I understand that it’s hard to get to grips with…given your own past unhealthy lifestyle choices and voluntary slide into a disease state…but there really are folk who’ve practised far healthier lifestyles and for a lot longer than you’ve managed to hack so far but who’ve been dealt the sort of genetic hand that can’t be beat by healthy living. Not nohow.
Even if the profits that the drug companies make nowadays come primarily from use of, say, statins or insulin in folk who’ve voluntary eaten, sat or smoked themselves sick as you did in the past, the fact remains that these were developed initially for use in folk who did not. Folk who, in spite of living the sort of healthy lifestyle most folk couldn’t dream of handling, died prematurely…or suffered significant morbidity (arguably worse) … anyway.
The successful diagnosis of the non-obvious groups of multiple symptoms is where, it would seem, advances in AI should help, a lot. (versus every relevant provider having to learn, retain, and apply this information repeatedly - and accurately - across all ailments and patients)
Someone will have to be diligent and thorough to gather the required data, because a large collection of accurate diagnoses of such cases will be needed in the first place to train such models.
But, it should be feasible.
Of course, there is a big difference between “feasible” and “successfully and broadly implemented.”
Maybe some of the software used now already does some of this?
There’s an app in the Netherlands that diagnoses skin cancer based on a photo from your smartphone, and does so with an accuracy attained by only the most senior and experienced dermatologists.