I am conducting a metaanalysis in R using the metafor package looking at the incidence of a disease that can be caused by different viruses and am doing a subgroup analysis to look at the incidence of the each specified virus. Is there a way to group papers that contribute data for more than one virus to ensure the denominator for the overall effect is correct? I have several papers that include huge numbers of patients and they contribute data for multiple viruses and so my overall estimate is inaccurate as the total number of patients for these papers are repeated.
Currently I have just produced separate forests for each virus but would like to have an overall figure with overall estimate and accurate estimates for each virus. When I use subgroups in a combined forest currently the overall estimate is incorrect and if I don't use subgroups then I have an accurate overall estimate but no subgroup data.