© 2002 AlphaMed Press The Cancer Patient and Quality of LifeEuropean Organisation for Research and Treatment of Cancer (EORTC) Data Center, Brussels, Belgium Correspondence: Andrew Bottomley, Ph.D., European Organisation for Research and Treatment of Cancer, EORTC Data Center, Avenue E Mounier 83, 1200 Brussels, Belgium. Telephone: 32-2-774-1661; Fax: 32-2-779-4568; e-mail: abo{at}eortc.be
Over the last decade, clinicians have accepted that while survival and disease-free survival are critical factors for cancer patients, overall quality-of-life is fundamental. This review considers recent developments in the field of quality of life, oncological challenges and future directions. Key Words. Health-related quality of life • rHuEpo • Fatigue • Subjective experience • Cancer
Increasingly, researchers are faced with situations where patients may not gain benefits in terms of traditional end points, such as survival or disease-free survival. However, it is possible to see significant changes in health-related quality of life (HRQOL) [1]. HRQOL, a multidimensional construct and an important concept, has, for many years, proven difficult to define. A number of definitions proposed by various authors as to the exact nature of HRQOL and the formulation of a defining consensus are shown in Table 1
Generally, HRQOL covers the subjective perceptions of the positive and negative aspects of cancer patients' symptoms, including physical, emotional, social, and cognitive functions and, importantly, disease symptoms and side effects of treatment [2]. Only 20 years ago, scant literature reported quality-of-life benefits. However, in recent years, there has been a large increase in studies reporting the assessment of HRQOL. At present, some 10% of all randomized cancer clinical trials include HRQOL as the main end point [3]. However, while the U.S. Food and Drug Administration now recognizes the benefits of HRQOL as a basis for approval of new anticancer drugs, and many international research groups include HRQOL in their studies, introducing HRQOL into oncology has been difficult. There are several reasons for this [46]. One problem involves understanding the subjective nature of the results that HRQOL studies generate and the barriers to acceptance by clinicians [1]. Furthermore, as Moinpour [7] points out, bringing these metrics into a busy practice is difficult. The purpose of this review is to help clinicians understand the value of HRQOL.
Fortunately, more clinicians are considering the importance of HRQOL as critical to cancer patients' care [8, 9]. Recently, Tanaka and Gotay [10] demonstrated this in a HRQOL awareness survey of U. S. clinicians and medical students. HRQOL was perceived to be as important as survival in making treatment decisions. However, medical students were more likely to emphasize HRQOL over survival than practicing clinicians. This emphasis was confirmed by Morris et al. [11], who undertook a survey of 260 senior oncologists to study how HRQOL is viewed in clinical practice. An impressive number (80%) of the questionnaire responders believe that HRQOL should be collected from patients; although, in practice, only some 50% managed to do this, claiming problems of limited time and resources. Importantly, confusion about the measure to use is the biggest barrier. The latter is a major problem in the use of measures and definitions, as Taylor et al. [12] noted when interviewing 60 oncologists in Canada and the U.S. The results are broadly in line with Tanaka and Gotay [10] and Morris et al. [11]: some 88% of clinicians thought that it was important to examine HRQOL, while over 33% of clinicians felt that the current measurement tools were inadequate. Bezjak et al. [13] found, in a survey of 357 Eastern Cooperative Oncology Group clinicians, that 84% felt that their individual knowledge of HRQOL was limited. However, over 82% believed that HRQOL data are appropriate for patient care. With more specific knowledge, oncologists are likely to increase their use of HRQOL. Detmar and Aaronson [4], in order to assess the value of individual HRQOL assessments, undertook a small-scale study with six clinicians and 18 cancer patients. Just prior to the consultation, clinicians were given the results of the patients' HRQOL. The results suggested that patients were more satisfied with the consultation, stating that the clinician showed greater understanding. The clinician also rated the consultation as more satisfying, without greatly increasing the time taken. The main inference from this small study is that HRQOL can be used in an individual setting, and clinicians may benefit from completing and using HRQOL data for treatment and clinical care-making decisions.
While there is increasing evidence for the value of HRQOL assessment, one of the most difficult tasks is actually measuring it. HRQOL is subjective and can prove a challenge to measure. Many of the components, such as social functioning and spirituality, cannot be directly observed. Therefore, these are measured using classical measurement paradigms. The measurement process draws on many different disciplines, including psychology and statistics. Fortunately, the majority of researchers recognize that when measuring quality of life, it is important to focus very clearly on specific domains important to cancer patients, rather than addressing a more general question such as, "How is your quality of life?" The inherent vagueness in the latter approach could lead to difficulties in interpretation.
