Well, I understand (and agree) that in order to solve real-life problems we have to simplify the reality down to the models we can reasonably operate on without having too many variables, simply to avoid “analysis paralysis”.
What I’m challenging is whether the current models explain the reality well enough to cover rare cases, like anatomical variations, psychological or autonomic system disorders, rare health conditions, genetic mutations and similar “deviations”.
A lot of these medical models are based on data collected at the laboratories, but have you ever carried real-time body monitoring devices? E.g. Holter monitor? They can show fascinating things, when you don’t feel any noticeable effect but your heart starts skipping beats. Or, for example, you start feeling totally dizzy and lack of blood flow in your brain, but neither blood pressure monitor nor data from Holter shows anything happening.
These are real-life data points but we still have very few continuous monitoring devices for everyday use, with the situation just recently starting to improve with wearable devices like smartwatches constantly monitoring SpO2, heart rate, skin temperature, and in some cases, perfusion index and estimated blood pressure.
I wish there was a simple scarf-like gadget performing continuous Doppler scanning on arteries (veins would be a bonus, but there are too many of them) in the neck and estimating blood supply 24/7 under normal physiological loads. Maybe one day we will have it… I know that now there are experimental compact implantable CSF pressure devices recording CSF pressure all the day, so there is some progress towards uncovering more details in how the body works.