© The Author 2007. Published by Oxford University Press.
Health-Related Quality of Life Measurement in Symptom Management Trials
Affiliations of authors: UCLA Schools of Medicine and Public Health and Jonsson Comprehensive Cancer Center, University of California, Los Angeles, Los Angeles, CA (PAG); Department of Medicine and Division of Clinical Epidemiology, Samuel Lunenfeld Research Institute, Mount Sinai Hospital, University of Toronto, Toronto (ON), Canada (PJG)
Correspondence to: Pamela J. Goodwin, MD, Department of Medicine and Division of Clinical Epidemiology, Samuel Lunenfeld Research Institute, Mount Sinai Hospital, 1284-600 University Ave, Toronto, ON M5G 1X5, Canada (e-mail: pgoodwin{at}mtsinai.on.ca).
 |
ABSTRACT |
|---|
 Top Abstract Overview and HistoryUtility and Rationale for...Conceptual Framework--Beyond the...Instruments, Measurement, and...The Food and Drug...Key MessagesMoving ForwardReferences |
|---|
There is increasing support for the incorporation of patient-reported outcomes (PROs) into clinical trials in cancer. While the need for inclusion of measures of target symptoms in symptom management trials is clear, arguments can also be made for measurement of a broader range of symptoms, for evaluation of symptom burden, and for evaluation of health-related quality of life (HRQOL) in these trials. What is key to their inclusion is a priori selection of instruments, provision of a theoretic basis for inclusion of instruments, and a clearly described plan of analysis. The federal Food and Drug Administration (FDA) has provided guidance regarding the use of PROs (symptom and HRQOL measures) to support treatment benefit claims in product labeling. Moving forward, research is needed to address methodological issues raised by the FDA and to increase understanding of relationships among symptoms, symptom clusters, HRQOL, and other outcome measures.
|
Overview and History |
|---|
|
TopAbstractOverview and HistoryUtility and Rationale for...Conceptual Framework--Beyond the...Instruments, Measurement, and...The Food and Drug...Key MessagesMoving ForwardReferences |
|---|
Elimination of pain and suffering has been the goal of medical care since the time of Hippocrates, yet it has taken several millennia to reach the point where it has become feasible to assess these outcomes in a reliable and valid way. In the aftermath of the Second World War, the World Health Organization acknowledged that health not only was the absence of disease but also required the maximization of physical, emotional, and social well-being (1). The latter definition of health has become the foundation for contemporary conceptualization of health-related quality of life (HRQOL). These two important perspectives provide a backdrop for the workshop convened by the National Cancer Institute (NCI) Community Oncology and Prevention Trials Research Group that was held in Bethesda in October 2005. The papers that are published in this monograph provide a comprehensive update on the state of our current knowledge about the tools and strategies that are available to measure the control of symptoms and maintenance or improvement in HRQOL. The questions that brought the workshop participants together focused on the relevance or rationale for HRQOL assessment in symptom management trials; the conceptual framework for HRQOL assessment in such trials; and issues related to instruments, measurement, and analysis.
As with any developing area of research, early efforts to design and conduct symptom management trials did not always demonstrate clarity of focus or a priori specification of the patient-reported outcomes (PROs) of interest (2). Challenges inherent in the conduct of these trials, especially within the mechanism of the US clinical cooperative groups, were the scarcity of behavioral science expertise, the limited number of reliable and valid tools for the assessment of PROs, and the underappreciation of the value of inclusion of PROs (as opposed to observer ratings) in assessing trial outcomes. The multidimensional, cancer-specific HRQOL tools that were developed in the late 1980s and early 1990s were among the few reliable and valid tools available for inclusion in these trials. And, although these HRQOL questionnaires clearly included items that measured patient-reported symptoms (3,4), their scoring focused on the broad domains of global health, physical functioning, and emotional and social well-being, with the exception of some modules that were explicitly described as reporting symptoms (e.g., European Organization for Research and Treatment of Cancer Questionnaire Core 30 items [EORTC QLQ C-30]). Nevertheless, without specific scales to address the symptom being examined in a symptom management trial (e.g., vasomotor symptoms, fatigue, vaginal dryness), trial outcomes and efficacy could not be accurately measured. The substitution of a multidimensional HRQOL scale, in these circumstances, added little value, unless there was a direct impact of control of the symptom on a specific domain of HRQOL—for example, relief of pain might be hypothesized to improve physical functioning. However, in patients with very advanced cancer, this might be difficult to achieve given other symptoms that might be contributing to decrements in physical functioning. Further, until the past decade, PROs were seldom directly incorporated into the parent trial protocol (5), limiting their integration into the goals, objectives, and hypotheses of the treatment or symptom management strategy.
