The New United States Workforce Designs
Robert C. Bowman, M.D.
Summary - More of the Same Designs Still Fail in Health Access, Primary Care, Care of the Elderly, Rural Health Care, Care of Lower and Middle Income Americans, and Care of the Underserved
The first decade of this century has given more than enough time to consider physician and non-physician workforce designs for primary care and for specialty care. These designs from the past have shaped the present and will shape workforce for the first half of the century even if the design is changed.
There are a number of complex approaches regarding workforce estimates, but simple measures suffice. These measures readily divide workforce into primary care and non-primary care workforce. Consistency in the designs for primary care training and the designs for primary care practice support allow the levels of physicians and non-physicians to be estimated for primary care for decades to come. The same consistency also limits primary care in 2040 to much the same level as found in 1980.
Common sense workforce measuring tools such as the Standard Primary Care Year and calculations of maximum primary care graduates age 30 65 illustrate that the United States workforce in 1980 and in 2000 will be much the same as in 2020 and in 2040. The level will remain about 225,000 primary care physicians worth of primary care.
These findings illustrate the fact that the United States has not had a significant change in the design of training or in the design of policy support for primary care since the 1970s decade. The primary care workforce design still fails to emphasize permanent primary care that remains 90% or more in primary care for 35 years of a career. With 22,000 of 28,000 annual primary care graduates emerging from flexible primary care training forms that graduate two to eight or more specialists for each primary care graduate, the United States design for primary care has been converted substantially into a specialty workforce supplier. The workforce design is not the only problem. The health care funding design continues to fail to support primary care facilities, primary care health care team members, and primary care practitioners. Most of all the design fails to support basic health access for more than a majority of the population of the United States in one or more dimensions such as geographic access or financial access.
The United States will continue to maintain the same primary care physician numbers despite the doubling of the population from 1980 to 2040 and the doubling of the old, older, and oldest populations that require multiple times more primary care. Simple incremental increases in primary care reimbursement will not address needed design changes or reverse the departures from primary care.
The poorly coordinated workforce design now produces record levels of specialty care with even more annual graduates coming due to expansions. Additional expansions of graduate medical education offer even more opportunity for choices outside of family medicine and primary care. In some cases primary care training positions are being converted directly to specialty positions which garner significantly more funding under the graduate medical education design.
Primary care is also substantially out of position with regard to the populations in need of primary care. Five forms of primary care compete for positions in 3400 zip codes that have top concentrations of physicians as well as saturations of primary care. Insufficient primary care at 30 60 primary care physicians per 100,000 is found in 28,000 zip codes outside of concentrations of physicians that also have 65% of the United States population. About 61% of primary care is found in 3400 zip codes with 35% of the population while 39% of primary care is found in 28,000 zip codes with 65% of the population.
Family practice forms distribute equitably when practitioners remain in family medicine and all other primary care forms concentrate together in existing concentrations. Internal medicine and pediatric training programs contribute only 30% of primary care graduates to zip codes outside of concentrations as compared to family practice MD, DO, NP, and PA forms that match up much better with 53 60% found in zip codes with 65% of the population. Grossly insufficient, poorly coordinated, and out of position represents major design problems with regard to primary care, stroke centers, heart attack centers, and other health needs for the majority of the people of the United States, especially the elderly.1, 2
With primary care fixed in place, the new workforce design results in every greater numbers entering specialty care. The increase is the result of higher percentages entering specialty care and expansions of the number of physicians entering the workforce annually. The growth rate of specialty physicians continues to exceed the growth rate of the United States population and the growth rate continues to increase. Each source of specialty workforce now contributes the highest levels of specialty workforce in the history of the United States. Each source has increased in production by 50% or more in the past decade alone. Allopathic United States medical school graduates and international graduates have increased beyond 80% specialty workforce. Osteopathic graduates have closed the gap with and increase to 72% specialty workforce. Non-physicians have converted from 70% primary care workforce to 70% specialty workforce. Physician assistant annual graduates have doubled in the past decade and the primary care percentage of entering graduates has increased from 46% to 72% entering specialty care. This combination represents an increase of 180% with regard to annual graduates entering the specialty workforce. There is no indication that specialization rates have stabilized in any source indicating even higher percentages to come.
