Objectives To explore methodological issues related to the use of surveillance data in burden of disease studies.

Methods and Results Burden of disease methodologies generally rely on the calculation of population attributable fraction (PAF). This in turn typically requires estimates of exposure prevalence and relative risk. An exposure surveillance system can provide the required estimate of exposure prevalence.

Alternatively, PAF can be estimated if the incidence of the disease in the population, and the incidence of the disease in the unexposed, are known. Improved disease surveillance can make an important contribution in areas where exposure prevalence measures are not good and difficult to improve. Good estimates of disease that can reliably be connected to work (e.g. occupational asthma, occupational dermatitis) in specific study populations would allow the estimate of incidence in the unexposed if the incidence in the population was known. This approach would be more difficult in circumstances where the direct connexion to work is harder to establish (e.g. many musculoskeletal disorders and noise-induced hearing loss). Alternatively, since estimates of the incidence of the disease in the general population will often be available, probably the key area of additional information required would actually be the incidence in unexposed persons. This isn’t the typical focus of occupational disease surveillance but could provide a useful contribution to the understanding of occupational disease and its burden on the community.

Thirdly, surveillance systems that cover disorders that are virtually uniquely occupational, such as a register of cases of malignant mesothelioma, can provide a direct estimate of the number of cases of a particular occupational disorder.

Conclusions Occupational surveillance systems can potentially contribute to estimations of burden of disease studies using several different approaches.