussion makes a strong argument for the inclusion of both global and relevant component QOL assessments so that the patient’s experience may be captured in sufficient detail while allowing for ready synthesis of the results by a global assessment. Clearly, the inclusion of the global QOL measure answered the question oft-asked by investigators when faced with a myriad of potentially conflicting individual component scores; namely, "how do we synthesize all this information into a singular result?" It would seem appropriate to use the global score as the primary piece of evidence for whether or not the patients felt, on the whole, that their QOL changed over the life of the study. REFERENCES
1. Hobday TJ, Kugler JW, Mahoney MR, et al: Efficacy and quality-of-life data are related in a phase II trial of oral chemotherapy in previously untreated patients with metastatic colorectal carcinoma. J Clin Oncol 20:4574–4580, 2002 2. Sloan J, Symonds T, Vargas-Chanes D, et al: Practical guidelines for assessing the clinical significance of health-related quality of life changes within clinical trials. Drug Inf J 37:23–31, 2003
Related Article
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2003 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