The current directions of physician and non-physician workforce, workforce research, selection of physicians, training design, and health policy design can be summed up as more exclusive and less normal. More normal distributions of funding, more normal physician origins, and more normal physician training all result in better distributions physicians and non-physicians. More exclusive funding, origins, training, and policy design all result in higher concentrations of physicians, non-physicians, people, income, and economics. More exclusive is winning and most Americans are losing out in the current health care designs.
The steady progressive changes over time should be causing great concern on the part of physicians who hope to preserve respect for the physician form of health care delivery as well as those hoping to reverse disparities in health access, cost, and quality.
The current approach to workforce prediction often involves projections of workforce based on the patterns of recent graduates. Projections based on 1990s changes are grossly inaccurate for one major reason. The 1990s represented an atypical policy period - one of very few favorable to primary care. One example would be the doubling of Medicaid spending from 1990 to 1995 with the focus placed in increased eligibility one of the ways that the nation spreads out health care support beyond current concentrations. Small reforms in primary care reimbursement coupled with cuts in specialty reimbursement represented a shift in direction. The new workforce design based on managed care was a major factor driving medical students away from hospital based careers that were expected to be in excess. Few anticipated the short duration of these changes. It was a brief natural experiment illustrating increased choice of primary care careers. Not surprisingly the government projections (assisted by organized medicine) of 2020 primary care based on 1990s primary care resulted in predictions of a 21% increase in physician primary care numbers from 2000 to 2020.3
These workforce projections are in error at a very critical time in the history of the United States. The major reason for the error involves internal medicine departures from primary care that are already evident at the national, state, and local level. By the end of the last century and since the beginning of this century, steadily more have departed primary care internal medicine than have entered. 4-9 Other methods must be used to determine workforce when those trained in primary care programs depart primary care during training, at graduation, and each year after graduation.
Various practices such as the use of first career choices (National Residency Matching Program) and first practice choices should not be tolerated. These reporting methods do not capture a representative cross section of national workforce. Often the first practice choices are the most atypical of all as graduates spend a few years in primary care or rural or underserved careers before making more representative choices. This allows medical schools, nurse practitioner programs, and physician assistant programs to keep the appearance of much higher primary care contributions than are found for the average graduate over a career.
It is common sense that types of primary care such as family medicine with 35 or more year careers with top activity (fewest part time or inactive or departed the United States), top retention in primary care for a career, and top volume of primary care compared to other forms would deliver far more primary care compared to primary care graduates with a minority entering primary care, steady departures after entering, lowest activity in practice, fewest years in a career, and lowest volume. So far studies using measuring tools that capture the primary care delivery differences such as the Standard Primary Care year have not been accepted even though they are quite logical.
The United States tolerates a wide variety of primary care definitions. This contributes to the confusion. Claims of primary care must be replaced by a focus on actual delivery. Internal medicine training produces 6 10 graduates to get one primary care physician. Nurse practitioner and physician assistant training requires 3 to 4 graduates to get 1 primary care practitioner delivering lowest volumes of primary care per graduate. Pediatrics training programs now graduate two to get one primary care physician. Four out of five primary care training programs graduate a minority into primary care. Only family medicine graduates result in close to a 1 to 1 situation with one average graduate resulting in 0.8 to 0.85 graduates delivering top volume primary care for 35 years.
A common theme in primary care is slow progressive cumulative changes. When compared year to year, the changes are small. Over decades, the changes are significant.
