© The Author 2007. Published by Oxford University Press.
Should Health-Related Quality of Life Be Measured in Cancer Symptom Management Clinical Trials? Lessons Learned Using the Functional Assessment of Cancer Therapy
Affiliations of authors: Evanston Northwestern Healthcare, Evanston, IL (DC, LW, TAH, SY); Northwestern University Feinberg School of Medicine, Chicago, IL (DC, LW, TAH, SY); University of Chicago, Chicago, IL (JC); Thoracic Oncology Program, Mount Sinai Comprehensive Cancer Center, Miami Beach, FL (RCL)
Correspondence to: David Cella, Center on Outcomes Research and Education, Evanston Northwestern Healthcare, 1001 University Place, Evanston, IL, 60201(e-mail: d-cella{at}northwestern.edu).
 |
ABSTRACT |
|---|
 Top Abstract IntroductionBackground on the Functional...Prevalence of Poor Performance...Relationship Between Performance...ConclusionNotesReferences |
|---|
There are several advantages to including comprehensive health-related quality of life (HRQL) in symptom trials in oncology. The most obvious is to test the hypothesis that HRQL will be improved in addition to the symptom benefit. We should not "require," however, that a successful symptom intervention also improve other dimensions of HRQL. On the other hand, we should expect that it will not make other dimensions worse through side effects or exacerbation of disease, even if it improves the symptom. HRQL assessment in the trial helps evaluate the competing risks of any therapy. Furthermore, assessment of HRQL is now accomplished with very brief assessment (usually 30 questions or less), and the knowledge gained is valuable. With HRQL, one can compare cancer patients with those with other conditions and can determine the contribution of symptoms and side effects to the more broadly defined HRQL. Examples using the Functional Assessment of Cancer Therapy measurement system will demonstrate how HRQL assessment has contributed to our understanding of common cancer symptoms and their place in the conceptualization of HRQL. The prevalence of clinically significant symptoms is greatest in poor performance status (PS) patients compared with patients with good PS. Symptom improvement trials specifically designed for these patients should be encouraged, particularly with interventions that can provide symptomatic relief and improve multidimensional HRQL.
|
INTRODUCTION |
|---|
|
TopAbstractIntroductionBackground on the Functional...Prevalence of Poor Performance...Relationship Between Performance...ConclusionNotesReferences |
|---|
In a recent article, Buchanan et al. (1) expressed some concern that multidimensional health-related quality of life (HRQL) assessment in cancer symptom management trials was rising without good justification. Specifically, they suggested that interventions aimed to improve a single symptom often fail to improve HRQL and that improvement of HRQL need not be an expectation in a symptom intervention trial. Therefore, it follows, measuring HRQL in a symptom management trial should not be considered an automatically approvable interest. Rather, the clinical investigator should provide a clear and compelling rationale for inclusion of HRQL as an outcome of a symptom intervention trial. Some examples of acceptable rationales include a desire to gauge the importance patients assign to symptom relief, an interest in gaining information about offsetting treatment impact (side effects), and an interest in testing a compelling conceptual model of the relationship between symptoms and quality of life.
This article illustrates several advantages to including comprehensive HRQL in symptom trials in oncology. Examples using the Functional Assessment of Cancer Therapy (FACT) measurement system will demonstrate how HRQL assessment has contributed to our understanding of common cancer symptoms and their place in the conceptualization of HRQL. Before expounding on this discussion, an important clarification is needed: symptoms are a component of HRQL. They are not separate physical and mental experiences that exist outside the realm of some more esoteric concept of HRQL. All commonly used HRQL questionnaires include self-reported symptoms [for reviews, see (2,3)], and most extant conceptualizations of HRQL include physical and mental aspects of health, the operational definitions of which include symptoms (4). The two core aspects of HRQL are its subjectivity and multidimensionality. Although there are several proposed models to depict HRQL dimensions, these models share much in common. The World Health Organization triad of physical, mental, and social health have become the basis for the HRQL framework of the National Institutes of Health Roadmap Patient-Reported Outcomes Measurement Information System. This framework prominently includes disease-related symptoms (http://www.nihpromis.org). Therefore, the debate ought not to be whether there is a place for HRQL assessment in symptom trials but rather whether or not it makes sense to measure anything other than the target symptom experience—which is usually one narrow slice of HRQL, but a component nevertheless.
