This report by The Swedish Council on Technology Assessment in Health Care (SBU) reviews, classifies, and grades the scientific literature on cancer chemotherapy in some major tumour types, describes the practice of chemotherapy in Sweden, compares practice with scientific knowledge, and analyses the costs and cost-effectiveness of chemotherapy. The report is intended primarily for decision-makers at various levels, both practitioners and administrators. It is also of interest for the medical profession. The extensive body of scientific literature was reviewed according to strict criteria that reflected the scientific weight of the literature. Sixteen experts representing different disciplines (oncology, surgery, internal medicine, health economy and quality of life research) participated in the literature review. Each section was discussed within the project group and was reviewed by at least one, but usually two international researchers. Additional input was provided by national experts representing different scientific disciplines. For the final evaluation to be as close to the objective truth as possible, a concerted effort was made to guarantee objectivity and thorough assessment of current knowledge about the effects of chemotherapy on the selected cancers. The tumour types selected for this assessment include firstly those types where three investigations had shown an increased use of chemotherapy in Sweden during the latest decade. These were non-small cell lung cancer (NSCLC), gastric cancer, pancreatic cancer, colorectal cancer and urinary bladder cancer. Secondly, the two tumour types comprising the greatest number of patients treated with chemotherapy in Sweden, breast cancer and haematological malignancies, were included. Among the haematological malignancies, the most prevalent ones, acute myeloid leukaemia (AML), chronic lymphocytic leukaemia (CLL), Hodgkin's disease (HD), aggressive non-Hodgkin's lymphoma (NHL) of the large B-cell type and indolent NHL of follicular type were evaluated. These constitute about 75%, of all haematological malignancies. Thirdly, ovarian cancer was included since chemotherapy has been extensively used and since, at the time of the planning of this overview, a group of very expensive drugs, the taxanes, had preliminarily shown promising results. A wealth of scientific literature has been published on cancer therapy. The review presented in this report is limited to scientific studies judged to be important for evaluating chemotherapy efficacy. Assessments of the content and quality of these studies, and a critical summary of the results in all stages of the selected tumours, have never before been attempted in this way. However, similar comprehensive overviews of certain stages of the tumours have previously been made. These overviews were also critically evaluated. Totally 1,496 studies involving 558,743 patients were reviewed. The survey of practice of chemotherapy use involved all departments of surgery, urology, gynaecology, internal medicine including haematologic units, pulmonary medicine and general and gynaecologic oncology at 16 hospitals in two health care regions in Sweden, covering 39% of the Swedish population. During the 4 weeks of the survey, all patients with the diagnoses concerned who received chemotherapy were registered. The study included 1,590 patients. The working group's general conclusions are summarised in the following points: The literature on the effects of chemotherapy is extensive. Chemotherapy has a well-documented role in the curative and palliative treatment of patients with several types of cancer. The use of chemotherapy is of utmost importance for the possibility of cure in certain tumour types. In other tumours, chemotherapy increases the possibility of cure when added to local and regional treatments, particularly surgery. In the instances of no possibility of cure, chemotherapy may to a variable extent improve both patient survival and well-being. In Sweden chemotherapy is largely used in accordance with that documented in the scientific literature. The extent of both over- and under-treatment seems to be limited but cannot be excluded at the individual patient level. The literature-based knowledge is scientifically of lower quality in the most chemotherapy sensitive tumours than in tumours showing more limited sensitivity. In the more sensitive tumours, positive effects on a symptomatic stage and survival were seen several decades ago. In those days, clinical treatment studies did not fulfil the current high quality requirements. Small life-prolonging effects of chemotherapy are sometimes very well documented in large, high quality scientific studies. Some of these s