The United States graduates entering and leaving primary care continue to remain the same in family medicine graduates and in physician assistant graduates. The two forms reach the same primary care numbers different ways. For the physician assistant numbers entering primary care to remain the same it has required a doubling of the annual graduates in the past decade. The numbers entering primary care remain the same but the percentage entering primary care has been cut in half.10
Pediatric primary care produced matches those departing primary care.11 There have been changes over the past 10 15 years in the proportion entering office based primary care compared to all pediatric graduates in secondary databases such as the American Medical Association Masterfile. Decades of pediatric graduates entered office based primary care at 70 80% levels before recent declines.12 The more recent studies involving residents indicate just below 50% of final year residents planning entry to primary care not long before their graduation.13
Nurse practitioner primary care has actually declined over the last decade of class years as there has been no change in annual graduates and the percentage remaining in primary care continues to decrease.14
Internal medicine training now results in less than 10% (2% of 25%) entering primary care rather than levels of nearly 60% entering office based general internal medicine for 1992 residency graduates. This steady decline is verified by comparisons with office based primary care physicians and surveys of residents and medical students.12, 15, 16 Osteopathic graduates maintained a higher proportion remaining in primary care but dipped below 30% (4.9/17) by 2004.17 In addition the internal medicine graduates that enter primary care move away from primary care to the expanding hospitalist, urgent, emergent, and locums careers along with the traditional departures over time to hospital, administrative, teaching, research, and specialty careers.9, 12
The United States has not made any moves to prevent the conversion of primary care to specialty care during primary care training, at primary care graduation, or during any subsequent year of primary care delivery. From the perspective of most physicians, a change makes little sense as over 75% of physicians, hospitals, emergency rooms, urgent care centers, medical schools, and residency programs benefit from the current designs. The design does not work well for the 65% of Americans that are found in 28,000 zip codes each with less than 75 physicians. These are locations with only 23% of physicians and half to one-third of the primary care level needed. Locations outside of concentrations receive 15% of the economics related to health care so they are left behind by the design in health services, health access, and health economics.18
The current design results in at least 80% if not 85% of the health care spending distributed to 3400 zip codes clustered together in 4% of the land area where 75% of physicians are found. The focal points for these concentrations are commonly medical schools or other largest systems. With hundreds of billions of dollars flowing annually to these current concentrations, it is easy to resist changes that would move funding elsewhere to primary care or to locations outside of current physician concentrations. Insurance companies satisfied with a situation with higher costs and a steady regular share of funding are also not likely to be interested in changes. In addition the concentrations of funding, expertise, social organization, and political organization allow all state and federal programs, even those devoted to health access, to be utilized in locations that already have top concentrations. Inside of concentrations every line of revenue is represented with the top reimbursement in each line. Outside of concentrations the facilities, physicians, billing codes, and funding sources are limited to the fewest and reimbursement is the lowest in each line of funding. Health care may be one of the largest factors dividing the United States into rich and poor in terms of distributions of income as well as health outcomes.
Non-physician workforce has followed the same pattern as physician workforce as a result of the same influences. The United States has tolerated the conversion of the nurse practitioner and physician assistant forms from their initial design to meet needed health access primary care to the non-primary care design with specialty choice and practice location in just 3400 zip codes. This conversion has worked well for non-physicians that also receive higher salary and more desirable working conditions. The design also results in financial benefits to employers (system or physician) due to the lower non-physician salary cost (compared to specialty physicians) along with the ability to generate substantially more reimbursement for an enhanced profit scenario.
The United States has not opposed the development of urgent, emergent, locums, specialty, and hospitalist care forms that have recruited continuity primary care practitioners away from primary care. None of these forms have saved overall health care costs. There is evidence that costs are increased. These are also mechanisms that concentrate physicians and non-physicians rather than distributing health access. A problem in health care is that more and more deliver no health care at all yet take an increasing share of health care spending.
Few have considered the destructive cycles that defeat primary care and health access. Since practitioners in these new forms of care benefit and their employers also benefit, even more office based primary care is lost. In turn this leads to more difficulty accessing continuity primary care with a new cycle of desperation, profiteering, and declining primary care workforce. At the root of the problem is lowest primary care support compared to other choices with departures of primary care during primary care training, at graduation, and each year after graduation. The design is actually no design with physicians and non-physicians following the pathway of highest concentrations of physicians, health resources, people, and income.
The lack of change in support for primary care in the design would be expected to result in no more primary care. This is exactly the result. United States primary care workforce has remained the same from the period of 1980 to the present period. The existing designs for training and support assure much the same primary care workforce delivery capacity until at least 2040.
Physician workforce changes take decades of effort. It takes years or decades to recognize the need for change. It may take longer if significant social or political organization is required to take on the status quo that has become dysfunctional. After the new design change is accepted takes 10 years to increase (or decrease) annual graduates to the new design level. Once this level is reached it takes 35 class years of graduates to replace the class years in the previous design. After about 40 45 years the new design can fully emerge. As will be confirmed, the primary care design remains unchanged while the specialty workforce design continues to expand.
Granted it is somewhat challenging to calculate primary care workforce. This is mainly because so many primary care graduates do not delivery primary care. About 22,0000 of 28,000 annual primary care graduates arise from flexible primary care training (IM, NP, PA) and since over 65% - 90% of these forms continue to depart primary care at graduation and after graduation, it is actually not possible to produce enough primary care under the current primary care design.19 A lower level of about 16,000 primary care annual graduates could address primary care workforce, but not using the current design based on flexible primary care training.
