A major concern of policy analysts regarding the Affordable Care Act is whether and how the country will be able to produce a sufficient supply of primary care physicians (PCPs) to meet the projected demand arising from extending healthcare coverage. But to what extent future demand for PCP services will be owing to demographics versus expansion in coverage requires the use of some rather subjective assumptions. While it is plausible to assume that removal of cost as an obstacle to healthcare utilization would increase demand among that portion of the population unable to afford coverage, such thinking can also be counterintuitive.
According to a 2012 article published in the Annals of Family Medicine, Projecting US Primary Care Physician Workforce Needs: 2010-2025, “with nearly 209,000 PCPs in 2010, the United States will require almost 52,000 additional PCPs by 2025—about 33,000 to meet population growth, about 10,000 to meet population aging, and about 8,000 to meet insurance expansion.” There are numerous similar studies using different methodologies and approaches and different (hypothetical) assumptions, but most all I have seen support the challenging reality that demand for PCP services is going to substantially outpace supply given the historical rate at which new physicians enter the workforce.
In reaction to this concerning challenge, the journal Health Affairs recently published a paper that argues the projected PCP shortage can be largely addressed by using teams, better information technology and sharing of data, and non-physician professionals (i.e., physician extenders, such as Registered Nurses, Physician Assistants and/or Nurse Practitioners). I fear again, this may be a situation where the reliance on subjective assumptions produces desirable findings from sound research practices that won’t bear out over time.
I think it also illustrates – and this is really the larger point I wanted to make with this post – where very often healthcare policy research methodologies inherently rely upon linear dynamics to study problems that really require a nonlinear dynamics approach. And understandably so. If you want to produce a movie using a still frame camera, you had either be extremely fast or quite imaginative. You work with the tools at your disposal.
As advances continue in information technology computing power and capacity (i.e., Big Data), the ability to model nonlinear relationships will increase. But the nature of unpredictability in human reactions to environment and circumstances will still be a difficult challenge. There is quite a body of interesting literature suggesting ways in which nonlinear dynamics (e.g., Chaos Theory) can be adapted in social policy research, which is well beyond my purpose here. But to be sure, the observations I offer on the subject are neither unique or original.
As a more practical matter, however, I think the ideas presented in the Health Affairs paper are viable and will probably result from being as much a function of necessity as requiring support of public policy. But the nature of how these clinician-patient relationships form and whether or not they will be sufficient to meet the projected demand for PCP services really cannot be predicted because of the modeling constraints of linear dynamics.
Unfortunately, there are usually significant limitations to what healthcare policy research can offer in terms of predicting the future benefit of what appear to be good ideas. On the other hand, fortunately, the lack of a projected empirical benefit has not been an obstacle to the pursuit of good ideas throughout the history of mankind. The historical resolution of these two realities has always been the economic reward for the risk taken in pursuit of an idea that lacks a demonstrable benefit. The challenge we face today is our inability to accept the consequences when that pursuit does not bear fruit.
We love being rewarded. Paying the Piper – not so much.