Can you use reasoning to determine an unknown diseases etiology based on demographics and symptoms

Factual Questions
1

Unsure if this is a dumb question (I obviously don’t know everything), but here goes.

Autoimmune diseases tend to disproportionately affect women
Chronic fatigue syndrome/ME disproportionately affects women
ergo, CFS/ME could be an autoimmune condition

The evidence seems to show this could be the case (but it might not too, plus it could be effect rather than cause).

Or inverse relations between diseases:

Rheumatoid Arthritis and schizophrenia are inversely correlated. People who get one are less likely to get the other. Ergo the etiology of one disease should help understand the other disease and create treatments.

Or matching symptoms to diagnose an unknown disease:

Infections cause fever. Patient A has a fever (as well as other symptoms that match infection like high WBC) but doesn’t have any known infections. Therefore an unknown infection is possible (if other causes of fever are ruled out).

Are there universal guidelines like this to determine potential causes of diseases or are they too vague to be of any real use to understanding disease and medicine (lots of these logical steps probably don’t work under the complexity of the human body)? Is this a valid branch of medicine or is this more just logical guidelines that could be right or could be wrong?

Granted, after you get the hypothesis based on observation and reasoning you can conduct an experiment to see if things are validated.

I guess what I’m asking is, can logical reasoning be used to understand the biology of disease, or is human biology so complex that these reasoning tool aren’t really that useful? Or are we at a point in medicine that this has its use, but all the low hanging fruit regarding diagnosis and treatment has been picked over the last 200 years and now it is better to focus on things like advanced computer models to better understand etiology rather than logical reasoning? ie, has all the simplistic A = B or A =/= B stuff related to medicine been solved, now we need more complex tools like data analysis to make sense of this instead.

2

The low hanging fruit has been picked over. What is left is complicated. The history of medicine is full of logical theories trumped by empirical evidence.

3

That was my assumption, the easy stuff is done and now there are so many variables to consider that you need data analysis rather than logical theories to make sense of it.

4

I think the answer is mostly “Yes” to all of it. Logically reasoning about a set of symptoms is called diagnosis and it’s what doctors do every day. Sometimes diagnosis is simple, sometimes it’s not.

Looking at evidence and making logical conclusions from it is what scientists/researchers do every day.

But of course, scientists then test their conclusions. And, yes, most fairly simple tests have been done already, so testing things can be complicated.

Your example about arthritis and schizophrenia being inversely correlated is absolutely the kind of thing a scientist would look at. Just that fact isn’t going to take you very far by itself; even if there really is a fundamental relationship, that doesn’t tell you much about how either disease works or how to treat them. But it might be a good clue – for instance if a particular gene has been strongly linked to arthritis, looking at what that gene does might give some insight into the cause of schizophrenia.

5

The logical guesses can still be useful in guiding you to the right experiments to conduct, or the right data to gather. It’s just that, in medical experiments, the answer is often “no”.

EDIT: On arthritis and schizophrenia, you’d also want to be careful in your controls, to try to isolate the reason for the anticorrelation. For instance, maybe that just means that schizophrenia decreases with age. Or maybe schizophrenia sufferers tend to die young, before they get old enough to get arthritis. Or maybe it’s not age, it’s year of birth, and forty years from now we’ll see a lot of schizophrenic old folks.