The shortage of primary care physicians in the US, which I have often discussed (most recently in “When is the doctor not needed? And who should take their place?”, January 5, 2013), has become a national theme. The Robert Graham Center of the American Academy of Family Physicians (AAFP) has done much of the work in documenting this shortage, such as in the article “Projecting US Primary Care Physician Workforce Need” by Petteson, et al., discussed in my post “Health reform, ACA, and Primary Care: Is there still a conundrum?”, December 24, 2012. Essentially the problem is we have too few primary care doctors for the current population, the demand for them will continue to grow, and the rate of production (medical students entering primary care specialties) is below that even needed to replace those who retire. The growth in demand is a result of (in order of impact): population growth, aging of the population, and a more-or-less-one-time blip from increasing coverage under ACA (although for the latter, the people with a need for care were already there; it is just that with having insurance they will be able to seek it more easily).
In a recent issue of Health Affairs, Green and colleagues argue that “Primary Care Physician Shortages Could Be Eliminated Through Use Of Teams, Nonphysicians, And Electronic Communication “.[1] This is not a new concept; it is a central component of what is known as the “Patient Centered Medical Home”. The article suggests that many functions now carried out by physicians can be done by others, ranging from nurse practitioners and physician’s assistants, to nurses, to others on the health care “team”. It also suggests that many problems that now require face-to-face communication (trips to the doctor’s office) could be done by phone or “virtually”, such as by structured email or web-based visits. Thomas Bodenheimer and his colleagues in San Francisco have done much of the work in this area, most recently published in Annals of Family Medicine “Estimating a Reasonable Patient Panel Size for Primary Care Physicians With Team-Based Task Delegation”.[2]
Green, et al., use computer simulation models to estimate the “panel size” (number of patients that can be cared for per doctor) by the employment of such techniques; they add calculations for “pooling” of physicians, that is, sharing of patients among a group of doctors. This allows greater efficiency by “smoothing out the bumps” that may occur when one physicians has more or fewer patients coming in for same-day care or not showing up for their appointments by allocating them among the group. Using these statistical models they estimate that the ability of patients to access care (get in to be seen) would be dramatically increased by the implementation of such policies.
The work done by both the Green and Bodenheimer groups is convincing, and provides a model for more efficient primary care practice that would help to address the problems our country faces from having too few primary care doctors. Indeed, these approaches utilize the “crisis” as a way to actually improve both access to and quality of patient care. There are, however, challenges to implementation of this model. One is payment; while health systems in many parts of the nation have demonstrated that it is possible to restructure their practices to achieve these advantages, this is most effective in settings in which the provider is also the insurer (notably Kaiser). In those parts of the country where this model of care is less prevalent, where most payment to medical providers is “fee for service” for face-to-face visits to doctors, there is not only no incentive to change, there is a large negative financial incentive since any non-face-to-face care is, essentially given out free.
A second challenge is that such models only work where there is a large enough concentration of patients and providers to achieve the benefits of scale; as with most such analyses, it leaves out the needs of rural populations. Some large systems, such as Geisinger in Pennsylvania, have been successful in creating such efficiencies in their clinics in rural areas, but Geisinger is atypical; there are not many like it. In addition, it is a financially integrated system (like Kaiser) – that is, it is also the payer -- and it works in a relatively-densely populated rural area of northeastern Pennsylvania, not like the vast empty frontier counties of the West.
It is interesting to me that so much of this emphasis on efficiencies, and particularly the use of professionals other than physicians to provide care, has been on primary care. This, I am sure, is due in part to the need for primary care in all settings, while much specialty care can be centralized in larger cities. It is also because there is not a shortage of many non-primary-care specialists for the needs of the population (although there are for some, such as general surgery, especially in non-urban areas). The reason usually given for this non-shortage is largely that these specialists make so much more money than primary care doctors, so medical students are attracted to them. To the extent that some specialties also have more regular work hours and a limited scope of work, it may also increase their attractiveness.
