Comparison of Self-Administered Questionnaire with Physician Diagnosis in the Diagnosis of the Sick Building Syndrome

Abstract

The aim of the study was to compare the results of a self-administered questionnaire, designed to estimate the prevalence of the sick building syndrome, with an independent medical interview and clinical opinion. Six buildings were chosen for study from a sample of 47 with known measurements of the building symptom index. One building with a high and one with a low score was selected from each of three ventilation classes (natural ventilation, air-conditioning with induction units and air-conditioning with variable air volume systems). A stratified random sample of 160 workers in these buildings was studied. Each received, in random order, a self-administered questionnaire, and a medical opinion based on a free medical history. The doctor had no access to the self-administered questionnaire at the time. The average number of work-related symptoms per worker (the building symptom index), which is used to compare one building with another, showed a good agreement between the two methods. There were, however, consistent differences between the two assessments in individual symptoms. The self-administered questionnaire produced a higher prevalence of work-re- lated runny nose and flu-like symptoms, which were often regarded as being due to infections in the medical opinion. In individuals, work-related symptoms on the self-administered questionnaire were validated by the medical opinion in over 75% of cases for eye and throat symptoms, lethargy and headache. Only 31% of work-related runny nose and 21% of work-related flu-like symptoms were thought by the medical opinion to be work-related. The medical opinion identified an extra 5% of work-related symptoms that were missed on the self-administered questionnaire. The self-administered questionnaire therefore produced a satisfactory estimate of the building symptom index, removing the potential bias of an interviewer. The questions on ninny nose and flu-like symptoms would be improved by including only those that occurred more frequently. The building symptom index was calculated for the six buildings twice, with separate random samples of workers completing the questionnaire two years apart. The buildings were ranked in the same order (for the building symptom index) on both occasions, again confirming the validity of the self-administered questionnaire.