A number of research articles suggest how clinicians can assess patient HRQOL more objectively by looking at symptoms and reporting on them, instead of the patient self-reporting. Stephens et al. [14], using the Rotterdam Symptom Checklist, surveyed over 700 lung cancer patients. Patients and clinicians reported on the same symptoms for each assessment period. When compared, clinicians frequently under-assessed the level of functioning of the patient and under-reported symptoms that the patient actually reported. This became more problematic: the worse the patient's symptoms, the less likely the clinician appeared to rate them as severe. Titzer et al. [15] noted, in a study of 163 cancer patients, that when they compared patient HRQOL with clinical assessment (rated as mild, moderate, or severe impairment), only 54% of physician assessments correlated with patient assessments. Therefore, the use of patient-reported questionnaires has become a standard practice in the assessment of cancer patient HRQOL. Some of the more commonly used measures are presented in Table 2
In many cases, if well-validated instruments have not been used in the correct manner, there may be problems with correct interpretation [16]. Kong and Gandhi [17] reviewed 265 articles reporting on HRQOL assessment in clinical trials; only 23% provided reliability data, and only 21% provided validity data. Given the multidimensional nature of HRQOL data, it is important that researchers provide information on all measures used, including the domain investigated, even if not significant [18]. It must be stressed that, to many researchers, HRQOL is not a single unidimensional concept. This is exemplified in the work of Demetri et al. [19] who undertook an open-labeled study with a community sample of 2,370 anemic cancer patients undergoing chemotherapy. Patients received human recombinant erythropoietin (rHuEpo) s.c., three times a week, at an initial dose of 10,000 units, for a period of up to 4 weeks. The researchers defined a clinical response as an increase in hemoglobin levels of 2 g/dl or above. Approximately 60% of patients responded to these criteria, yet there were no improvements on the Karnofsky Performance Status (KPS) measure. However, HRQOL was assessed using the Functional Assessment of Cancer TherapyAnemia (FACT-AN) measure, which demonstrated significant improvements in fatigue and general quality of life. This suggests that a single unidimensional measure such as the KPS is not sensitive enough to detect subtle changes in patient HRQOL improvement. Some researchers argue for the aggregation of scores, obtained from various HRQOL domains, as a way of expressing and interpreting overall HRQOL. The advantage of this is a single-number representation [20]. Yet a single score lacks detailed information about how the various domains are influenced by the disease and treatment. One way to address this limitation is the method adopted by the European Organisation for Research and Treatment of Cancer (EORTC) Quality of Life Group, in which researchers ensure they are able to obtain both a single global score as well as detailed knowledge of the various domains [21]. The EORTC quality-of-life questionnaire, QLQ-C30, not only collects details on the domains, but also assesses two global HRQOL items independent of the domain scores. While global quality-of-life items are easier to interpret clinically, they can have disadvantages [22]. For example, it is possible that patients focus on limited physical functioning and possibly spiritual or psychological issues as their condition becomes more serious [23].
Recent work has involved looking at minimal clinical importance difference, a measurement noted as the smallest difference in a score of a domain a patient perceives as beneficial. This indicates, in the absence of adverse scale effects, a change in patient treatment and care [24]. However, when comparing mean score differences between two groups, some difficult interpretations could easily arise [2527]. Fortunately, several methods have been proposed to help interpretation. Perhaps the most common approach is to anchor the changes seen in disease-specific questions to one that asks about overall HRQOL, such as, "In general, how would you rate your quality of life?" Then, clinicians could look at changes in answers to such a global HRQOL question over time and compare this with changes seen on the disease-specific questionnaire [28]. In effect, the changes in the disease-specific measures are thus anchored to changes in overall health status.
Osoba et al. [29] investigated this issue using the EORTC QLQ-C30 [21], a subjective-significance questionnaire, asking patients about perceived changes in physical, emotional and social functioning, and global HRQOL. The patients were receiving chemotherapy for metastatic breast cancer or extensive disease small-cell lung cancer. The investigators interpreted changes on the QLQ-C30 as meaningful, depending on differences in scores. On a scale of 1-100, a difference of 5 to 10 was rated as little change. Patients who fell in the 10-20 range were rated as having experienced a moderate change; and those with a
Collecting HRQOL data in a clinical trial can be challenging. It requires resources and a survey in the patients' own language. Important data may also be difficult to obtain if the patients' performance status declines. One example is the study reported by EORTC investigators [3032]. Kramer and co-workers [33, 34] studied the effects of HRQOL on patients with advanced breast cancer. In this population, less than two-thirds of the patients completed baseline assessment, and compliance rates deteriorated continually thereafter. The investigators concluded that this compliance problem was related to several factors, including the unique challenges of collecting HRQOL data from within a multinational, multilanguage organization. In addition, collecting HRQOL data from patients with advanced and progressive disease may be more difficult because of their failing health. Other studies have come to similar conclusions. Coates et al. [32] investigated HRQOL in advanced breast cancer patients and found that only 44% of patients (133 of 305) were available for analysis after three cycles of chemotherapy. Patient compliance was poor both before and during treatment. In another study of patients with advanced breast cancer [35], only 71% of patients provided data at baseline and at the end of the third cycle 3 months later. One can reach erroneous conclusions from incomplete data. Several studies have failed to take these issues into consideration [26, 3638]. Clearly, these problems make it imperative that methods are adopted to ensure high-quality data are collected and appropriately analyzed [39]. Yet not all studies have had such problems. An impressive study, coordinated by the Scandinavian Breast Group [39], had 283 metastatic breast cancer patients randomized to either docetaxel or sequential methotrexate and 5-fluorouracil. While no differences in HRQOL were seen, compliance levels were over 96% at baseline and overall compliance was 82%. This clearly suggests that, in some studies, high compliance, even in advanced patients, can be achieved [40]. Sadura et al. [41], from National Cancer Institute of Canada, investigated levels of HRQOL compliance across melanoma, breast, and antiemetic therapy trials. In all trials, compliance was high and all assessments greater than 95%. They attributed their success to the development of a comprehensive education program prior to and during each trial. One other key factor was that baseline questionnaires were part of the eligibility checklist so that patients who refused to complete a baseline measure were not randomized into the trial. In addition, pre-trial workshops were conducted to help train nurses and data managers in collecting HRQOL data. This approach may represent a model requiring more resources to follow in selecting future trials.