The limited scope of work (although not, necessarily, less difficult work, especially when considering surgical interventions) also makes them, in many ways, more appropriate fields to use non-physician professionals than primary care. This is the reverse of the usual assumptions that sub-specialists are seeing difficult problems, while primary care providers see mostly colds and blood pressure checks. In fact, primary care is complex, as it sees both undifferentiated patients and those with multiple chronic diseases. Most specialty care is more routine, seeing a much more limited set of diagnoses with a more limited set of interventions; for the typical subspecialist, less than a half dozen diagnoses may account for 80% of visits, while for a family doctor the top 20 are probably 30%. Thus, the breadth of knowledge and skills in making complex decisions and appropriately prioritizing problems, require a level of sophistication and training not taught or developed in most other health professionals (family nurse practitioners are one other provider group where there is at least an effort to have this breadth of training). It is, then unsurprising that most of the tasks suggested for nurses and others to increase the efficiency of primary care practices have limited scope: maintaining disease registries, calling for recommended preventive care, screening a small set of diagnoses.
This type of narrow, in-depth scope of work is much more characteristic of subspecialty care, and it is one of the reasons why expanded-scope nurses and physician’s assistants have found so much use in these practices. They follow people with congestive heart failure for cardiologists or diabetes for endocrinologists, they manage chemotherapy recipients for oncologists, they use algorithms to care for people in intensive care units, they do pre- and post-operative care for orthopedists and other surgeons. And they do not go outside of the set of diagnoses and treatment options with which they are familiar; following the model of the physicians with whom they work, when a patient’s problem is not in their narrow area, it is referred.
The targeted but limited expertise of such nurse specialists have explains why they function so well clinically in subspecialties. What explains why it works financially is that the doctors (or hospitals, or health systems) that employ them are reimbursed at subspecialist physician rates (already very high) for work that is done by others; thus they can afford to pay such “physician extenders” relatively well compared to folks working in primary care. Reimbursement for “teams” follows the model of reimbursement for physicians: care for a limited set of diagnoses in a detailed way, especially when it involves procedures, is paid much better than management of complex sets of interactive diagnoses.
Unfortunately, the problem with such practice is challenging because the same person often has multiple conditions, and interventions that help one may make another worse. While efforts to build teams, and have each professional work at the “top of their license”, is important, so is payment. As long as primary care is reimbursed at lower rates it will continue to face challenges in recruitment of physicians, nurses, and other team members.
We need to develop and implement great strategies for team-based care. We also need to dramatically decrease the ratio of income for subspecialists and their subspecialist teams relative to those working in primary care.
[1] Green LV, Savin S, Lu Y, “Primary Care Physician Shortages Could Be Eliminated Through Use Of Teams, Nonphysicians, And Electronic Communication”, Health Affairs, 32, no.1 (2013):11-19
doi: 10.1377/hlthaff.2012.1086
[2] Altschuler J, Margolius D, Bodenheimer T, Grumbach K, “Estimating a Reasonable Patient Panel Size for Primary Care Physicians With Team-Based Task Delegation” , September/October 2012 vol. 10 no. 5396-400 doi:10.1370/afm.1400
As far as team based care - only 20 - 30% more primary care revenue above primary care delivery costs will retain the health care team members as a team. Stagnant primary care spending plus increasing costs result in problems for health care teams. Members can leave as fast as they can be untrained and trained again in team based care.
ReplyDeleteNP and PA have highest turnover already, making 4% more income for changing to a different primary care practice and over 10% more when leaving primary care. Squeezing practice revenues squeezes the workforce that can deliver primary care - squeezing out clinicians.
With only one-third of NP and PA in primary care and falling (for the two-thirds that are active direct care clinicians), the question is already answered. NP and PA are even better at generating subspecialty revenue - for specialists and for health system employers.
The current JAAPA has a survey of PA dermatologists and AAPA has pdf files. The revenue generation reaches $600,000 after 4 years of experience. The costs are about $100,000 to 110,000 in PA income or at the top with cardiothoracic PA.
This margin is so high that dermatologists would have to think hard about hiring another physician to do dermatology. They would cost too much and cut into profits and take away the more interesting cases referred to them by their PA colleagues.
NP and PA are moving to more specialties with more in each specialty - leaving employed family practice behind at below 25% and falling.
The only significant health access solution for 200 million Americans with lower to lowest local or adjacent zip code workforce is employed family practice (MD, DO, NP, and PA. When NP or PA grads depart family practice or fail to enter, they join specialist practices in locations with top concentrations leaving primary care and primary care where needed behind.
NP and PA are not strong solutions for primary care. FM is a strong solution - not expanded for 32 class years.