DOI link for Comparing Prevalences
Comparing Prevalences book
One of the more interesting aspects of palaeoepidemiology is the comparison of frequency of disease within and between different groups to try to make inferences about the changing pattern of disease, perhaps, suggests possible causes for any observed fluctuations. Generally, calculation of the odds ratio is to be preferred, because the age-specific prevalences of many of the common diseases encountered in human remains are likely to vary considerably. A rough and ready way of comparing prevalence is to plot out the 95% confidence intervals around a prevalence; if the confidence intervals overlap, then there is no statistically significant difference between them at the 5% level. This is usually not a very satisfactory way of going about things, since the crude prevalence is almost always a poor summary statistic. There are two principal methods of standardisation, direct and indirect. A problem of particular relevance to palaeoepidemiologists is the choice of rates to use for indirect standardisation.