Training examiners for a national epidemiological survey of oral mucosal lesions.

Ensuring the validity and reliability of data collected in epidemiological surveys is an important consideration. The purpose of the present report is to describe a training and calibration programme for 16 examiners taking part in a national survey of oral mucosal lesions and to present an evaluation of the results. The programme included the distribution of a pictorial manual to participants and a series of lectures followed by three diagnostic sessions, two using slides and the last involving patients. At the final session, the trainees classified 88 per cent of 16 patients correctly in comparison with the definitive diagnoses of the trainer, and their sensitivity on recording oral carcinoma, leukoplakia and lichen planus was at least 0.88. However, correctly classifying submucous fibrosis on the basis of slides alone proved problematic. At the conclusion, the diagnostic accuracy of two examiners for all types of lesion remained appreciably lower than the majority. Training strategies for various types of study are discussed. The method reported is considered to represent a model approach to training and calibrating examiners for this type of survey work.