|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
Journal of Clinical Oncology, Vol 25, No 34 (December 1), 2007: pp. 5345-5347 © 2007 American Society of Clinical Oncology. DOI: 10.1200/JCO.2007.13.3330
Should Patient-Reported Outcomes Be Mandatory for Toxicity Reporting in Cancer Clinical Trials?School of Nursing, University of Pennsylvania, Philadelphia, PA There is mounting evidence that patient-reported outcomes (PROs) provide significantly more toxicity and symptom data than clinician or Common Terminology Criteria of Adverse Events (CTCAE) assessments.1,2 The Food and Drug Administration recently defined a PRO as a measurement of any aspect of a patient's health status that comes directly from the patient, without interpretation of the patient's responses by a clinician or research associate.3 Basch et al have pioneered the adaptation of CTCAE into PRO measurements. In this issue, they discuss the feasibility of long-term toxicity monitoring using an electronic, patient self-reported, modified version of the CTCAE.4 They have carefully laid the groundwork by adapting the CTCAE symptom items into lay language and selecting salient items particular to the patients studied with thoracic malignancies receiving chemotherapy. This report documented that the CTCAE-PRO version is indeed highly feasible across sex, age, education level, baseline performance status, disease type and stage, or chemotherapy regimen.5 The work of Basch et al4 is more than just an interesting exercise. It is, in fact, the beginning of a process that may change the way in which toxicity in cancer clinical trials is measured and reported. The CTCAE has been a useful tool; yet, as with most tools, improvements are desired. There are several flaws in the CTCAE that have long been recognized, and there is recent progress in addressing some of these issues. One issue is the inability to summarize severity of both acute and long-term toxicity. A novel method to better capture the burden of such toxicity experienced by patients on clinical trials has recently been published.6 Another drawback to the CTCAE is that there is currently no standardized training for the implementation of this tool. This is highly problematic because the grading of toxicities is highly variable as a result of the differences in training and education related to patient assessments of those who use the tool (such as research associates, only some of whom may be nurses and many of whom may be from various, even nonmedical, backgrounds). In addition, the way in which CTCAE is recorded varies among institutions. Whereas some institutions may have the research associate meet with and assess the patient, other institutions have the research associates abstract the medical record. This latter approach means that the research associate must interpret the clinician's own interpretation of the patient's experience. Each iteration takes us further from the scientific rigor of assessing the patient's actual experience with toxicities. Variations in interpretation of the CTCAE also occur as a result of numerous ambiguities in wording. For example, fatigue can be recorded as mild, moderate, severe, or disabling. Such terms have been shown to elicit high variability in interpretation. Furthermore, common patient complaints such as "I feel worn out most of the time" would require the rater to explore this complaint using the wording of the CTCAE fatigue item to ensure that the rater is capturing the patient experience and recording that experience in a consistent manner. However, this rarely, if ever, occurs. Neither clinicians nor research associates use the CTCAE as a direct assessment tool, which requires reviewing the list of pertinent symptoms with the patient and using the wording of the CTCAE. The CTCAE was never developed to be used in this manner and would be prohibitively time consuming for clinician and patient. Hence, current use of the CTCAE necessitates interpretation of the patient's toxicity experience and, in most cases, additional interpretation of the assessment performed by the clinician.
