As already mentioned on this thread, it's a very complicated question with no real answer. You can argue both for and against a current shortage, or a future shortage, or no shortage at all. It's much like asking what we should do about the current economy and inflation. We can try to predict what's going to happen, make best guess moves to try to address the issue, and never know whether what was done made things better or worse.
We could look at it from a purely numbers perspective. How many patients can a full time PCP manage? 2000-2500 is probably a reasonable number, potentially more if you have a very efficient practice. Let's use 2500 for this example. Next, what's the population of the US? That's easy - 330 million. 330 million / 2500 = 132,000. So with perfect distribution, we would need 132K PCP's to manage everyone. This oversimplifies things -- I'm lumping kids in with adults, and the census may be undercounting people, etc. But it's at least a number to start with.
So how many PCP's are there? That's not so easy to measure, interestingly. Google reports all sorts of numbers. The AMA Masterfile in 2010 reported 200K. Some estimates are higher, some are lower.
AAMC has some data, (reading more about this, it's actually all just a report of the same AMA Masterfile) suggesting 120K each of both IM and FP, so 240K together + another 60K in peds. Whether we should count NP/PA's is also complicated. And any counting system usually does not adjust for how much FTE / clinical activity they have. And whether Hospitalists count as primary care or not is also complicated. But we can see that we might actually have enough PCP's based on the raw numbers.
Distribution is often mentioned as a big problem, and it almost certainly is. Also hard to measure. The common counterargument is that PCP's are being hired everywhere, so there must be a shortage everywhere. But there's an interesting problem with that argument -- there's good evidence that healthcare is supply sensitive - adding more supply drives more demand. Nobody wants an open schedule - that's just lost income. So when a doc opens a new PCP practice and sees their first patients, their schedule is wide open. They are very likely to recommend closer follow up. Why see that pt with HTN trying to lose weight in 6 months when you can see them in 2 and try to encourage this good behavior? The same is true for MRI machines, cath labs, and any other healthcare resource -- if you build it they will come. And studies that have looked at this usually show no improved outcomes -- more MRI's just costs more money without making anything any better. So just because everyone's schedule is full and practices are hiring doesn't necessarily mean that we really need more PCP (or any other) resource.
The Baby Boomers are often portrayed as a disaster on the horizon. And they may be - a bubble of people all of similar age. But if we "staff up" to deal with them, then what happens after they all die off? Then we won't need all of those docs any more - yet if we open more residency programs we will have them in the pipeline for 30+ years. And closing residency programs rarely happens.
I've already commented on the oft quoted figure about physician ages. The AAMC data above breaks it down between over/under 55. Overall, 55% are under 55 and 45% are over. That data makes me worry that the database is not capturing when physicians actually retire -- which doesn't surprise me greatly. If I were to retire this year, how would the AMA know that? I'm certainly not updating them, and I'd likely keep my medical license -- cost is small, CME is easy, and it's nice to have. So I expect their data is too high, they are including some retired docs. Even with that, there may not be a huge shortage once we include NP/PA's. And sure people may threaten to retire early, but unclear if that will really happen or not.
Some fields have big age skews. New fields trend towards more young physicians - ED is a great example. Didn't exist as a separate field when I was a resident, so not surprising there are less older physicians in it. Other fields swing the other way and may be a sign of impending collapse - Pathology.
And then there's the question of specialization. As more fellowships open and more residents leave IM for a subspecialty, that leaves less for PCP. Is the number of specialists right? Too low? Too high? There's no central management of this -- any program can open a fellowship.
And last is asking whether just pumping out more doctors is the right answer, or thinking about the whole "system". A good example is colonoscopy. Do we really need to train a zillion gastroenterologists to do screening colonoscopies? Or should we train techs to do them, leave the GI folks to do the complicated endoscopies? But screening colo's are the bread and butter for a GI practice, mention "taking them away" and you'll get lots of screaming of how unsafe it would be.
So, complicated.