Fortunately, with support from the NCI and specifically the Division of Cancer Prevention, symptom management has become a priority for the health and welfare of cancer patients and survivors (6). Tools to measure various symptoms have proliferated, and there is an acknowledgment that new tools must be developed and validated alongside treatment and symptom trials (7,8). This is particularly true given the new and previously unreported side effects of cancer treatments (e.g., skin rash associated with epidermal growth factor receptor antagonists). This has also led to the earlier identification of symptoms that may impair tolerance to anticancer therapies, as well as their recognition and PRO assessment as early as the phase II test setting. With clear identification of patient-reported symptoms early in the evaluation of new therapies, phase III trials can incorporate relevant PROs as a key secondary outcome of the trial and symptom management strategies can be developed as part of cancer control efforts using these same PRO assessment tools. Whether or not to include a HRQOL outcome (domain specific or global) in symptom management trials should be driven by the expected impact of the particular symptom(s) on aspects of HRQOL. If HRQOL is incorporated in a treatment trial evaluating drug toxicity and outcomes, then a preview of the relationship may be available for those designing symptom management trials.
To this end, the 2005 Workshop heightened our awareness of the need to have clearly specified goals and objectives for the incorporation of PROs in symptom management trials. This workshop preceded what has now been widely distributed as federal Food and Drug Administration (FDA) guidance on PROs in pharmaceutical trials made available in the early part of 2006 (9). In this guidance, the FDA has elevated the voice of the patient or participant to an equivalent status of other trial outcomes, including observer ratings of symptoms or toxic effects. Importantly, the guidance recognized the need for careful planning at the earliest stages of a pharmaceutical trial to select or develop appropriate measures of PRO that have good reliability and validity, are relevant to the outcomes of the agent being testing, and are included in an a priori framework or model for the entire trial. The FDA guidance, while particularly relevant for cancer researchers, is also important for other conditions in which pain and symptom control may be the primary outcome for trials, for example, arthritis and skin and musculoskeletal disorders. In oncology pharmaceutical trials, PROs may be an important primary or secondary endpoint and may be a critical part of the drug approval process (e.g., safety of chemopreventive agents) depending on the patient population, the goal of therapy, and the conceptualization of patient benefit. However, the guidance is clear in advising prespecification of PRO endpoints with testable hypotheses and adequate sample size, accounting for missing data. In summary, PROs have come of age with this guidance as an important endpoint in the drug approval process.
In the sections that follow, we highlight and reflect upon the discussions at the Symptom Management Workshop and the papers published in this monograph. We were fortunate to have the participation of leading experts in this field, and thus, our task in this paper has not been burdensome. The work of our colleagues has facilitated our discussion of the relative importance of symptom assessment versus HRQOL as PROs and the challenges that we face in improving the design and implementation of future symptom management studies.
|
Utility and Rationale for Health-Related Quality of Life Assessment in Symptom Management Trials—What is the Added Value? |
|---|
|
TopAbstractOverview and HistoryUtility and Rationale for...Conceptual Framework--Beyond the...Instruments, Measurement, and...The Food and Drug...Key MessagesMoving ForwardReferences |
|---|
Collectively, the papers in this monograph have provided a strong rationale for inclusion of HRQOL assessment in symptom management trials. The authors point out that HRQOL is a multidimensional construct composed of at least four dimensions (physical function, psychologic function, social role function, and treatment or disease symptoms) (10). Ferrans (3) reminds us that many of the HRQOL instruments commonly used in symptom control trials (e.g., Functional Assessment of Cancer Therapy—General [FACT-G], EORTC QLQ-C30, Medical Outcomes Study Short Form 36) contain items that are symptoms; in fact, over half of the items on the FACT-G and EORTC QLQ-C30 are symptoms, underscoring the important contribution of symptoms to our current conceptualization of HRQOL. These multidimensional instruments can provide estimates of overall quality of life (QOL) in one of two ways—by summing across items in a QOL questionnaire (e.g., FACT) or by adding unique items that ask the patient to rate their perception of overall QOL (e.g., EORTC QLQ-C30). The latter approach allows patients to incorporate their personal values into the QOL assessment; such an approach may be particularly important in trials aimed at "providing comfort, such as symptom trials." Ferrans (3) points out that perceived health status is most affectd by physical functioning, while global QOL is most affected by emotional well-being, leading to the potential for lack of concordance between perceived health status and overall QOL scores.
