Concerns over the rising costs of health care in the United States have recently focused attention on the young discipline of health services research.
By measuring health outcomes, such as consumer satisfaction, evaluated health status, and perceived health status, health services researchers are helping to assist physicians in determining appropriate treatment strategies that are tailored to individuals. Perceived health status, or quality of life (QOL), is particularly important in determining efficacious treatment protocols.
Quality of life is a multidimensional construct, affected by both treatment and disease, that includes somatic symptoms, functional ability, emotional well-being, social functioning, sexuality and body image, treatment satisfaction, and global quality of life.[1-4]
While other measures, such as pain scales, mood scales, measures of ability to fulfill the activities of daily living, and toxicity ratings, provide useful information, they are often unidimensional in nature.
QOL measures emphasize the impact of symptoms on patients' functional status and well-being, providing a more comprehensive evaluation of the impact of illness and its treatment on patients than does a unidimensional measure.
The application of QOL studies to the clinical practice of oncology begins with the identification of baseline patient characteristics that are predictive of prognosis, response to treatment, and the likelihood of experiencing treatment-related toxicities.
This, in turn, allows oncologists to advise patients on the differing effects on QOL of different treatments with comparable survival benefits, as well as assist patients in choosing between palliative therapies that involve tradeoffs between survival and QOL.
The ability of QOL assessment to predict patient prognosis and response to treatment has been demonstrated by Ganz et al, who observed a significant relationship between QOL at diagnosis and subsequent survival in patients with metastatic non-small-cell lung cancer.
Using the Functional Living Index-Cancer (FLIC), a validated cancer-specific measure of QOL, Ganz et al found that patients scoring high on the FLIC at baseline (prior to any treatment) had a median survival of 24 weeks, compared with 11.9 weeks for patients who scored low. This finding suggests that patients with metastatic non-small-cell lung cancer and low QOL at diagnosis are likely to have a worse outcome regardless of treatment.
In less severely ill patients, such as newly diagnosed breast cancer patients, baseline QOL assessment may assist in the identification of patients at high risk for difficulties coping with treatment and later on with "cancer survivorship."
Localized Prostate Cancer
The treatment of localized prostate cancer is another area in which the results of QOL research can be used to help guide medical decision making. Nonran-domized studies of patients with localized prostate cancer have demonstrated similar 10-year disease-specific survival rates and overall survival rates (approximately 85% and 60%, respectively) for patients treated with radical prostatectomy or radiation therapy. Thus, QOL is an extremely important tool for patients deciding which therapy to pursue.
In a descriptive study, Lim et al asked all patients with localized prostate cancer who underwent either radical prostatectomy or radiation therapy to complete several QOL measures, including the FLIC and the Profile of Mood States.
Patients treated with radical prostatec-tomy had significantly worse symptoms of urinary incontinence and worse sexual function scores, while patients treated with radiation therapy were more likely to report problems with loose stools.
In both groups, problems with incontinence, sexual functioning, and bowel functioning were significantly associated with higher scores for depression, tension, and fatigue on the Profile of Mood States, so these problems do appear to impact QOL.
In a similar study, Litwin et al compared QOL outcomes in men treated with radical prostatectomy, radiation therapy, or observation, as well as with a randomly selected, age-matched control group of men without prostate cancer.
The researchers found no significant differences in overall QOL between the treatment groups or between men with prostate cancer and the age-matched controls. However, men with prostate cancer did report more problems with sexual functioning than men without prostate cancer. And problems with urinary function were reported more often in patients treated with surgery, while patients undergoing radiation therapy experienced more problems with bowel function.
Tradeoffs: Survival vs QOL
Making decisions that involve tradeoffs between survival and QOL is a frequent part of oncology practice. For many patients, treatment decisions are based exclusively upon a desire to prolong survival, while for other patients preserving QOL, even at the expense of length of life, is of paramount importance.
Randomized clinical trials that compare the effects of palliative treatments on survival and QOL can provide important information to physicians and to their patients when it is necessary to make a very difficult and personal choice regarding various possible treatments.
