Validity and reliability of self-reported retrospectively collected data on sick leave related to musculoskeletal diseases

Objectives The aim was to study the reliability and validity of retrospective data, collected by self-report, on sick leave related to musculoskeletal diseases. Methods The study groups consisted of 66 and 306 subjects, for the reliability and validity studies, respectively. They were all part of a wider study of risk factors for musculoskeletal disorders, the REBUS study, conducted in Stockholm in 1993. Reliability was tested using a test-retest design regarding self-reported sick leave related to musculoskeletal diseases in 1970-1993. The validity study comprised the period 1990-1994. Self-reported and registered sick-leave data related to musculoskeletal diseases were collected and analyzed regarding concordance. Data about current musculoskeletal disorders and different work-related conditions were collected and analyzed regarding possible effect- and exposure-dependent misclassification. Results The test-retest reliability study showed the percentage of agreement to be between 0.88 and 0.97, and the kappa values were between 0.73 and 0.93. The validity study of the concordance between the self-reported and registered data showed high agreement and specificity, but the sensitivity was sometimes lower. All the kappa values exceeded 0.50. No effect- or exposure-dependent misclassification was found. Conclusions The validity of retrospectively collected self-reported sick-leave data was sufficient for use as a measure of musculoskeletal morbidity in the analyses of associations with work-related conditions. Because of the relatively low sensitivity, such data will underestimate the prevalence of sick, leave and should not be used for surveys of morbidity.