Osoba (4) and Ferrans (3) discuss the model of HRQOL of Wilson and Cleary (11). This model links biologic and physiologic variables, symptom status, functional status, general health perceptions (integration of the preceding health concepts, as perceived by the patient), and overall QOL (satisfaction with life as a whole as evaluated by the patient). A modified model incorporates characteristics of the individual and the environment, as well as nonmedical factors. The original model suggested that influences of factors flow in a unidirectional fashion from biologic and physiologic variables through to overall QOL. Both Ferrans (3) and Osoba (4) recognize that at least some of these relationships may be bidirectional. For example, symptoms such as weakness, tiredness, and feeling depressed may reflect QOL (i.e., they are indicator variables), while many other symptoms (e.g., nausea, vomiting) are causal variables (i.e., they directly influence QOL). This is an important distinction because treatment is ideally targeted at the source of the problem rather than at downstream effects; for causal variables, treatment would be directed at the symptom, while for indicator variables, treatment would be directed at other aspects of HRQOL. Recognition of this complexity underscores the importance of including HRQOL measurement in symptom control trials when symptoms are potentially indicator variables of HRQOL.
Osoba (4) provides informative examples demonstrating the impact of symptoms (such as nausea and vomiting) on functioning, notably physical functioning. He argues that symptoms and HRQOL are interrelated and that measurement of symptoms only, without measurement of HRQOL, results in failure to understand the full impact of treatment. Although global QOL may be too far "downstream" to reflect important changes in target symptoms in many instances, Osoba (4) also provides an example of a study (temozolide in recurrent anaplastic astrocytoma) in which treatment had a similar effect on proximal symptoms and on distal global QOL. Other authors provided examples of situations in which effects of treatment on symptoms and HRQOL either converged or diverged, as well as examples in which symptom improvement was demonstrated without an associated effect on HRQOL.
Thus, HRQOL assessment adds value in symptom management by allowing a broader understanding of the impact of symptom management on a range of symptoms beyond the targeted symptom, on functioning, and on overall HRQOL. As such, HRQOL measurement allows for unexpected observations, enhances understanding of interrelationships among these variables (including exploration of whether symptoms variables are indicator variable or causal variables), and allows assessment of whether there is a net beneficial or adverse effect of treatment on HRQOL. Cella et al. (12) point out that these issues are particularly important if symptom interventions introduce unanticipated side effects or are associated with side effects that may counteract treatment benefit. They also point out that HRQOL assessment can capture the longitudinal trajectory of symptoms; allow comparisons across groups of patients with different types of cancers; and facilitate exploration of the association between symptoms, symptom severity, and performance status. They provide data illustrating many of these associations.
Cleeland (10) discusses symptom burden, a concept that "encompasses both the severity of the symptoms and the patient's perception of the impact of the symptoms." He differentiates symptom burden from HRQOL, saying that HRQOL incorporates more abstract concepts than those involved in symptomatology; however, because he includes the impact of symptoms on function in the concept of symptom burden, one cannot equate symptom burden with simple description of symptoms. Cleeland (10) argues that it is not necessary to demonstrate an improvement in generic measures of HRQOL if symptom burden is improved. He suggests that improvement of symptom burden may be a sufficient outcome in many clinical trials of symptom management, guiding decision making about use of specific treatments. Given that symptom burden incorporates the impact of symptoms (not just the severity of symptoms), this is probably reasonable. This apparent dichotomy of approaches—measurement of symptoms or symptom burden only versus measurement of HRQOL—may relate more to the ultimate use of the data (e.g., regulatory decisions versus understanding of patient experience) than to the validity of the approach as will be discussed further below.
