Assessment of quality of life during chemotherapy.

Increasingly more aggressive chemotherapy together with expected small differences between treatments with respect to objective endpoints has heightened awareness about the importance of addressing how patients experience and value the impact that treatment has had on their overall life situation. Assessment of a patient's quality of life (QoL) is now conceptually viewed as an important complement to traditional objective evaluation measures. It was therefore considered important to review the basis for the assessment of this endpoint when The Swedish Council of Technology Assessment in Health Care (SBU) performed a systematic overview of chemotherapy effects in several tumour types. The group came to the following conclusions: QoL assessments, mostly by patient self-reporting in questionnaires, have come increasingly into use during the past decade. A number of general, cancer-specific and cancer diagnosis-specific instruments have been developed. There is at present little need for development of new cancer instruments, although specific treatment modalities and tumour types may need new additional modules. A predefined hypothesis should determine the instrument to be used. Since the selection of a QoL instrument in a specific study influences both the results and the conclusions, it is essential to carefully select the instrument or instruments that have the greatest likelihood of identifying relevant differences between treatment alternatives. Interpretation of QoL data is more difficult than interpretation of objective endpoints such as survival time, objective response rates or toxicity. Despite these difficulties, QoL analyses have provided new insights into the advantages and disadvantages of various treatments not provided by traditional end-points. Some palliative treatments seemingly increase patients' QoL despite side-effects or the lack of, or marginal, increases in survival. When using potentially curative chemotherapy, it is not a matter of when the treatment should be started, but rather when it should be concluded. When using less active chemotherapy, the expected small therapeutic gains must be weighed against the QoL costs of using the therapy: does the toxicity and/or the inconvenience of the proposed treatment justify the expected gain? When it is found that the strain on the patient is greater than the effects of the cancer, treatment must be discontinued. It is not possible to determine whether or not the advantages of palliative chemotherapy are worth their costs without knowledge about patients' personal values regarding the influence on factors of relevance for QoL. The mostly used QoL questionnaires do not consider individual preferences, which therefore need to be addressed in the dialogue with the patient. QoL assessment is clearly in need of further methodological improvement before this endpoint can be regarded as fully established with respect to ability to provide unequivocally useful data in clinical trials. The multitude of questionnaires, missing data, lack of pre-study hypotheses of relevant differences between treatments and data multiplicity giving a risk for chance findings are examples of serious methodological problems. Patient response-shifts over time further complicate the interpretation of the data. Thus, QoL data, also from seemingly well-performed clinical trials, have to be interpreted cautiously. The international development during recent years has aimed at creating increased standardization of QoL measures. This has created greater possibilities to compare results from different trials. Hopefully, this also implies that it will be possible to draw firmer conclusions from QoL measurements in recently completed or ongoing trials than has been the case previously. QoL assessments are resource demanding even when short standardized questionnaires are used. Since cancer patients also generally give priority to anticancer effects over toxicity and convenience, QoL assessments in clinical trials are motivated mainly in study settings comparing treatments without expected major differences of outcome in objective endpoints.