How to calculate incidence rates from proportionate data

This paper describes the methodological aspects of calculation of incidence rates from incomplete data in occupational epidemiology. Proportionate measures in epidemiological studies are useful e.g. to describe the proportion of slips, trips and falls compared to other types of injury mechanisms within single age-strata. However, a comparison of proportions of slips, trips and falls among the different age-strata gives no meaning and can hamper the conclusions. Examples of a constructed example and some selected studies show how estimates of incidence rates can be calculated from the proportionate data by applying estimates of denominators available from other information. The calculated examples show how the risks based on the incidence rates in some cases differ from the risks based on the proportionate rates with the consequence of hampering the conclusions and the recommendations for prevention. In some cases the proportionate rates give good estimates of the incidence rates, but in other studies this might cause errors. It is recommended that estimates of the incidence rates should be used, where this is possible, by estimation of the size of the population. The paper is intended to be useful for students and teachers in epidemiology by using the attached Excel training file.