Razib Khan and the commenters at his site are pretty smart. But in a recent post called Up with nurses! Down with doctorates!, the commentary fell into a common error. That is, the assumption that greater medical training and ability is necessary to deal with more serious diagnoses.
Razib's original point was that "degree creep" is a bad thing that hurts the economy and adds nothing to the medical field. Basically, he's right. Charles Murray said it a long time ago, and John Durant also points us to a Peter Thiel interview where he says much the same about an "educational bubble."
However, Razib's commenters then riff on his point by musing that we could use NPs and PAs for simple family medicine, whereas MDs are needed for more serious medicine. The commenters even mention dermatology as a specific area not requiring a medical doctor. The truth is probably almost the opposite.
If you're having a heart attack, and especially if you're having a heart attack with cardiac arrest, in terms of diagnostics, the most you need to deal with the problem effectively is non-degreed emergency medical personnel who have gone through nothing more than a training course. But if you have a skin lesion, an EMT isn't going to have any idea what the possible differentials and causes are.
Most people would probably have trouble with this idea because a heart attack is scary and we've been trained to turn to MDs to deal with fear and doubt rather than for expertise.
Once you move beyond immediate life-saving diagnostics, you wouldn't want an EMT managing your cardiac care, but even at that point, is an MD necessary? Let's say you need a cardiac catheterization. Having seen several of these and managed the patients post-procedurally, I doubt an MD is necessary. I suspect a BS in biology or nursing with a two-year Masters focused completely on cardiology and catheterizations could perform them as effectively as most cardiologists. And the outstanding cardiologists are that way from intrinsic skill factors and experience, not training.
Once you get beyond the catheterization lab and into the territory of secondary level prevention and long-term management of heart failure, etc, the issues get more complicated, but with increasing technology making a memory for drug interactions, etc less strictly necessary, and with the increasing team orientation of health management, I'm not sure a cardiologist is called for universally.
The theme here is that training (and especially the experience of residencies, etc) is needed to deal with complexity and vagueness, not with severity.
One of Razib's commenters complains about a misdiagnosis of an intestinal parasite. Ironically, this is the type of problem you deal with in simple "family medicine." When you have an office practice, anybody can walk in the door with anything from a heart attack to domestic abuse to a stubbed toe to a tropical disease to end-stage cancer.
When your toddler is sick with vague symptoms, do you really want a two-year associates trained nurse with a three-year online "bachelor-to-masters" practitioner degree trying to figure out if the kid has a mental illness, environmental poisoning, or an infection?
This is the reason doctors get out of family practice: it's very hard, doesn't pay as well, and doesn't get the respect. Why would you stay?