Real-world occupational epidemiology, Part 3: an aggregate data analysis of Selikoff's '20-year rule'

A number of different statistical techniques can be used to analyze potential associations in occupational epidemiology, such as the calculation of the odds ratios, correlations, or chi-square statistics. Most of these calculations, however, rely on having a complete data set, even though the collection of data is rarely straightforward in real life. In many practical situations when researching in the field of Environmental and Occupational Health (EOH), one may have knowledge of the row and column totals, for example, but have little or no information on the value of the cells themselves. This may be because such data were not recorded at the time the study was taken, the data that were collected may have been unreliable or incomplete; or simply because the data one needs to undertake a thorough analysis were not made public for reasons of confidentiality. In such situations, an analysis of aggregate data can be very useful for deriving a meaningful result from previously “incomplete” data sets