Assessment of oral mucositis in clinical trials: Impact of training on evaluators in a multi-centre trial M.A. Stokman a, * , S.T. Sonis b , P.U. Dijkstra a , J.G.M. Burgerhof c , F.K.L. Spijkervet a a Department of Oral and Maxillofacial Surgery, University Medical Center Groningen, P.O. Box 30.001, 9700 RB Groningen, The Netherlands b Division of Oral Medicine, Brigham and WomenÕs Hospital, Boston, USA c Department of Epidemiology and Statistics, University of Groningen, University Medical Center Groningen, The Netherlands Received 22 February 2005; received in revised form 14 April 2005; accepted 18 April 2005 Available online 20 July 2005 Abstract In the assessment of mucositis, the inter-evaluator variability needs to be minimised and would likely to be best accomplished by training. The aim of this study was to evaluate the effect of training on concordance of evaluators in scoring oral mucositis. The evaluators were informed about the pathobiology and clinical appearance of mucositis and were trained in scoring mucositis accord- ing the Oral Mucositis Assessment Scale (OMAS). The effect of the training was evaluated by a pre- and post-training test. Each test consisted of 15 slides depicting oral mucositis. The pre- and post-training scores were compared to the reference standard. During 8 months at 6 meetings, 65 evaluators were trained. The mean percentage correctly scored slides according the OMAS increased sig- nificantly between the pre- and post-training test (P < 0.001). Training evaluators in scoring oral mucositis has a significant improve- ment on the outcome of mucositis assessment. Ó 2005 Elsevier Ltd. All rights reserved. Keywords: Inter-observer; Mucositis; Scoring; Training; Multi-centre trial 1. Introduction Mucositis of the oral mucosa is a frequent cause of morbidity in cancer therapy with a serious burden on patients. Severe mucositis causes considerable pain and discomfort, leading to a higher need of pain medication, parenteral nutrition and length of hospitalisation [1,2]. Many studies in which agents are tested with potential useful outcomes on mucositis have their shortcomings. Studies were underpowered, lacked an adequate control arm, were not investigator or patient blinded or had other major design flaws [3]. To determine the value of new agents aimed at prevention of mucositis, well de- signed, sufficiently powered and appropriately executed studies are needed. Due to the lack of sufficient patient numbers at single study site, often a multi-centre design is necessary to obtain data within an acceptable time frame. Moreover, time frames for preventive studies on sequelae of cancer therapy are very tight because of new developments or changes in ablative therapies like changes in fractionation schedules in radiotherapy or new combinations of cancer cytotoxic therapies [4]. One of the major concerns in controlled multi-centre trials with mucositis is the establishment of adequate in- ter-evaluator reliability to reduce outcome variability. In the implementation of the evaluation method in a multi- centre trial, standardisation between the different study sites and evaluators is essential for decreasing error var- iance and reducing type II error, i.e. failing to detect true differences between active drug and placebo. Further- more a poor inter-evaluator reliability decreases statisti- cal power, resulting in necessity for larger sample sizes to be able to detect significant differences between drug and placebo [5]. 0959-8049/$ - see front matter Ó 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.ejca.2005.04.030 * Corresponding author. Tel.: +31 50 3613840; fax: +31 50 3611136. E-mail address: m.a.stokman@kchir.umcg.nl (M.A. Stokman). www.ejconline.com European Journal of Cancer 41 (2005) 1735–1738 European Journal of Cancer