ORIGINAL ARTICLE Doctors’ participation in randomized trials of adjuvant systemic therapy in breast cancer: how does it relate to their recommendations for standard therapy in breast cancer? P. M. Ellis,* P. N. Butow,* R. J. Simes, † M. H. N. Tattersall, † S. M. Dunn † and C. MacLeod ‡ Medical Psychology Unit, *Royal Prince Alfred Hospital, Camperdown, NSW, Australia 2050; † NH&MRC Clinical Trials Centre, Sydney University, Mallet St Campus, Sydney, NSW, Australia 2006; Sydney University, NSW, Australia 2006; ‡ Department of Radiation Oncology, Royal Prince Alfred Hospital, Camperdown, NSW, Australia 2050 S U M M A R Y. A cross-sectional survey of all medical and radiation oncologists in Australia was undertaken, plus surgeons listed as participants of the ANZ Breast Cancer Trials Group, to examine whether doctors’ participation in randomized trials of adjuvant systemic therapy for breast cancer, is associated with their recommendations for adjuvant therapy in two clinical scenarios. Two-hundred and sixty-nine questionnaires were returned (response rate 71%). Fifty-six per cent of respondents, were participating in current adjuvant systemic therapy trials. Radiation oncologists were significantly more likely than surgeons or medical oncologists to recommend radiotherapy, while medical oncologists and surgeons were significantly more likely than radiation oncologists to recommend chemotherapy, in both clinical scenarios. In a multivariate analysis adjusting for the differences between specialist groups, respondents recommending chemotherapy were more likely to be high accruers to clinical trials (OR 3.6, 95%CI 0.93 to 13.9, P=0.08) in scenario 1, or participants of a breast cancer trials organization (OR 3.1, 95%CI 1.6 to 5.9, P=0.001) in scenario 2. Some of the variation in adjuvant systemic therapy recommendations is associated with participation in adjuvant systemic therapy trials. However, this study is unable to determine if trial participation influences opinions about adjuvant therapy, or opinions influence trial participation. © 1999 Harcourt Publishers Ltd The Breast (1999) 8, 182–187 © 1999 Harcourt Publishers Ltd 182 INTRODUCTION Randomized clinical trials or meta-analysis of randomized trials are considered the most appropriate way of assessing the efficacy of therapeutic interventions. 1 However, it has been noted previously that results from randomized trials may not be readily incorporated into routine clinical prac- tice. 2,3 Patients, therefore, may be deprived of beneficial treatments, or continue less effective treatments, because of a failure to incorporate effective therapy, based on high- quality, published evidence, into routine clinical practice. A study examining the use of thrombolytic therapy for myocardial infarction among UK regional health authori- ties, found a delay of approximately 3 years between publi- cation of the first large randomized trial demonstrating a therapeutic benefit and an increase in usage of thrombolytic therapy. 4 Of note, previous participation in multicentre trials of thrombolytic therapy was predictive of current usage. In recent years several authors have demonstrated that results of randomized trials in cardiovascular medicine can lead to rapid changes in clinical practice. 5,6 There is conflicting evidence about the impact of ran- domized trials on routine practice within oncology. Several authors, using doctors as surrogate patients, report that treatment recommendations for lung and genitourinary can- cers are influenced predominantly by specialist training and practice location and not by the results of published data. 7–10 In addition, a survey of North American doctors involved in the treatment of non-small cell lung cancer concluded that management was guided more by personal beliefs than published data. 11 However, a survey of UK medical and radiation oncologists treating lung cancer demonstrated that doctors participating in clinical trials for lung cancer are more likely to adopt trial results into their practice. 12 Overviews of randomized trials of local and systemic therapies for early breast cancer provide high-quality evi- dence on which to base treatment recommendations. 13,14 However, there is little information looking at the incorpo- ration of these data into clinical practice. This paper reports Address correspondence to: Dr Peter Ellis, Medical Psychology Unit, Royal Prince Alfred Hospital, Missenden Road, Camperdown, Sydney, NSW, Australia 2050. Tel.: 612 9515 7618; Fax: 612 9515 5697; E-mail: pellis@mail.usyd.edu.au