|
Background on the Functional Assessment of Cancer Therapy Measurement System |
|---|
|
TopAbstractIntroductionBackground on the Functional...Prevalence of Poor Performance...Relationship Between Performance...ConclusionNotesReferences |
|---|
Initially guided by direct patient input, clinical expert input, literature review, including the World Health Organization three-dimensional framework for self-reported health, version 1 of the FACT-General (FACT-G) was developed to sample the full content domain of HRQL as viewed by people with cancer. Factor analysis of version 1 response data led to a five-scale (version 2) instrument that measured physical, functional, emotional, and social well-being and relationship with physician (5). The last of these domains, relationship with physician, was subsequently removed from the core FACT-G scale and expanded as a noncore subscale to measure treatment satisfaction. This resulted in today's 27-item FACT-G (version 4) that provides scores for physical, functional, emotional, and social well-being. Each question is answered with regard to the past 7 days on a Likert-type scale ranging from 0 (not at all) to 4 (very much). Responses can be summed to create a total FACT-G score, or four domain scores can be calculated, depending on the research question. The FACT-G can be used by itself or in conjunction with any of 50 additional subscales that measure more targeted questions associated with specific cancer diagnosis or treatment. These disease- or symptom-specific subscales have been developed and validated as part of the larger Functional Assessment of Chronic Illness (FACIT) Measurement System (http://www.facit.org). A subscale such as the 13-item FACIT-Fatigue scale (6) can be administered by itself to produce a fatigue score sufficiently precise for individual patient monitoring, or it can be administered with the FACT-G (i.e., 40 items total) to provide a complete picture of HRQL that includes a focused emphasis on the symptom of fatigue. The stand-alone (13 item) decision carries less patient burden and provides a reliable fatigue score; the 40-item option adds burden but at the same time adds the ability to determine if a change in fatigue is matched with (or offset by) change in other symptoms or function. Disease-, treatment-, or symptom-specific subscale scores have often been added to the FACT-G physical and functional well-being subscale scores to create a single Treatment Outcome Index score. The Treatment Outcome Index is a useful endpoint for clinical trials because it provides an assessment of the domains most directly affected by cancer and treatment interventions. When justified by the data, reduction of relevant endpoints to one summary score eliminates multiple comparisons and as a result improves power. In practice, this enhances responsiveness to meaningful change.
The performance status (PS) rating is a universally accepted and practiced classification system for documenting functional status. The two main PS scales routinely used in oncology are the Karnofsky Performance Status (KPS) (7) with 11 hierarchic functional categories and Eastern Cooperative Oncology Group (ECOG) Performance Status Rating (8) with five functional categories. Because of the way the FACT questionnaires are constructed, including questions about symptoms and questions about functional status and perceptions of well-being, we are able to understand these important relationships whenever a FACT questionnaire is included in a clinical trial. Later we will illustrate what this has enabled us to appreciate about the relationship between symptoms and PS.
Health-Related Quality of Life Is a Relevant Endpoint for Symptom Intervention Studies
As noted by Buchanan et al. (1), many studies aimed at improving cancer symptoms have shown benefit to the individual symptoms and yet not to overall (total or global) HRQL. While this has been the case in some studies, there are notable exceptions. For example, Velikova et al. (10) showed that overall HRQL as measured by the FACT-G was clinically and statistically significantly improved based on a symptom management intervention as compared with a usual care control arm. Patients assigned to the intervention arm demonstrated a 9.2-point increase on the FACT-G, indicating improved HRQL, in comparison to a decrease of 2.0 points among participants on the control arm. These findings support the relevance of overall HRQL as a symptom management clinical trial endpoint. Having said this, it is unclear how lack of impact on overall HRQL would be a plausible argument against including HRQL in symptom trials. Perhaps we should value a symptom intervention with a more expansive benefit than merely the symptom at hand, when compared with an intervention that improves only the symptom. While there is clear value in relieving a symptom absent any other benefit, is it not even better to expand that benefit to other areas of HRQL?
