The aim of this study was to evaluate the validity of a questionnaire and medical anamnesis to identify persons with dermatitis in an occupational setting. The design was a clinical epidemiological cross-sectional study. The study was performed between the second and fourth week of January 2001. A questionnaire was followed a week later by a medical occupational interview and a clinical dermatological examination, including a comprehensive patch test with potential workplace chemicals. The anamnesis and the clinical examination were made independently by occupational and dermatological physicians, and the skin examination was performed blinded to anamnestic data. The setting was the mother plants of a Danish-based international company producing wind turbine systems. The study population was a workplace cohort, highly exposed to epoxy resin systems and other chemicals, and totalled 724 production workers at 4 facilities. The rate of participation was 84.7%. Using enquete questions of current skin rash against the clinical presence of dermatitis, we found a sensitivity of 22% and a specificity of 89%, compared to 45% and 87%, respectively, when the anamnestic work history, taken by an occupational physician, was the screening parameter. Using 'workplace periodic prevalence' of dermatitis, we found sensitivities in the range of 63-76% by a questionnaire and 70-83% by medical anamnesis. Questionnaire screening by skin symptoms gave the highest values for redness, a sensitivity of 33% and a specificity of 76%, and decreasing validity parameters as more symptoms were added to the list of screening questions. We found that the use of a questionnaire and medical anamnesis were problematic, when the purpose was screening for contact dermatitis and allergy, in this industrial cohort manufacturing reinforced plastic products. But these instruments might be useful for epidemiological surveillance, when the questionnaire has been validated in the given occupational setting.