|
Conceptual Framework—Beyond the Multivariate Model |
|---|
|
TopAbstractOverview and HistoryUtility and Rationale for...Conceptual Framework--Beyond the...Instruments, Measurement, and...The Food and Drug...Key MessagesMoving ForwardReferences |
|---|
Enhanced recognition of the complexity of relationships among symptoms and between symptoms and HRQOL is reflected in the papers by Miaskowski et al. (13), Cleeland (10), and Moinpour et al. (14). Miaskowski et al. (13) and Cleeland (10) discuss the concept of symptom clusters, a relatively new concept that seeks to understand the grouping of symptoms into reproducible clusters that may or may not share a common etiology. Early observations suggest that symptom clusters do, indeed, exist and that the presence of certain symptom clusters may place patients at increased risk of poorer outcomes, including poorer HRQOL. Miaskowski et al. (13) discuss two conceptual approaches to the identification and subsequent validation of symptom clusters, while Cleeland (10) explores the relationship between symptom clusters and symptom burden, noting that not all symptoms or symptom clusters increase or decrease together and that sets of symptoms or symptom clusters may have different temporal relationships to treatment and disease progression. Cleeland (10) also argues that symptom measures may provide information that is more indicative of treatment differences than global HRQOL measures. He cites results of the National Surgical Adjuvant Breast and Bowel Project Study of Tamoxifen and Raloxifene trial in which no differences in overall physical or mental health were identified in women receiving tamoxifen or raloxifene, whereas important differences in symptom profiles were readily apparent.
Moinpour et al. (14) tackle the difficult issue of statistical analysis of HRQOL data. In an elegant reanalysis of Southwest Oncology Group 9916, a randomized trial of two chemotherapies in advanced prostate cancer, they used latent trajectory modeling to "tease out" relationships between pain (a major symptom outcome of the trial) and HRQOL. Despite the fact that the initial analysis found no evidence for differences in HRQOL or in pain palliation between the study arms, this reanalysis demonstrated that the relationship between pain and HRQOL differed quantitatively between the two study arms and that treatment had an effect on HRQOL after controlling for pain (i.e., pain-free HRQOL). This difference was qualitative, pain-free HRQOL improving in patients receiving docetaxel and estramustine chemotherapy (associated with the best medical outcomes) and worsening in patients receiving mitoxantrone and prednisone. This group also explored individual variability in HRQOL. Although the results of these analyses supported the overall conclusions of the trial (that docetaxel and estramustine was the preferred treatment based on overall survival advantage, median time to progression advantage, and percentage of patients showing at least a 50% reduction in prostate-specific antigen), they enhance understanding of the effect of the treatment on QOL and provide additional evidence of benefit for the preferred regimen.
Finally, Watkins-Bruner (15) presents a model for incorporation of outcomes research into cancer symptom management trials, using the Radiation Therapy Oncology Group conceptual model. For example, the model guided the investigators in their analysis of six locally advanced non–small-cell lung cancer trials. Specifically, two quality-adjusted survival analyses revealed that patients older than 70 years experienced increased toxic effects. With these data, the investigators developed a symptom management trial of esophagitis. This higher level approach links clinical, humanistic, and economic outcomes.
|
Instruments, Measurement, and Statistical Analysis |
|---|
|
TopAbstractOverview and HistoryUtility and Rationale for...Conceptual Framework--Beyond the...Instruments, Measurement, and...The Food and Drug...Key MessagesMoving ForwardReferences |
|---|
As noted above, many of the commonly used HRQOL instruments include items that measure symptoms. Cella et al. (12) review the development of the FACT measurement system, emphasizing that symptoms are a component of HRQOL and demonstrating how the use of more broad-based HRQOL measures like the FACT has enhanced understanding of commonly identified cancer symptoms, including their impact on functioning, performance status, and HRQOL. Ferrans (3) compares three of the most commonly used HRQOL instruments (FACT, EORTC QLQ-C30, and Medical Outcomes Study Short Form 36). She differentiates symptoms, functioning, general health perceptions, and overall QOL, illustrating how each of these instruments addresses these attributes. She notes that global ratings of QOL may be influenced by factors such as spirituality that are not under consideration in a symptom management trial. She recommends head-to-head comparisons of instruments to enhance understanding of differences among instruments and to clarify distinctions between symptoms, functional status, general health status, and global QOL. Finally, Osoba (4) discusses the challenging problem of using HRQOL data in the clinic and outlines a model for incorporating HRQOL assessment into clinical practice, providing evidence that HRQOL measurement in clinical practice leads to improved patient outcomes. Osoba (4) also recommends that statistical analyses focus on the proportion of patients meeting some predetermined level of improvement (or deterioration) rather than on mean scores, thereby enhancing the clinical meaning of HRQOL data.