Health-Related Quality of Life Benefit Observed in the Absence of Symptom Benefit
Some studies suggest there may be an overall HRQL benefit even when a more logical target symptom does not appear to be improved substantially. In a study comparing epoetin beta with placebo for the management of chemotherapy-induced anemia, Osterborg et al. (11) found no demonstrable benefit to fatigue, and yet there was a statistically significant benefit to the more general HRQL. Because anemia is associated with a constellation of symptoms, including but not limited to fatigue, the overall HRQL assessment proved more sensitive to therapeutic benefit. Studies like this argue in favor of including general HRQL in symptom studies as a secondary endpoint to avoid mistakenly concluding a lack of therapeutic benefit when treatment arm differences are not observed on a single symptom endpoint.
Symptom Interventions Can Introduce Side Effects
A substantial concern in symptom management trials that target a single symptom is that symptom relief may come at a cost to other symptoms or HRQL. An analgesic intervention that reduces pain is only of value if the competing problems with constipation, drowsiness, or other side effects do not outweigh that benefit. An assessment of overall HRQL can facilitate an analysis of symptom relief versus symptom burden associated with symptom management interventions. This is a critical comparison because treatment side effects can compromise HRQL enough to counteract any treatment benefit. Using a single item from the FACT-G ("I am bothered by treatment side effects"), total FACT-G scores were examined from breast cancer (n = 529), colorectal cancer (n = 254), and head and neck cancer patients (n = 233) to quantify the impact of side effect burden on overall HRQL. Patients who reported that they were "not at all" bothered by side effects reported statistically significantly higher HRQL than those who reported that they were bothered by treatment side effects "very much" or "quite a bit" [(11); see also Fig. 1]. A general HRQL assessment helps provide reassurance that the intervention was efficacious without detriment to HRQL. Therefore, a lack of treatment arm differences on HRQL endpoints is informative when evaluating a symptom intervention because it provides evidence that symptom relief can be achieved without compromising HRQL.
|
Health-Related Quality of Life Assessment Can Capture the Trajectory of Symptoms Longitudinally
Oncology patients experience many disease- and treatment-related symptoms. Most patients, particularly those with advanced disease, experience more than one symptom. These multiple symptoms can occur in clusters, or symptoms may be independent and follow different trajectories. Focusing exclusively on a single symptom endpoint may not adequately capture the trajectory of disease- and treatment-related symptoms over time and therefore may not provide an adequate assessment of the efficacy or breadth of symptom management interventions. HRQL instruments that include items to assess specific symptoms, such as the FACT-G, can be used to better understand patients' symptom experience over time. As an example, Wagner et al. (13) selected items a priori from the FACT-Lung (FACT-L) questionnaire (FACT-G plus Lung Cancer Subscale) to examine the trajectory of clinically significant neuropsychiatric symptoms among non–small-cell lung cancer patients with provider-rated PS of 0–1 at baseline who participated in a cooperative group oncology trial. Symptoms were considered to be clinically significant if participants reported experiencing the symptom "quite a bit" or "very much" over the previous 7 days. Despite having a PS rating of 0 or 1, approximately 20% of participants had clinically significant fatigue and fatigue steadily increased during treatment to 33% at 6 months. Twenty-five percent of patients had clinically significant anhedonia at baseline, which ranged up to 34% throughout treatment. In contrast, 17% of patients had clinically significant depressed mood at baseline, and this decreased to 9% at 6 months. The percentage of patients who reported clinically significant general worry also decreased over time, from 25% at baseline to 9% at 6 months. Similarly, 21% of patients had clinically significant pain at baseline, and this percentage decreased to 9% at 6 weeks and 10% at 12 weeks and was 14% at 6 months (12). These findings demonstrate the usefulness of the FACT-L, a general HRQL instrument that also assesses a set of targeted lung cancer symptoms, to understand lung cancer patients' symptom experience throughout treatment. Results indicated that up to 25%–35% of lung cancer patients with good PS (0/1) report symptoms that are clinically significant and warrant intervention. This information has informed investigators at ECOG regarding priorities in symptom management and clinical research with lung cancer.
