On the usefulness of the diagnosis of the cause of death in church records as source material for social medicine