The medical science in DF seems to be close to the
barber-surgeon tradition, which fits dwarves thematically as they treat the body much like a machine in need of repair. Add in bloodletting for injected toxins and you're basically done.
But surgery is not the only kind of medicine. Even when most people accepted that haircuts and amputations ought to happen in the same shop, there was still a rival tradition of
physicians who specialized in drugs and therapies. This sister field to herbalists and alchemists seems like the elven way of healing. Humans did both (badly) during DF's target time period.
Getting physicians to work at a basic level would require a simple raw update: the ability to add a
negative concentration to a syndrome. Generic symptomatic relief could also be accomplished with syndromes with negative severity ratings, though this could easily get overpowered. Some of these remedies might be directly in plant extracts, while others would need to be produced using reactions with multiple reagents.
Bringing physicians on par with surgeons would require some medical procedures, in particular the ability to prescribe a specific remedy to a specific patient (rather than relying on cumbersome burrow assignments to get the poor sap to drink his medicine). A prescription for a non-existent remedy would queue up a job to produce it. Fort settings would control whether the job sticks around if the reagents are not yet available, or if the original patient recovers/dies.
Some prescriptions could be for simple therapies rather than specific medicines, for example a cold bath to reduce fever.
Adding in side-effects would be fairly simple (short-lived numbness with your cold bath, etc.), but to really get the feeling of medieval medicine the surgeons and physicians should be able to make mistakes. Physically messing the procedure up is a distinct possibility, but without modern imaging and chemical testing the most likely fault will be misdiagnosing the problem.
The only previous hit I got on misdiagnosis was
here. It's reasonable for a physician to mistake The Undulating Planes Fever for Scorchedhair Fever, but it would be silly for a surgeon to miss a compound leg fracture. What I'm proposing as a fairly equivalent system for both is to let the diagnostician generate a craftsmanship level for the diagnosis, which then governs how effective the treatment would be. For example, the same traction bench can be used appropriately or inappropriately, depending on the diagnosis informing treatment.
Does this sound reasonable to everyone? Did my search miss some megathread where all of this was already hashed out?