Health-Related Quality of Life Measures Allow for Informative Comparisons Across Cancer Sites
Although cancer is really a collection of more than 100 diseases, most share the same set of common symptoms: fatigue, pain, nausea, anorexia, etc. In a survey conducted by the National Comprehensive Cancer Network (13), pain and fatigue were found to be the most important symptoms to monitor across nine solid tumors in the advanced disease setting. However, to a degree, the most prevalent and troublesome cancer- and treatment-related symptoms can vary by cancer type. For example, shortness of breath, fatigue, and pain are the most important symptoms to lung cancer patients, whereas difficulty swallowing, pain, and fatigue have highest priority for head and neck cancer patients (13). Multidimensional HRQL instruments allow comparisons among patients with various types of cancer because instruments such as the FACT-G include items to assess the more common cancer-related symptoms (e.g., pain, fatigue, nausea) as well as items to assess general HRQL that are relevant to all patients.
Health-Related Quality of Life Data Can Provide an Informed Assessment of Symptoms and Function
HRQL instruments can also be used to compare oncology samples with other diseases. Using the FACT-G, Cella and Nowinski (8) compared oncology patients with patients with various other diseases. Oncology patients reported a total FACT-G score that was 14 points lower than the average score for individuals with no history of illness, and this corresponded to an effect size of approximately 0.90 (Fig. 2, bottom). Cancer patients reported better HRQL than patients who had a history of stroke, hepatitis, anxiety, depression, and cardiac problems and poorer HRQL than those with diabetes, asthma, and hypertension (see Fig. 2). These findings can assist clinicians with understanding the impact of illness on their patients' well-being relative to the larger chronic disease landscape.
|
Further understanding of the relationship between symptoms and HRQL can be gained by plotting individual symptom levels against the broader, more inclusive HRQL. Figure 3 provides an illustration. The data for Fig. 3 came from the first 1163 people enrolled into the Q-score study described in the next section. The study compared and pooled data from four common HRQL questionnaires, including the FACT-G and the European Organization for Research and Treatment of Cancer Quality of Life Questionnaire (EORTC QLQ). Detailed clinical and demographic information can be found in Chang and Cella 15). Having data from more than one HRQL questionnaire enabled us to examine the relationship between symptom severity on selected EORTC items and overall HRQL as defined by the total FACT-G score (16). Visual inspection of these strong linear associations between HRQL and individual symptoms (in this case, pain, insomnia, fatigue, nausea, and diarrhea) illustrates their interrelatedness and provides a basis for anticipating that a good symptom intervention should improve or at least not worsen HRQL.
|
Health-Related Quality of Life Data Can Help Appreciate the Symptom Severity Associated With Poor Performance Status
Global ratings, such as PS, can be an informative basis for comparison of patients with different types or stages of cancer. Surprisingly, this has rarely been done systematically in such a way that relative burden of cancer can be compared across sites. If symptom data are captured alongside PS, this sets up an opportunity to better appreciate the relationship between the two across various cancer sites and stages of disease. To illustrate this, we pooled data from two large databases. Both studies included Institutional Review Board-approved survey results from outpatients and inpatients at the participating institutions. The first database, "Bilingual Intercultural Oncology Quality of Life," is from a study conducted between 1994 and 1996. Participating institutions included Rush-Presbyterian Medical Center, Chicago, IL; Cook County Hospital, Chicago, IL; San Juan Oncology, San Juan, PR; San Juan VA, San Juan, PR; San Juan City Hospital, San Juan, PR; Grady Memorial Minority Community Clinical Oncology Program, Atlanta, GA; and Emory University, Atlanta, GA. This study evaluated the impact of language, culture, and literacy upon HRQL of patients with cancer and HIV disease. Specific aims included 1) completion of the adaptation of FACT scales for use with Spanish-speaking cancer patients and 2) testing of the psychometric properties and statistical equivalence of the FACT scales across the following four variables: language, culture, literacy, and mode of administration. This study demonstrated linguistic, conceptual, and technical equivalence between the English and Spanish language versions of the FACT (17) and demonstrated that language, culture, and literacy by themselves are not drivers of HRQL (18).
The Q-score database contains data collected in 1995 and 1996 from Rush-Presbyterian Medical Center, Chicago, IL; Northwestern University Medical School, Chicago, IL; Johns Hopkins University Medical Center, Baltimore, MD; Fox Chase Cancer Center, Philadelphia, PA; and the Medical College of Ohio, Toledo, OH. These sites participated in a project to evaluate the comparability of four HRQL questionnaires commonly used in oncology: The Cancer Rehabilitation and Evaluation System (19), EORTC QLQ C-30 (20), the FACT-G (5), and The Medical Outcomes Study Short Form 36 (21).