The issues of variability in assessing patient toxicities using standardized clinician-rated toxicity criteria have been well documented. A well-designed study reported by Brundage et al7 of inter-rater reliability of the National Cancer Institute of Canada Clinical Trials Group toxicity scale and the WHO toxicity scale found modest levels of inter-rater agreement when the assessment was related to a laboratory test and low to poor agreement when the assessments were clinical. Seven highly experienced data managers separately interviewed 12 simulated patients (actors who delivered scripted health complaints) and scored their toxicities. Inter-rater agreement (agreement between multiple raters of the same patient) ranged from 0.50 to 1.00 ( In addition, studies have shown that there is not only modest to poor agreement among raters of patient toxicities, but that there is also poor agreement among patients and raters of symptoms.2,9 The more subjective the symptom is, the lower the clinician-patient agreement. For example, it has been shown that clinician assessments using standard toxicity ratings of erectile dysfunction are no better than chance in predicting patient self-report of the same symptom.10 This discordance raises an important question. If there is low concordance between raters and patients of patients' toxicity experience, whose rating counts? Over more than a decade, it has become widely accepted by the medical community that the most valid measure of patients' pain is their own report of pain.11 It seems a bit baffling as to why there is a lag in the recognition by the medical research community and authoritative agencies that other equally subjective symptoms (such as nausea, fatigue, and sexual satisfaction) are equally well measured by a PRO. For any symptom for which we do not have a scan, laboratory test, or other reliable biomarker or for any symptom for which the patient experience is not concordant with the testing we currently have available (eg, xerostomia v salivary flow), the only truly valid measure is the self-report of the person experiencing the symptom.12 For those who may argue that patient self-reports are soft end points, let us note that PROs have been found to be independent prognosticators for several important traditional end points such as survival or disease-free survival across a number cancer sites including non–small-cell lung cancer,13 esophageal cancer,14 cancer of the cervix,15 metastatic breast cancer,16 and advanced bladder cancer.17 Furthermore, for PROs to be used in large clinical trials they must have undergone rigorous psychometric testing. The US Food and Drug Administration PRO Guidance delineates the meticulous methodology for the validation of PRO measures. However, neither the CTCAE version 3.0 nor any earlier version has ever undergone rigorous systematic testing. We have already demonstrated the low inter-rater and rater-patient agreement, yet other concerns arise with a lack of psychometric validation. For most CTCAE toxicities, we have no formal evidence that what separates grade 1 toxicity from grade 2 and 3 toxicity is clinically significant or has meaning relative to the patient experience. As a member of both the Common Toxicity Criteria version 2.0 and CTCAE version 3.0 development committees, I can attest to the fact that the categorization of symptoms or toxicities into each grade level was primarily developed by expert opinion rather than by systematic evaluation of the evidence. In defense of this approach, much of the data that would be needed to validate the grading of toxicities does not yet exist. What is needed is a large government-funded initiative to validate this important tool. Furthermore, objective toxicities should be measured objectively with laboratory tests and procedures and scored using clinician or research associate ratings; however, the time has come to recognize that subjective symptoms should be measured with validated PRO instruments or a patient-reported, validated version of CTCAE. Basch et al4 have brought progress toward providing the latter solution. By adapting CTCAE subjective symptoms to a patient self-report, we eliminate the chance for rater misunderstanding, interpretation, and filtering of the patients' experience of a symptom. However, this is indeed a first step. As with all other scientifically acceptable PRO measures, the CTCAE-PRO version has only been adapted for a limited number of symptoms. More importantly, the PRO version will need to be judged by the same psychometric criteria as any other PRO measure to be used in a clinical trial. However, before we completely reinvent the wheel, let us recognize that tremendous work has gone into the development and testing of instruments with which to measure patient-reported toxicities. Although many of these instruments have been useful in particular settings and some have demonstrated high validity and reliability, the lack of a common lexicon across all cancer clinical trials in the way we use the CTCAE is a major scientific drawback. However, there are ongoing efforts to develop a common lay lexicon of symptoms focused on adapting validated quality-of-life and symptom measures. The work of Basch et al may benefit from a collaboration with the National Institutes of Health–funded Patient-Reported Outcomes Measurement Information System (PROMIS) initiative. The goals of the