A natural experiment illustrates the problem. The doubling of physician assistants from 3100 annual graduates in 1998 to 6500 annual graduates in 2008 did not result in significantly more primary care as physician assistants in 1998 started at 54% in primary care and have declined to 28% starting in primary care as of 2008 graduates.10 Declines in non-physician primary care retention (NP and PA) have averaged 2 3 percentage points each passing calendar year although in the past decade this has slowed to 1 percentage point per year.
Each passing year the primary care retention rates set new record low levels. The new graduates of recent class years also continue to enter primary care at known lower levels as compared to established graduates for internal medicine, pediatric, and physician assistant forms. This indicates further declines to come.
It is difficult to estimate future primary care when primary care graduates fail to remain in primary care. Also there are differences in the primary care volume, the percentages remaining active in health care, and in career length.
Measuring Primary Care Considering Primary Care Retention
The Standard Primary Care year measuring tool is specific to the type of primary care and the class year of graduates. The estimate of future workforce estimate is the product of years in a career age 30 to 65, the percentage remaining in primary care for a career, the percentage remaining active in the United States workforce, and volume of primary care delivered relative to the top delivery of primary care in pediatric and family medicine primary care forms also expressed as a percentage. For the 2008 graduates the estimates for primary care delivery were 25 SPC years for family medicine residency graduates, 14 SPC years for pediatric graduates, and 2 4 Standard Primary Care years for internal medicine, nurse practitioner, and physician assistant graduates. Ideal primary care is 35 Standard Primary Care years. Minimal primary care is 1.2 Standard Primary Care years for non-citizen international medical graduates completing internal medicine training with lowest primary care retention (all forms), fewest years in a career (all forms), fewest remaining active in the United States workforce (all forms), and lowest volume of primary care delivered (among physicians).
The United States has not increased primary care delivery capacity in at least 30 years across the span of numerous political administrations, Congresses, and health workforce leaders. Future primary care delivery will remain at the same level under the current design despite an increase in total primary care graduates from 28,000 to 45,000. This expansion involves predominantly projected increases in internal medicine, nurse practitioner, and physician assistant graduates with family medicine and pediatric primary care remaining steady.19
The United States did increase primary care workforce capacity in the 1970s and briefly in the 1990s. The 1970s growth involved internal medicine and family medicine and the 1990s growth involved more family physician and nurse practitioner graduates and higher rates of retention in the flexible IM, NP, and PA forms. During the 1990s all primary care was needed to capture market share and teaching volume. Now market share considerations involve specialty and hospital market share and IM, NP, and PA forms have converted this direction.
Some of the 1970s gains were maintained for the long term but the 1990s changes did not remain. The doubling of annual physician graduates was able to address an increase in primary care from 1970 to 1980. During this period the government added specific primary care funding to medical education with new schools and expansions of existing schools. Also more primary care was supported by the initial expanded coverage designs of Medicare and Medicaid. For the 1980 to 2010 period the United States has failed to improve upon the 1970s gains.
The Standard Primary Care year is a new measuring tool that represents a significant advance over assumptions as it does address the major determinants of primary care delivery years in a career, primary care retention, activity in United States practice, and primary care volume differences.
Another advance over simple projections involves the concept of maximal primary care. These methods indicate that the United States will have only a workforce of physicians and non-physicians equivalent to about 200,000 to 250,000 primary care physicians for decades to come rather than the 400,000 primary care physicians needed in the workforce by 2040.
The Process of Primary Care Workforce Design
The first task of workforce design is to ignore the current situation and time period. In particular the anomalies of the 1990s class years do not provide a sound basis for the typical workforce projections. As noted previously it takes decades to move from design to implementation. Workforce design must focus on 40 50 years into the future. Typically once a nation decides on a design change (if it does make a decision) it takes about 45 years to allow the new design to fully emerge.
As a notation, the United States has rarely if ever attempted such a coordinated planning and implementation over its entire history. Of course the system designs for child development, early education, public health, and primary care all require such coordination. Not surprisingly the nation is failing in child well being, education, public health, primary care, physician distribution, 20and numerous other ways dependent upon these basics.1, 19, 21-24 For the United States to continue to exist as a viable nation, these areas must be addressed.