The combined sample size consisted of 3329 patients. Patients with HIV disease (n = 444) were excluded, leaving a pooled cancer sample of 2885 patients. In both source studies, providers and patients independently completed identical versions of the ECOG PS scale [0 = normal activity, 1 = some symptoms, but no bed rest during daytime, 2 = bed rest for <50% of daytime, 3 = bed rest for >50% of daytime, 4 = unable to get out of bed"; (8)]. Relative risks of having poor PS for each tumor type were computed using localized breast cancer as the reference group.
Patients in both source studies completed the FACT-G as a tool for HRQL assessment. The responses to the 10 negatively worded symptoms (both physical and emotional) of the FACT-G were used to compare the relationship between patient-reported PS and symptom severity. A single question assessing bother with side effects was also included. These questions are reproduced in Table 1. Each question of the FACT-G requests an answer using the following scale: 0 = "not at all"; 1 = "a little bit"; 2 = "somewhat"; 3 = "quite a bit"; and 4 = "very much." Any symptom rated by a patient as either 3 or 4 was classified as "clinically significant." All symptoms rated as 0, 1, or 2 were classified as not clinically significant. This produced a conservative estimate of the number of clinically significant symptoms experienced by each patient. The average numbers of symptoms for every level of ECOG PS were compared using an exploratory analysis of variance with P value set at .05.
|
There was a similar number of men and women in the pooled cohort, and the majority of patients (83%) were treated as outpatients (see Table 2). Just over half (57%) of the patients were White, with most of the remainder being either African American (22%) or Hispanic (19%). Twenty-eight percent of patients had a diagnosis of breast cancer, followed in order by cancer of the colon, head and neck, lung, lymphatic system, and prostate. Among all 2885 patients, comparing PS groups 0 through 4, there was a monotonically increasing mean number of clinically significant symptoms for more impaired PS scores (see Table 3). The average number of clinically significant symptoms (emotional and physical combined) steadily climbed from mean = 0.5 at PS0 to mean = 3.2 at PS4. Means were compared using analysis of variance and a Scheffe test for pairwise comparisons (a = 0.05). For physical symptoms, all adjacent PS groups differed from each other in the expected direction; the only statistically significant difference between adjacent PS levels for emotional symptoms was between PS2 and PS3.
|
|
|
Prevalence of Poor Performance Status |
|---|
|
TopAbstractIntroductionBackground on the Functional...Prevalence of Poor Performance...Relationship Between Performance...ConclusionNotesReferences |
|---|
Among all patients with cancer, the majority (61%) were classified as having a good PS, defined as ECOG PS = 0/1 (Table 3). We determined relative risk of having poor PS (defined as PS = 2–4) by tumor type and disease severity. Both provider and patient ratings were evaluated. Because they had the lowest risk of poor PS, patients with localized breast cancer were set as the reference group for these analyses. Using this cohort as the benchmark, the risk of having poor PS was then calculated for other tumor types using both patient and provider ratings of PS (Fig. 4). Compared with localized breast cancer, the risk for poor PS was about fivefold in advanced lung cancer patients and almost threefold in localized lung cancer.
|
|
Relationship Between Performance Status and Clinically Significant Symptoms |
|---|
|
TopAbstractIntroductionBackground on the Functional...Prevalence of Poor Performance...Relationship Between Performance...ConclusionNotesReferences |
|---|
Patients' responses to HRQL questionnaires were used to compare the relationship between PS and clinically significant symptoms. Comparing PS groups 0 through 4, there was a monotonically increasing mean number of clinically significant symptoms for more impaired PS scores (see Table 3). The average number of clinically significant symptoms (emotional and physical combined) steadily climbed from mean = 0.5 at PS0 to mean = 3.2 at PS4. This increase was statistically significant (P<.05) with the largest difference between PS2 and PS3.