|
The Food and Drug Administration Perspective |
|---|
|
TopAbstractOverview and HistoryUtility and Rationale for...Conceptual Framework--Beyond the...Instruments, Measurement, and...The Food and Drug...Key MessagesMoving ForwardReferences |
|---|
Rock et al. (16) discuss challenges to the use of HRQOL for FDA approval of anticancer products. They state that claims of clinical benefit are made on the basis of "substantial evidence" demonstrated in "adequate and well-controlled investigations." Direct measures of treatment effect, including disease-related symptoms, are sufficient on their own to support drug approvals; however, the FDA has rarely used HRQOL data as primary evidence to support an anticancer product approval.
In 2006, the FDA released draft guidance regarding the use of PROs to support treatment benefit claims in product labeling (9). This guidance advocated for "clear, prospective delineation of measurement concepts underlying health labeling claims and outlined principles for development, validation, implementation and interpretation of reliable and accurate PRO measures." While overall QOL measures are not acceptable in product labeling claims (because they incorporate attributes that go beyond health), HRQOL measures are potentially acceptable. The focus of the paper of Rock et al. (16) is on the use of HRQOL data for product approval of anticancer drugs. Some of the issues raised may be more (or less) relevant to symptom management trials that focus on the effect of symptom control agents. Additionally, the issues raised by Rock et al. (16) have a much narrower focus than those raised by the other authors who explore and call for additional research to understand associations between symptoms, HRQOL, performance status, and other variables.
In their discussion, Rock et al. (16) review methodological shortcomings, resulting in the potential for bias, that have been associated with use of HRQOL data in anticancer drug trials. These include lack of randomization, lack of blinding, missing data, multiplicity (of hypotheses and comparisons), and failure to address intrinsic meaning of the HRQOL data. Some of these concerns, such as lack of randomization, are relatively easily overcome in symptom management trials. Others are more problematic. For example, it may be difficult or impossible to blind studies when treatment side effects, administration schedules, and routes make blinding impractical or unethical (this may be more problematic in anticancer treatment trials than in symptom management trials). It may also be very difficult to avoid missing data in palliative trials, where many patients become ill and die before full data are collected. This problem cannot be overcome through the use of proxy data, given that the FDA does not consider such data appropriate for product labeling. Many of the other limitations discussed by Rock et al. (16) can, over time, be addressed. For example, the issue of multiplicity of outcomes, hypotheses, and comparisons can, and should, be addressed through detailed planning of the statistical analysis. Research should be undertaken to address concerns relating to intrinsic meaning, including whether all relevant information is being captured, whether findings within a set of HRQOL results are consistent, and whether statistically significant findings are clinically relevant in the population studied. The call for overall consistency of endpoint results is potentially problematic—one statement "if worsening is observed in some domain(s), conclusions regarding improvement in overall QOL are not valid" is difficult to understand. For example, a pain medication may improve pain but worsen constipation—it is the net benefit that is of interest rather than consistent directionality of effects. Additionally, concordance with other endpoints, such as overall survival or response rates, may be of limited relevance in symptom management trials.
|
Key Messages |
|---|
|
TopAbstractOverview and HistoryUtility and Rationale for...Conceptual Framework--Beyond the...Instruments, Measurement, and...The Food and Drug...Key MessagesMoving ForwardReferences |
|---|
Although, at first glance, it may appear that there are inherent disagreements and contradictions in the opinions expressed in the papers included in this monograph, there is probably more agreement than disagreement among the authors. First and foremost, all authors, including the FDA, have recognized PROs, including symptom measures and HRQOL measures, to be valid as observable outcomes. Furthermore, there is broad recognition that symptoms and HRQOL are related entities, with symptom items being included in the majority of HRQOL measures. Although related entities, the effects of treatment on symptoms and HRQOL may converge, diverge, or be seen in only one, but not both, of these entities. There is general consensus that the inclusion of measures that go beyond the target symptoms—either a broader range of symptom measures or measurement of function and HRQOL—will allow broader effects of treatment to be identified and will provide richer information on patient experience and on longitudinal effects of treatment within groups of patients; this approach will also facilitate comparisons across groups of patients with different types of cancer and different medical conditions. This broader evaluation of the impact of treatments that target individual symptoms was felt to be, in general, of benefit. A range of research activities needed to move the field forward were identified. This included activities that will enhance understanding of the relationships between symptoms, functioning, HRQOL, performance status, and other important outcomes; explore relationships among symptoms, notably the existence of symptom clusters; investigate novel approaches to statistical analysis that will provide insight into these relationships; evaluate clinical meaning of results; and advance incorporation of these measures into clinical practice.
