In this study, the applicability of a symptom-based questionnaire on hand dermatitis was assessed in a population of rubber workers. The questionnaire was previously validated in a study among nurses. 224 subjects employed in 9 different companies completed a questionnaire on skin complaints. Subsequently, 202 workers attended the physical examination of the skin by a dermatologist. The ascertainment of skin complaints according to the questionnaire was compared to the medical evaluation. The 2 different diagnostic tools used for assessing dermatitis resulted in dissimilar estimates of the prevalence of active hand dermatitis, ranging from 6.9% to 38.1% of all workers. Using the medical evaluation as 'gold standard' we observed a moderate sensitivity and specificity (respectively 71.4%; 95% CI: 47.7-95.1 and 76.1%; 95% CI: 70.0-82.2), a low positive predictive value (18.2%; 95% CI: 8.0-28.4) and a high negative predictive value (97.3%; 95% CI: 94.7-99.9) for the classification based on the self-administered questionnaire. When evaluated against 'first symptoms of dermatitis' the sensitivity decreased, while the specificity remained almost the same. The deviant findings between the present and the original validation study of the same questionnaire among nurses hamper its applicability in populations with different occupations. Therefore, if questionnaires are to be used, validity studies have to be carried out to evaluate differences in perception of skin diseases between different (occupational) populations.