Reliability of a questionnaire on sickness absence with specific attention to absence due to back pain and respiratory complaints.
OBJECTIVE--To evaluate the performance of a questionnaire on sickness absence due to back pain and respiratory disorders with the view of using sickness absence as a measure of morbidity. METHODS--A cross sectional survey was conducted among 511 male workers aged 20-65, drawn from the personnel register of an animal feed mill. The response was 404 (79%) participants. Data on sickness absence in the six months before the survey were collected by a self administered questionnaire and by sickness absence records. Correlation between both methods was examined for prevalence, duration, and frequency of overall sickness absence, back pain absence, and absence due to respiratory complaints. RESULTS--The questions about the prevalence of sickness absence from all causes, back pain, and respiratory complaints showed a high specificity of 91%, 97%, and 98%, respectively. The sensitivities of these questions were 79%, 88%, and 13%, respectively. The survey found a moderate agreement between the questionnaire and the medical register for duration and frequency of overall sickness absence with kappa values of 0.54 and 0.50. A good agreement was found for back pain absence with kappa values of 0.65 and 0.61. Respiratory absence showed a poor agreement of 0.16 and 0.13. CONCLUSION--In epidemiological studies questionnaires might be considered a valuable source of information on overall sickness absence or absence due to back pain which lasted for at least two weeks in the past six months. Whether a questionnaire survey is a reliable source of data on sickness absence due to respiratory complaints remains to be seen.