The FDA focus is, understandably, much narrower than that of many of the other authors. Their focus is primarily on what is needed for drug approval. Although the FDA does, in principle, accept PROs as primary evidence to support product approval, this has rarely been the case for anticancer product approval. It has, however, been the case for symptom control agents in a number of medical conditions. Many of the methodological concerns outlined by the FDA can be overcome. Some, notably missing data, are likely to continue to be problematic in symptom management trials, particularly in the palliative setting. Furthermore, the desire for concordance among outcomes is potentially problematic when treatments improve some symptoms or aspects of HRQOL, possibly even overall HRQOL, but result in some toxic effects that are reflected in deterioration in other symptoms or aspects of HRQOL.
Overall, the positions put forward by the FDA and in other papers included in this monograph are not mutually exclusive—the overarching belief is that PROs are valid as observable outcomes and that, for the most part, identified methodological shortcomings can be addressed to allow inclusion of PROs (either symptoms or HRQOL or both) as key endpoints in symptom management trials. The specific outcomes included will reflect the study hypothesis and design.
|
Moving Forward |
|---|
|
TopAbstractOverview and HistoryUtility and Rationale for...Conceptual Framework--Beyond the...Instruments, Measurement, and...The Food and Drug...Key MessagesMoving ForwardReferences |
|---|
There are some advantages to having this monograph prepared some 18 months after the workshop where these issues were first discussed. With the passage of time, the issues that were discussed in late 2005 have become even more salient. The field of PRO measurement has come of age with the elaborate National Institutes of Health–funded Patient-Reported Outcomes Measurement Information System (PROMIS) initiative (http://www.nihpromis.org/), in which item banks and psychometric strategies are being developed to address the key domains of HRQOL and several common symptoms (e.g., pain, fatigue, anxiety depression), across a range of chronic diseases. With this effort and the taxonomy of PROs that has been developed by the PROMIS investigators, we will soon have available a wide range of items and short forms to measure reliably many of the symptoms and domains that were of concern to the participants in the workshop, including the FDA. However, many of the unusual toxic effects of cancer therapies may still require novel strategies for tool development and validation.
At this point in time, we have the very scholarly FDA guidance as a model for how we should approach the assessment of PROs in cancer patients, especially if a product labeling claim is being considered as part of the drug approval process (9). It draws heavily from the social sciences and incorporates the interaction of various sources of data (e.g., demographics, biomedical factors, behaviors, expert-rated and patient-rated outcomes) that might influence survival or that could affect treatment and outcomes. Nevertheless, the design and measurement issues discussed in the guidance are consistent with the views of the experts who contributed to this monograph. What is apparent is that symptom management trials must clearly state the expected goal of the therapeutic strategy, with the expected effect size and appropriate measurement tool. Instruments that are used to measure outcomes should be reliable, valid, and responsive to change so that benefit (or lack thereof) of the symptom management strategy can be successfully evaluated. If there is strong a priori evidence to suggest an impact of the symptom on a domain of HRQOL, then its inclusion may be warranted as a secondary or exploratory endpoint. What one wants to avoid is the inclusion of multiple domains (and multiple comparisons) without some hypothesized relationship. Careful thought and planning will identify domains that warrant inclusion.
In spite of the caution regarding inclusion of HRQOL measures in symptom management trials, there still may be benefit to their inclusion, from a safety perspective, if the symptom management strategy could potentially lead to a decrement in PROs of specific domains of HRQOL. Symptom and HRQOL outcomes may diverge or converge in different situations, and to the extent that divergence is anticipated, measurement of both types of PROs may be important. Thoughtful design and execution of symptom management trials, as well as treatment and prevention trials, that include PROs is what needs emphasis. Efforts to match the measures to the patient population under study, to minimize missing data, and to facilitate lack of bias in assessment when possible (use of placebo controls) are particularly salient and consistent with good study design. These facts reflect the contributions of the papers in this monograph as well as the FDA guidance and demonstrate how much has been learned during the past two decades of conducting symptom management trials.