There is one basic assumption that can be made in estimates of primary care workforce. Using national populations of physicians and non-physicians over many decades, the direction of primary care has been steadily lower with decreasing percentages remaining in primary care over subsequent class years. While it is true that some specialty physicians and non-physicians convert to primary care, the bulk of annual change is primary care to specialty care. For example even during the 1990s time period favorable to primary care with physician assistants (all graduates and new graduates) increasing from 44 to 54% in primary care, studies of physician assistants that were known to be in primary care in 1990 steadily departed primary care across the decade.25, 26 This natural experiment exposes the folly of thinking that simple percentage increases in primary care reimbursement can influence primary care capacity significantly.
Maximal Primary Care Calculations of Age 30 to Age 65 Practitioners
When annual graduates entering the workforce are constant or declining, estimates of maximum primary care can be made. With a constant annual graduate level, the future levels are predicted well. With a declining annual graduate number, the use of the most recent data at least allows some reality to set in regarding coming changes as for example a decline from over 3000 internal medicine graduates entering primary care to less than 1000.
· Constant Annual Graduates Entering Primary Care With a constant level entering primary care, the calculation of maximal primary care is the simple product of the number entering times the average career length from 27 years (NP, FIMG) to 35 years (US origin physician, PA). For family medicine with 2500 entering since 1976 the level is 35 times 2500 or about 90,000 to 100,000.
· Stable or Declining Graduates Entering Primary Care With steady declines in annual graduates entering primary care it is also possible to generate a maximum level by the same method of annual graduates times career length. However this is a level greater or much greater than the actual level. For internal medicine a maximum estimate would be 800 entering primary care for 35 class years for 28,000 internal medicine primary care physicians age 30 - 65. This may be a shock for those used to seeing over 110,000 as an internal medicine primary care number or projections of 155,000 in internal medicine primary care by 2020. The declines in the number entering primary care have been steady for over 15 class years without much slowing and it is likely to continue to fall until there are major design changes involving those selecting internal medicine, training, and policy support.
A final method of calculating maximum workforce involves the market-based limitations of the distribution of a particular graduate. For general internal medicine and general pediatric physicians and non-physicians, about 70% are found in zip codes with 75 or more physicians with only 10% found in rural or in underserved locations. This is a limited distribution not much different than the 75% of United States physicians found in these 3400 zip codes. Also for pediatric workforce, this limited market where pediatricians choose to locate is saturated. It is also a market that remains glutted from all five forms attempting to deliver primary care and because those graduating from health professional training are predominantly born, raised, and trained in top concentrations and prefer to remain in these locations, even if this requires lower salaries or part time work to make it to full time. Family physicians take over more pediatric care as age increases past 4 years and as practice locations have fewer physicians, facilities, people, and income. Family medicine could also be considered a reason for limitations of pediatric and internal medicine workforce distribution through market effects.
The end result of a number of factors is to limit pediatric distribution to locations with higher densities of physicians, income, and people. The pediatric primary care workforce is limited by stagnant levels of United States children, care limited to age 4 and lower, and limited distribution outside of zip codes with 75 or more physicians. Increases in pediatric graduates cannot result in more primary care due to limitations in children, age range, practice support, and practice location.27-30 Indeed the pediatric percentage of graduates remaining in office based care has declined in conjunction with saturations of pediatric primary care and limitations in the distributions of pediatricians.
The primary care numbers until 2040 are actually fixed in place by the design.
· Family medicine will remain steady at 95,000 determined by a steady 2500 annual graduates entering primary care. The level could be lower if expansions of graduate medical education positions allow even more to bypass this permanent primary care form or if the percentage leaving office primary care increases more rapidly.
· Pediatric primary care will remain saturated at 55,000. Increased support for primary care for children could slightly increase pediatric primary care.
· Internal medicine primary care age 30 to 65 will continue to move toward 800 times 35 or 28,000 graduates by 2040 or sooner. The actual level still needs a reduction for the 45% of internal medicine graduates that have only 28 year careers due to later entry into the workforce.
· With nurse practitioners relatively stable at 7000 annual graduates per year for the past decade the 27 year career length results in 7000 times 27 for 189,000 but declining primary care percentages, only 60% active, and limitations in volume of primary care (60% of the volume of a primary care physician) results in only about 25,000 physicians worth of primary care. Once again without a major training design change that would result in changes in all four parameters, nurse practitioner primary care contributions are relatively fixed at lower levels.
· The physician assistant primary care retention percentage remains lower than nurse practitioners but PAs tend to have fewer in part time and inactive categories and also deliver slightly higher volume. The end result for physician assistant primary care is about 25,000 physicians worth of primary care as found in nurse practitioners and as found in internal medicine for 2040.