PROMIS network seem congruent with this discussion insofar as they seek to develop and test a large bank of items measuring PROs and to create a computerized adaptive testing system that allows for efficient, psychometrically robust assessment of PROs in clinical research. PROMIS uses item response theory to generate a list of patient self-report questions based on initial cues. The purpose was to develop a more fluid, yet consistent, measurement system for PROs, but it is not clear whether the intent of PROMIS was ever to become the subjective patient-report companion of the CTCAE. A parallel process with slightly different aims may be either necessary or redundant. A collaboration among those developing a lay version of the CTCAE and PROMIS may help to avoid redundancy of effort and a drain on the limited funding environment. The thought of the energy and enthusiasm of these groups coming together to explore such a collaboration would be exciting and may help to speed progress in this area. So should PROs be mandatory for toxicity reporting in cancer clinical trials? When the end point is subjective, absolutely. And Basch and colleagues have begun the process of formatting the CTCAE to make psychometric testing of the PRO version possible. High five to Basch and colleagues! AUTHOR'S DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST The author(s) indicated no potential conflicts of interest. REFERENCES 1. Parliament M, Danjoux C, Clayton T: Is cancer treatment toxicity accurately reported? Int J Radiat Oncol Biol Phys 11:603-608, 1985[Medline] 2. Varricchio CG, Sloan JA: The need for and characteristics of randomized, phase III trials to evaluate symptom management in patients with cancer. J Natl Cancer Inst 94:1184-1185, 2002 3. US Department of Health and Human Services: Guidance for industry: Patient-reported outcomes measures—Use in medical product development to support labeling claims, 2006. http://www.fda.gov/cder/guidance/5460dft.pdf 4. Basch E, Iasonos A, Barz A, et al: Long-term toxicity monitoring via electronic patient-reported outcomes in patients receiving chemotherapy. J Clin Oncol 25:5374-5380, 2007 5. Basch E, Artz D, Iasonos A, et al: Evaluation of an online platform for cancer patient self-reporting of chemotherapy toxicities. J Am Med Inform Assoc 14:264-268, 2007 6. Trotti A, Pajak T, Gwede CK, et al: TAME: Development of a new method for summarising adverse events of cancer treatment by the Radiation Therapy Oncology Group. Lancet Oncol 8:613-624, 2007[CrossRef][Medline] 7. Brundage MD, Pater JL, Zee B: Assessing the reliability of two toxicity scales: Implications for interpreting toxicity data. J Natl Cancer Inst 85:1138-1148, 1993 8. Kaba H, Fukuda H, Yamamoto S, et al: Reliability at the National Cancer Institute-Common Toxicity Criteria version 2.0. Gan To Kagaku Ryoho 31:1187-1192, 2004[Medline] 9. Movsas B, Scott C, Langer C, et al: Randomized trial of amifostine in locally advanced non-small-cell lung cancer patients receiving chemotherapy and hyperfractionated radiation: Radiation Therapy Oncology Group trial 98-01. J Clin Oncol 23:2145-2154, 2005 10. Bruner DW, Scott C, Lawton C, et al: RTOG's first quality of life study–RTOG 90-20: A phase II trial of external beam radiation with etanidazole for locally advanced prostate cancer. Int J Radiat Oncol Biol Phys 33:901-906, 1995[CrossRef][Medline] 11. Jensen MP: The validity and reliability of pain measures in adults with cancer. J Pain 4:2-21, 2003[CrossRef][Medline] 12. Bruner DW: Outcomes research in cancer symptom management trials: The Radiation Oncology Group (RTOG) Conceptual Model. J Natl Cancer Inst (in press) 13. Langendijk H, Aaronson N, de Jong J, et al: The prognostic impact of quality of life assessed with the EORTC QLQ-C30 in inoperable non-small cell lung carcinoma treated with radiotherapy. Radiother Oncol 55:19-25, 2000[CrossRef][Medline] 14. Fang F, Tsai W, Chiu H, et al: Quality of life as a survival predictor for esophageal squamous cell carcinoma treated with radiotherapy. Int J Radiat Oncol Biol Phys 58:1394-1404, 2004[CrossRef][Medline] 15. Monk B, Huang H, Cella D, et al: Quality of life outcomes from a randomized phase III trial of cisplatin with or without topotecan in advanced carcinoma of the cervix: A Gynecologic Oncology Group study. J Clin Oncol 23:4617-4625, 2005 16. Efficace F, Biganzoli L, Piccart M, et al: Baseline health-related quality-of-life data as prognostic factors in a phase III multicentre study of women with metastatic breast cancer. Eur J Cancer 40:1021-1030, 2004[CrossRef][Medline] 17. Roychowdhury D, Hayden A, Liepa A: Health-related quality-of-life parameters as independent prognostic factors in advanced or metastatic bladder cancer. J Clin Oncol 21:673-678, 2003
Related Article
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2007 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
statistic, values of 1.0 indicate perfect agreement) in laboratory-based categories and from –0.04 to 0.82 for clinically based categories. There have been no reported studies of inter-rater reliability of the CTCAE version 3.0; however, a small single-institution Japanese study evaluated the reliability of the National Cancer Institute Common Toxicity Criteria version 2.0. Five research associates independently reviewed medical records from 17 patients and graded toxicities. As in the study by Brundage et al,