One trial by Chodak et al examines the differences in survival and QOL, measured every three months for one year, for men with advanced prostate cancer treated with antiandrogen therapy (bicalutamide, Casodex) or castration.
All patients reported increased physical capacity and vitality, less limitation of activity, less time in bed, and less pain, regardless of treatment. However, bical-utamide-treated patients maintained baseline sexual interest and functioning throughout treatment while patients treated with castration did not. The trade-off is that disease progression and survival analysis both favored castration.
The Measurement Tools
The measurement of QOL is a complex task that draws on the fields of social science research and psychometrics.[11,12] The tools used to measure QOL, in general, are self-administered questionnaires that have undergone extensive reliability and validity testing.
The tools currently available for clinical research include measures of general health status that can be applied to a variety of clinical situations, cancer-specific instruments, and symptom-oriented scales .
Instruments Used to Measure Quality of Life in Cancer Patients
General health status instruments
Cancer site-specific instruments
In developing or selecting a QOL instrument for use in a research or clinical setting, several methodological considerations need to be addressed to ensure the validity of the data obtained, as well as its clinical relevance.
Problem of Missing Data
Many of the cancer-specific QOL instruments have been tested primarily in research settings with adequate staff to ensure completion of the questionnaires. Use of these same QOL measures in clinical treatment trials has presented frequent problems with missing data, especially in patients with deteriorating physical status for whom there is the greatest interest in measuring QOL.
In a study by Ganz et al of QOL in lung cancer patients, the Karnofsky Performance Status, an expert rating of patients' functional status, showed that patients with the lowest performance status had a disproportionately low rate of self-administration of the QOL questionnaire, with 30% of the questionnaires being completed by the interviewer.
As patients' functional status deteriorated during the course of their disease, self-administration rates declined, and missing data was an even greater problem. Because of missing data, these investigators were unable to perform their intended comparison of quality of life between two treatment arms (supportive care alone and supportive care with combination chemotherapy).
This difficulty in obtaining complete data on quality of life has been noted in a number of trials, and stems from a combination of patient and staff-related problems.[13-15]
In July 1996, the major international cooperative clinical trials groups met at a workshop in Switzerland, with representation of methodologists and statisticians, and a forthcoming supplement to Statistics in Medicine based on the workshop will address in more detail the problem of missing QOL data in clinical trials.
Another meth-odologic concern in QOL research lies in the interpretation of differences in scores over time. In clinical trials, differences between groups of patients can be statistically significant over time. These differences, however, are not always applicable to an individual patient's scores in clinical practice. What does a two-point change in emotional well-being on a particular QOL instrument mean in an individual patient assessed four weeks apart?
Clinically Important Differences
While work has been done in trying to ascertain the minimal clinically important difference on a QOL scale for patients with other chronic illnesses, such as congestive heart failure, it remains to be established for QOL instruments commonly used in oncology.
One additional concern of QOL instruments is that they are language and culture specific. Patients' responses to illness are culturally mediated[18,19], and their perception of illness and its effect on a QOL dimension are also subject to culture-specific constructs.
Validity must therefore be established in each new patient population, which presents an important challenge in linguistically and culturally diverse cities like Los Angeles.
Important Issues Remain
The time is not too far off when oncologists will be able to order a "QOL Test" to help them evaluate a patient, much in the way they order bone scans or chest x-rays today.
Using QOL instruments in clinical practice will allow oncologists to obtain baseline information about patients and, by applying the results of QOL research as we described earlier, to help predict prognosis, plan treatment, and anticipate patients' needs for social or psychological support.
Although QOL assessment is on the verge of becoming available for use in clinical practice, further testing in the clinical setting is required, and several issues need to be addressed.
First, QOL instruments need to be reliable for individual subjects, which is particularly challenging due to the focus on group outcomes in the clinical research setting.
Quality of life instruments must also be sensitive enough to detect changes over time in an individual patient. To be applicable in clinical practice, QOL instruments must be capable of determining and interpreting a clinically significant change, while maintaining "user-friendly" characteristics.
Improving oncology outcomes requires a comprehensive, multidimensional approach, of which QOL assessment is proving to be a valuable asset.