The military medical records of the nineteenth century are a rich source, particularly because they were often collected following international agreements as to what should be included. One might therefore hope to compare the data by disease, territory, and territorial subdivisions – but this is not always possible. Medical reporting was carried out for conventional military units. Sometimes these units were too small to be statistically viable for a single year or a short period of years. The solution is to use an annual average for a series of years, but disease terminology has changed with time, and it is hard to create multiyear aggregates when the original compilers of the records changed the names of categories.

Military medical records are also different from many other archival records – at least in England and France. The official manuscript sources for political and military history, and for many kinds of social history, are carefully preserved in the Public Record Office and the Archives Nationales. Medical records were either collected and printed, or else they were destroyed. Neither the French nor the British archives kept the manuscript reports or the intermediate calculations that led to the published statistics. This makes it impossible to disaggregate data that were once aggregated for publication, or to search for records that seem valuable in retrospect but were not printed at the time.

As a result, one often has to accept the aggregates that are offered. From the 1880s onward, the British published medical indicators for the current year alongside those for the preceding five or ten years. Earlier, however, the compilers followed less systematic policies.