doi:10.1016/S0360-3016(02)04400-0 PHYSICS CONTRIBUTION PERFORMANCE OF DIFFERENT RADIOTHERAPY WORKLOAD MODELS LISA BARBERA, F.R.C.P.C.,* LYNDA D. JACKSON, M.R.T.(T),* KARLEEN SCHULZE, M.MATH.,* PATTI A. GROOME,PH.D.,* FARSHAD FOROUDI, F.R.A.N.Z.C.R.,* GEOFF P. DELANEY, F.R.A.N.Z.C.R., † AND WILLIAM J. MACKILLOP, F.R.C.P.C.* *Radiation Oncology Research Unit, Division of Cancer Care and Epidemiology, Queen’s University Cancer Research Institute, and Kingston Regional Cancer Centre, Kingston General Hospital, Kingston, Ontario, Canada; † Department of Radiation Oncology, Cancer Therapy Centre, Liverpool Hospital, Liverpool BC, NSW, Australia Purpose: The purpose of this study was to evaluate the performance of different radiotherapy workload models using a prospectively collected dataset of patient and treatment information from a single center. Methods and Materials: Information about all individual radiotherapy treatments was collected for 2 weeks from the three linear accelerators (linacs) in our department. This information included diagnosis code, treatment site, treatment unit, treatment time, fields per fraction, technique, beam type, blocks, wedges, junctions, port films, and Eastern Cooperative Oncology Group (ECOG) performance status. We evaluated the accuracy and precision of the original and revised basic treatment equivalent (BTE) model, the simple and complex Addenbrooke models, the equivalent simple treatment visit (ESTV) model, fields per hour, and two local standards of workload measurement. Results: Data were collected for 2 weeks in June 2001. During this time, 151 patients were treated with 857 fractions. The revised BTE model performed better than the other models with a mean observed –predicted of 2.62 (2.44 –2.80). It estimated 88.0% of treatment times within 5 min, which is similar to the previously reported accuracy of the model. Conclusion: The revised BTE model had similar accuracy and precision for data collected in our center as it did for the original dataset and performed the best of the models assessed. This model would have uses for patient scheduling, and describing workloads and case complexity. © 2003 Elsevier Science Inc. Radiotherapy, Workload, Efficiency, Administration. INTRODUCTION In any radiotherapy department, there is a need to balance the quantity of services provided against the quality of services provided. Health care administrators, be they pub- lic or private, want to deliver the largest volume of services possible in the most efficient way. Therapists, on the other hand, struggle with the pressure of increasingly complicated treatment plans and large volumes of patients requiring treatment. This can threaten the quality of services delivered (1). Disturbance of the balance may lead to poor morale, critical errors in treatment delivery, or an inefficient, expen- sive department. These factors are the main reason investi- gators have undertaken studies of radiotherapy workload and complexity. The specialty has turned to the rational arbiter of science to define fair expectations. In this way, all health care providers should feel that they are achieving optimal efficiency with a manageable workload. There have been two studies done using similar method- ology where patient- and treatment-related information was collected prospectively and used to construct a multivariate model that predicts treatment time (2, 3). Delaney et al. in Australia have developed a new measure of linear acceler- ator (linac) workload called a basic treatment equivalent (BTE). In this model, 1 BTE unit is the work it takes to treat a simple 1- or 2-field treatment with up to two shielding blocks. Higher BTE units represent more complicated treat- ments. One BTE unit takes approximately 10 min to treat. Burnet et al. in the U.K. have used similar methodology to develop their own model. However, most cancer centers or treatment providers have their own internal workload standards, which may or may not be based on measurements. In other cases, profes- sional associations have set standards that are adopted by individual providers (4). The purpose of this study was to assess the fit of different linac workload models developed elsewhere, as well as the standards used locally, with patient and treatment data col- lected prospectively in our own center. Reprint requests to: Dr. William J. Mackillop, Division of Cancer Care and Epidemiology, Queen’s University Cancer Re- search Institute, Kingston General Hospital, 76 Stuart Street, King- ston, Ontario K7L 2V7 Canada. Tel: (613) 548-6149; Fax: (613) 548-6150; E-mail: william.mackillop@krcc.on.ca Dr. Barbera is a Research Fellow of the National Cancer Insti- tute of Canada, supported with funds provided by the Canadian Cancer Society. Received Jul 3, 2002, and in revised form Oct 29, 2002. Ac- cepted for publication Nov 11, 2002. Int. J. Radiation Oncology Biol. Phys., Vol. 55, No. 4, pp. 1143–1149, 2003 Copyright © 2003 Elsevier Science Inc. Printed in the USA. All rights reserved 0360-3016/03/$–see front matter 1143