Future progress in the area of assessing PROs within symptom management and treatment trials will likely benefit from the recently established Steering Committee for Symptom Management and Health-Related Quality of Life Steering Committee within NCI's Clinical Trials Working Group implementation process. This group is charged with review of phase III symptom management clinical trials and the quality of life components in the disease treatment trials. The Scientific Steering Committees are designed to provide robust analysis of proposed concepts and facilitate the sharing of ideas among a broad range of clinical investigators, basic and translational scientists, NCI staff, community oncologists, and patient advocates in the development of those concepts. Broad representation from the extramural community of investigators with PRO expertise should facilitate the development and implementation of high-quality trials.
|
REFERENCES |
|---|
|
TopAbstractOverview and HistoryUtility and Rationale for...Conceptual Framework--Beyond the...Instruments, Measurement, and...The Food and Drug...Key MessagesMoving ForwardReferences |
|---|
(1) Constitution of the World Health Organization. World Health Organization. In: Handbook of basic documents (1952) 5th ed. Geneva (IL): Palais des Nations. 3–20.
(2) Buchanan DR, O'Mara AM, Kelaghan JW, Minasian LM. Quality-of-life assessment in the symptom management trials of the National Cancer Institute-supported Community Clinical Oncology Program. J Clin Oncol (2005) 23:591–8.
(3) Ferrans CE. Differences in what quality-of-life instruments measure. J Natl Cancer Inst Monogr (2007) 37:22–6.
(4) Osoba D. Translating the science of patient-reported outcomes assessment into clinical practice. J Natl Cancer Inst Monogr (2007) 37:5–11.
(5) Ganz PA, Moinpour CM, Cella DF, Fetting JH. Quality-of-life assessment in cancer clinical trials: a status report. J Natl Cancer Inst (1992) 84:994–5.
(6) National Institutes of Health State of the Science. Symptom management in cancer: pain, depression and fatigue: State-of-the-Science Conference Statement. In: J Pain Palliat Care Pharmacother (2003) 17:77–97.[Medline]
(7) Kopec JA, Land SR, Cecchini RS, Ganz PA, Cella D, Costantino JP, et al. Self-reported neurotoxicity in patients with operable colon cancer receiving oxaliplatin. J Support Oncol (2006) 4:W1–8.
(8) Land SR, Kopec JA, Cecchini RS, Ganz PA, Wieand HS, Colangelo LH, et al. Neurotoxicity from oxaliplatin combined with weekly bolus fluorouracil and leucovorin as surgical adjuvant chemotherapy for stage II and III colon cancer: NSABP C-07. J Clin Oncol (2007) 2205–11.
(9) U.S. Food and Drug Administration. Guidance for industry patient-reported outcome measures: use in medical product development to support labeling claims. In: Food and Drug Administration draft guidance; 2006. Available at: http://www.fda.gov/cder/guidance/5460dft.htm. [Last accessed: May 10, 2007.].
(10) Cleeland CS. Symptom burden: multiple symptoms and their impact as patient-reported outcomes. J Natl Cancer Inst Monogr (2007) 37:16–21.
(11) Wilson IB, Cleary PD. Linking clinical variables with health-related quality of life: a conceptual model of patient outcomes. J Amer Med Assoc (1995) 273:59–65.
(12) Cella D, Wagner L, Cashy J, Hensing TA, Yount S, Lilenbaum RC. Should health-related quality of life be measured in cancer symptom management clinical trials? Lessons learned using the Functional Assessment of Cancer Therapy. J Natl Cancer Inst Monogr (2007) 37:53–60.
(13) Miaskowski C, Aouizerat BE, Dodd M, Cooper B. Conceptual issues in symptom clusters research and their implications for quality-of-life assessment in patients with cancer. In: J Natl Cancer Inst Monogr (2007) 37:39–46.
(14) Moinpour CM, Donaldson GW, Redman MW. Do general dimensions of quality of life add clinical value to symptom data? J Natl Cancer Inst Monogr (2007) 37:31–8.
(15) Bruner DW. Outcomes research in cancer symptom management trials: the Radiation Therapy Oncology Group (RTOG) conceptual model. J Natl Cancer Inst Monogr (2007) 37:12–5.
(16) Rock EP, Scott JA, Kennedy DL, Sridhara R, Pazdur R, Burke LB. Challenges to use of health-related quality of life for Food and Drug Administration approval of anticancer products. J Natl Cancer Inst Monogr (2007) 37:27–30.
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||