Table 4 reports the HRQL scores of patients according to ECOG PS rating. With the exception of social well-being, HRQL scores of each domain and aggregated score were statistically significantly worse in poorer PS patients relative to better PS patients.
|
These results illustrate that by comparing PS across conditions, one can estimate relative risk of poor PS by site and severity of disease. The number of clinically significant symptoms, the primary driver of poor PS, is higher in patients with poor PS. These data, all obtained using the FACT-G in the context of clinical care, provide targets for intervention on each of the individual symptoms that contributes to poor function and well-being in people with cancer. Clinically significant and treatable symptoms are more common among those patients with poor PS. Nearly two of every five (39%) cancer patients in this study of 2885 patients rated their PS as poor (i.e., ECOG 2, 3, or 4). The risk of poor PS was greatest for patients with advanced disease, particularly those with advanced lung cancer and other/unknown primaries (Fig. 4). Increasing numbers of disease-related symptoms correlated with declining PS. The total number of physical and emotional symptoms, as determined by the patients' responses to relevant items of the FACT-G, increased with worse patient-assessed PS. Among all cancer patients, there was a substantial increase in the average number of symptoms when going through nearly each step of the PS scale; the only step for which the difference was not statistically significant was from PS3 to PS4.
Chang et al. (22) performed a similar analysis of the relationship between symptoms and PS using the KPS scale in a group of 240 US Department of Veterans Affairs patients with cancer. They determined the prevalence of symptoms using the patients' responses to the Memorial Symptom Assessment Scale. The median number of symptoms was 5 for patients with KPS 90–100, 9 for those with KPS 60–80, and 13 for those with KPS <60 (P<.001). The number of symptoms also correlated statistically significantly with the FACT-G score, supporting a relationship between increasing symptoms and decreasing HRQL. Importantly, that study used a lower threshold for indexing a given symptom and assessed more symptoms than this study. We therefore consider our estimates of symptom burden to be very conservative. Together these data support the inclusion of HRQL measurements in clinical trials conducted in patients with poor PS to test the hypothesis that symptom improvement indeed improves HRQL.
|
Conclusion |
|---|
|
TopAbstractIntroductionBackground on the Functional...Prevalence of Poor Performance...Relationship Between Performance...ConclusionNotesReferences |
|---|
In summary, there are many good reasons to include multidimensional HRQL assessments in a clinical trial. Only one reason is the possible hypothesis that HRQL will be improved in addition to the symptom benefit. Just as we do not require that a cytotoxic intervention associated with improved survival should also improve HRQL, we should not require that a successful symptom intervention also improve other dimensions of HRQL. On the other hand, perhaps we should expect that it will not make other dimensions worse (through side effects or disease exacerbation) to a degree that counteracts the symptom benefit. HRQL assessment in the trial helps evaluate the competing risks of any therapy. Lessons learned about HRQL based on the inclusion of the FACT Measurement System have added to our ability to compare degree of disability across diseases, to understand the natural history of symptoms in clinical trial participants, and to appreciate the relationship between patients' PS and their symptom severity. This can be helpful when planning priority symptom management trials and fully evaluating their outcomes.
|
NOTES |
|---|
This work was supported in part by grants from the National Cancer Institute (CA61679 and CA60068).
|
REFERENCES |
|---|
|
TopAbstractIntroductionBackground on the Functional...Prevalence of Poor Performance...Relationship Between Performance...ConclusionNotesReferences |
|---|
(1) 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.
(2) Cella D. Quality of life. In: Psycho-oncology—Holland J, ed. (1998) New York: Oxford University Press. 1135–46.
(3) Cella DF, Bonomi AE. Measuring quality of life: 1995 update. Oncology (1995) 9(Suppl):47–60.[Medline]
(4) Spilker B, ed. Quality of life and pharmacoeconomics in clinical trials. (1996) 2nd ed. Philadelphia (PA): Lippincott-Raven. Chap. 2, 4–6, 10.
(5) Cella DF, Tulsky DS, Gray G, Sarafian B, Linn E, Bonomi A, et al. The Functional Assessment of Cancer Therapy scale: development and validation of the general measure. J Clin Oncol (1993) 11:570–9.
(6) Yellen SB, Cella DF, Webster KA, Blendowski C, Kaplan E. Measuring fatigue and other anemia-related symptoms with the Functional Assessment of Cancer Therapy (FACT) measurement system. J Pain Symptom Manage (1997) 13:63–74.[CrossRef][Web of Science][Medline]
(7) Karnofsky DA, Burchenal JH. The clinical evaluation of chemotherapeutic agents in cancer. In: Evaluation of chemotherapeutic agents—Macleod CM, ed. (1949) New York: Columbia University Press. 199–205.
