S18 Posters patients (pts), in order to complete its clinical evaluation. Contrary to conventional grafi, this intensive chemotherapy may be administered in outpts. Methods: Pts were included in this study between l2/94 and 09/96. Physical symptoms due to chemotherapy were assessed in terms of frequency, duration/severity and distress using a self-administered questionnaire including 19 side-effects and completed by pts at the 4th cycle. The multidimensional QL was evaluated by means of the EORTC QLQ-QO administered : (a) prior the start of the treatment (b) at the 4th cycle of chemotherapy and at the end of radiotherapy (c) 1 month after the treatment and every year until 3 years. Variables for the evaluation of monetary costs were collected both in case report forms and by using a specially designed indirect cost form completed by pts at the end of the treatment and during the 3 years follow-up. m: 100 pts were included in the protocol. The current estimation of the return rate of questionnaires is 84% during the treatment and 65% during the follow-up. At the 4th cycle, tiredness, alopecia, lack of appetite, nausea, vomiting, change in taste, fever and weight loss were reported by 96 % to 63% of pts. Most of symptoms were distressing for patients, the most distressing being mucitis, vomiting, change in taste and stomach pain. At the end of chemotherapy the scores of the EORTC functioning scales (physical, role, cognitive, social and global QL) were statistically significantly lower than baseline scores. At the end of radiotherapy, these scores regained baseline values except for physical and role scores (p<O.O5). One year after inclusion in the treatment, all functioning scores were not statistically different from pre-treatment values excepted role score (P’O.05). Discussion: The protocol proposed to pts had an important adverse effect on QL, but this effect disappeared at the end of radiotherapy for several QL dimensions and for nearly all QL dimensions 1 year alter inclusion. The cost study will be performed on the 100 pts, when all case report forms will be filled, to assess the ratio cost-effectiveness of this therapy. These results will be compared with those of 2 historical groups of IBC pts treated by high dose chemotherapy with conventional graft or by conventional chemotherapy + radiotherapy sequence. Bouscary ML, INSERM U379, Paoli-Calmettes Institute, 232 Bd Sainte- Marguerite, 13273Marseille Cedex 09, France PPlO. Economic evaluation of adjuvant treatment for early breast cancer Brown J’, Karnon J’, Eldabi T2, Paul RJ2 ‘Health Economics Research Group, Uxbridge, UK; ‘Centre for Applied Simulation Modelling, Brunei University, Uxbridge, UK Background: The Adjuvant Breast Cancer (ABC) trial is a national collaborative randomised controlled trial to determine the value of adding cytotoxic chemotherapy and/or (in pre/perimenopausal women) ovarian suppression to prolonged adjuvant tamoxifen. The cost-effectiveness of this potential advance in cancer treatment will be critically dependent on a number of factors, including patient selection, the methods of treatment delivery and the impact of the treatments on quality of life. Methods: The approach taken was that of a pre-trial modelling exercise, using available data and ‘expert’ opinion, to determine the key determinants of the cost-effectiveness, Further data collection can then be concentrated on these key parameters during the trial. Thus data collection during the trial is kept to a minimum. A discrete event simulation model has been built describing the treatment pathways and possible progression of the disease. The events modelled are the administration of adjuvant treatment, local- regional recurrence, bone and non-bone metastases. Four categories of both menopausal and toxicity side-effects of the adjuvant treatments have been modelled as attributes, and cost and quality of life attributes have been attached to the events, The study is now at the stage of testing the sensitivity of cost-effectiveness to the values of the model parameters. Discussion: Although the methods presented are applied specifically to the ABC trial, the approach is applicable to economic evaluations integrated into clinical trials generally. one of the main advantages of the approach is seen as reducing the time and effort needed to collect large amounts of data and thus minimising disruption to everyday clinical activity. It does, however, require close collaboration between the health economists and modelling analysts. Brown J, Health Economics Research Group, Brunei University, Uxbridge, Middlesex UB8 3PH, UK PPll. Summary recommendations from the conference on purchasing oncology services: Methods & models in the marketplace Cygfalose C$%‘, Ballard DJ’, Hayes RP’, Blair AE’, Borchardt JS’, Ades T , Liberati A ‘Emory Center for Clinical Evaluation Sciences and the Kerr L. White Institute for Health Services Research, Atlanta, Georgia, USA, ‘American Cancer Society, Atlanta, Georgia, USA, ‘Mario Negri Institute, Milan, Italy Background: The annual cost of cancer care (prevention through terminal care) in the United States has been estimated to exceed more than $ 100 billion. The aging population is expected to significantly increase cancer incidence and associated costs. Accordingly, there is considerable concern about the high cost of oncologic services among purchasers and suppliers. Furthermore, there is currently very limited information regarding the clinical content of and quality of care reporting requirements within risk/capitated oncologic service contracts. In order to improve information flow between purchasers and suppliers of oncologic services, the American Cancer Society and the Kerr L. White Institute for Health Services Research have organized a multidisciplinary conference to synthesize clinical evidence, patient preferences, system performance and resource constraints to develop recommendations to assist health care purchasers in making better informed decisions when procuring benefits packages for oncologic services. Methods: A two-day conference will be held in Chicago, Illinois, on September 1l-12, 1997 that will include presentations and discussions focused on developing recommendations for purchasing oncologic services. The conference attendees will include health care purchasers and suppliers, clinicians, insurance companies, and members of the academic community. A multidisciplinary task force is responsible for planning the conference and writing the summary recommendations and papers describing current methods and models for purchasing oncologic services. These papers will be published in the peer-reviewed literature in a supplement to the journal Cancer. The task force members will include purchasers of oncologic services including government and private sector employers; suppliers of oncologic care such as clinicians, managed care organizations, and cancer centers; insurance companies; evidence-based health care methodologists; and consumer/advocacy groups. Results: The conference will conclude with a presentation of the summary purchasing recommendations and discussion. This summary report is intended to aid purchasers by recommending basic guiding principles to be used in defining oncologic benefits packages and will incorporate a purchaser’s view from a population-based perspective. At the First European Conference on the Economics of Cancer, we propose to describe the methods and models for purchasing oncologic services that are being employed across the United States and the implications of the American experience for the European Community. Cangialose CB,IOI W. Ponce de Leon Avenue, Suite 610, Decatur, GA 30030, USA PP12. Decision model for the cost of using opioids to treat cancer pain Guest I’, Hart W’, Fallon M*, Cookson R’ Catalyst Healthcare Communications, Pinner, Middlesex, UK; *Palliative Medicine Department, Western Infirmary, Glasgow, UK; ‘Janssen-Cilag Ltd, Saunderton, Buckinghamshire, UK Background: Patients requiring treatment for cancer pain often require strong opioids. A number of different drugs and delivery systems are available and patients may be switched between different forms during the course of their illness. The acquisition cost of opioids represents only a small part of the true cost. We are therefore developing a model to