AnemiaHRQOL Benefits of rHuEpo in Cancer Patients Some influential HRQOL studies have focused on cancer-related fatigue, an extremely common and distressing symptom during and following chemotherapy. Glaspy et al. [42] reported the results of an open-label nonrandomized study of rHuEpo in 2,342 cancer patients undergoing chemotherapy. Specifically, the study aimed to examine the impact of treatment on the effectiveness of rHuEpo in anemic patients over a 4-month period. The results suggested that mean hemoglobin levels increased progressively with a significant increase in hemoglobin levels of 1.8 g/dl from baseline to final assessment in 2,019 patients, i.e., all those with both baseline and final assessments. Overall, some 53.4% of patients demonstrated a substantial increase (defined as a >2.0 g/dl) in hemoglobin levels. Quality of life was assessed using the visual analog scale (VAS) (Abels et al.), and patients had energy, activity levels, and overall quality of life assessed before beginning and at the conclusion of rHuEpo treatment [43]. When compared with baseline scores for both groups of patients, significantly better scores were seen at the end of treatment on all three scales. A significant correlation between the magnitude of the improvement on each scale and the increase in hemoglobin level from baseline was noted. In a subanalysis of hemoglobin levels, patients who showed no improvement in hemoglobin levels had decreased quality-of-life scores (n = 252). Littlewood et al. [44] reported a large-scale double-blind randomized controlled trial, examining the effects of rHuEpo on anemic cancer patients. In total, 375 patients of mixed sex and diagnosis, receiving nonplatinum chemotherapy, were recruited. Patients were defined as anemic while receiving chemotherapy if their hemoglobin level was equal to or below 10.5 g/dl. Patients were then randomized to receive either 150 IU/kg rHuEpo s.c. or a placebo, three times a week for six cycles and for 4 weeks after completing treatment. Quality of life was assessed with three measurement tools: the three-item VAS, as used in the Abels et al. study [43]; the FACT-AN; and the SF 36. The authors reported a significant increase in the mean hemoglobin level of 2.2 g/dl in the rHuEpo group, while only a 0.5 g/dl increase was observed in the placebo group. Scores on all three items of the VAS, the Functional Assessment of Cancer TherapyGeneral (FACT-G), and the FACTAN system showed significant improvements in the rHuEpo group. However, on the well-validated SF 36, no significant improvements were reported. Such information may guide clinicians in the management of one component of cancer-related fatigue.
Example of the Use of Gemcitabine in Lung Cancer Patients One key study, undertaken by Anderson et al. [48], examined gemcitabine plus best supportive care (BSC) versus BSC in inoperable non-small-cell lung cancer. Patients were randomized to receive either gemcitabine 1,000 mg/m2 on days 1, 8, and 15 of a 28-day cycle, for a maximum of six cycles, plus BSC or BSC alone. The authors' main aim was to compare a predefined subset of commonly reported symptoms (SS14) from the EORTC QLQ-C30 scale with those from the LC13 scale. The EORTC QLQ-C30 and LC13 subscales both showed greater improvement in the gemcitabine plus BSC arm (in 11 domains) than in the BSC alone arm (one symptom item). Greater deterioration was observed in the BSC alone arm (six domains/items) than in the gemcitabine plus BSC arm (three quality-of-life domains). No difference in overall survival was reported, suggesting future treatments with gemcitabine could indeed lead to improvements in patient HRQOL.
While there are a number of barriers to its assessment, HRQOL is now more accepted by clinicians as a possible way to collect more meaningful data about patients' subjective experiences on cancer therapy. Increasing numbers of studies with quality-of-life outcome assessment as either a secondary or a primary end point are appearing. It is becoming clear that HRQOL information may, in some settings, lead to improvements in the status of the individual cancer patient. While many challenges exist in this field, it is hoped that the future will provide more acceptance of HRQOL and a more universal understanding of the concepts of HRQOL for patients and clinicians alike.
This article has been cited by other articles:
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||