(8) Cella D, Nowinski C. Measuring quality of life in chronic illness: the Functional Assessment of Chronic Illness Therapy measurement system. Arch Phys Med Rehabil (2002) 83(Suppl 2):S10–17.[CrossRef][Web of Science][Medline]
(9) Zubrod CG, Scheiderman MA, Frei E, Brindley C, Gold LG, Shnider B, et al. Appraisal of methods for the study of chemotherapy in man: comparative therapeutic trial of nitrogen mustard and triethylene thiophosphoramide. J Chronic Dis (1960) 11:7–33.[Medline]
(10) Velikova G, Booth L, Smith AB, Brown PM, Lynch P, Brown JM, et al. Measuring quality of life in routine oncology practice improves communication and patient well-being: a randomized controlled trial. J Clin Oncol (2004) 22:714–24.
(11) Osterborg A, Brandberg Y, Molostova V, Iosava G, Abdulkadyrov K, Hedenus M, Messinger D. Epoetin Beta Hematology Study Group. Randomized, double-blind, placebo-controlled trial of recombinant human erythropoietin, epoetin Beta, in hematologic malignancies. J Clin Oncol (2002) 20:2486–94.
(12) Cella D. Assessment methods for quality of life in cancer patients: the FACIT measurement system. Int J Pharmical Medicine (2000) 14:78–81.[CrossRef]
(13) Wagner LI, Rue M, Fisch M, Cleeland C, Bonomi PD, Schiller JH, et al. Neuropsychiatric symptoms during cancer treatment: an examination of quality of life data from two ECOG lung cancer trials. Proc ASCO (2002) 21:357a.
(14) Cella D, Paul D, Yount S, Winn R, Chang C-H, Banik D, Weeks J. What are the most important symptom targets when treating advanced cancer? A survey of providers in the National Comprehensive Cancer Network (NCCN). Cancer Investigation (2003) 21:526–35.[CrossRef][Web of Science][Medline]
(15) Chang CH, Cella D. Equating health-related quality of life instruments in applied oncology settings. Phys Med Rehabil State Art Rev (1997) 11:397–406.
(16) Chang CH, Peterman A, Cella D, Mo F. Relationship between symptoms and overall quality of life [abstract]. Qual Life Res (1997) 6:631–2.
(17) Cella DF, Hernandez L, Bonomi AE, Corona M, Vacquero M, Shiomoto G, et al. Spanish language translation and initial validation of the Functional Assessment of Cancer Therapy (FACT) quality of life instrument. Med Care (1998) 36:1407–18.[CrossRef][Web of Science][Medline]
(18) Wan GJ, Counte MA, Cella D, Hernandez L, McGuire D, Deasy S, et al. The impact of socio-cultural and clinical factors on health-related quality of life reports among Hispanic and African-American cancer patients. J Outcome Meas (1999) 3:200–15.[Medline]
(19) Schag CA, Ganz PA, Wing DS, Sim MS, Lee JJ. Quality of life in adult survivors of lung, colon, and prostate cancer. Qual Life Res (1994) 3:127–41.[CrossRef][Web of Science][Medline]
(20) Aaronson NK, Ahmedzai S, Bergman B, Bullinger M, Cull A, Duez NJ, et al. The European Organization for Research and Treatment of Cancer QLQ-C30: a quality-of-life instrument for use in international clinical trials in oncology. J Natl Cancer Inst (1993) 85:365–76.
(21) Ware J, Kosinski M, Keller S. SF-36 physical and mental health summary scales: a user's manual (1994) Boston (MA): The Health Institute, New England Medical Center.
(22) Chang VT, Hwang SS, Feuerman M, Kasimis BS, et al. Symptom and quality of life survey of medical oncology patients at a veterans affairs medical center: a role for symptom assessment. Cancer (2000) 88:1175–83.[CrossRef][Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  P. A. Ganz and P. J. Goodwin Health-Related Quality of Life Measurement in Symptom Management Trials J Natl Cancer Inst Monographs, October 1, 2007; 2007(37): 47 - 52. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||