The calculations using maximal primary care figures result in about 228,000 primary care physicians worth of primary care. This is a level similar to 1980s values. The gains of the 1990s were negated by internal medicine departures to hospitalist careers (over 30,000 lost). The graduates 2000 to 2040 should remain steady or should decline.
Primary care delivery will also be complicated by increasing complexity and the rising costs of delivering health care. Simple measures such as technology interventions cannot replace the insufficient primary care physician workforce, the insufficient nursing workforce, the insufficient health care team support, the increasing cost of delivering care, and the sad state of primary care facilities.
New forces also complicate basic health access primary care. In states such as Alaska the costs of delivering the same primary care are increasing by at least two million additional dollars with each passing year. Locum tenens increases are one million per year and recruitment and retention costs are another one million dollar increase each year. The burden of the additional cost falls on Alaskas most vulnerable facilities delivering care to the most vulnerable populations. In addition the state will triple the level of elderly in the next twenty years.31
Alaska receives insufficient help from the national designs of medical education and graduate medical education and also loses what it attempts to produce to other states. Alaska has already begun the process towards a new design for oral health. Eventually it may also figure out that about 25 annual permanent primary care graduates with medical education paid in exchange for 15 year instate obligations could meet about 70% of the workforce needs at a cost of 8 10 million dollars. This is about the same cost as currently paid with only about 30 40% of the workforce needs met.
Family medicine represents another natural experiment to illustrate the impacts of the entire range of health policy influences at a single decision point in time. Students selecting a permanent primary care make one permanent decision on a career that lasts for 35 years. The US allopathic medical school graduates represent a defined population that has relative freedom to choose family medicine or not. Other graduates may be forced to choose a primary care training form just to enter the US workforce. The primary care health policy impact of the United States can be illustrated by the percentage of US MD students choosing the permanent family medicine choice in each medical school class year. An additional control is needed for physician origins as the probability of family medicine choice also varies by parent influences such as income and geographic location (linear starting at 5% for the most urban and highest income counties and 28% for most rural and lowest income).
Foreign born US MD physicians share the most urban origins prior to beginning medical school and have increased from less than 3% to 16% of US physicians.12 The US MD grads of lower and middle population density and lower and middle income levels have been steadily replaced by admissions of higher income and more urban origin physicians.32, 33 The national focus on minorities in some ways has distracted the focus from replacement of lower and middle status children with admission of those most exclusive. Children from families that have been in America for generations in America are being replaced by those born in other nations or relatively new to the United States. In this matter, the birth to higher education process is exposed as flawed in the United States.
The levels of family medicine choice are also modified by the type of training. The schools with top MCAT scores most selective in admission and training have lower family medicine choice for each birth origin and each policy period. With a more normal MCAT score for entering classes the choice of family medicine shifts higher. Family medicine, primary care, rural, and underserved workforce are more closely associated with lower and middle income children and first generation to college, but the United States is moving a different direction. Parent income levels of admitted students set new records yearly. The MCAT scores of allopathic matriculants have continued to rise steadily since the 1993 standardization of the MCAT. Osteopathic MCAT scores have been increasing at a rate 30% faster than allopathic scores in the past 7 years.
Exclusive policy design, exclusive birth to admission, and exclusive training all shape decreasing primary care as best indicated in family medicine. Policy represents a set point with admission and training influences moving family medicine choice above or below the set point.
Health policy influences impact medical student choice of family medicine and impact primary care retention in other forms of primary care. Family medicine changes match up to 1980s decreases and 1990s increases in physician assistant primary care retention and in the proportion of internal medicine graduates remaining in office based primary care.34 Pediatric primary care retention appeared to be stabilized by 1990s policies and only decreased with saturations in the past decade. The doubling of Medicaid from 1990 to 1995 with the emphasis on expanded eligibility may have been a key reason for stabilization of pediatric primary care during the early 1990s. Another reason likely is that pediatric specialization has not rewarded pediatric graduates to the same degree as found in those landing internal medicine fellowships.
Overall family medicine remains the same level of 3000 annual graduates although this has involved increasing supplementation by osteopathic graduates, United States origin international medical graduates (Caribbean), and a 20% fraction of foreign origin international medical graduates.35
Few understand that the top single medical school source of primary care for the United States is actually outside of the United States. Ross University in the Caribbean leads with about half of graduates entering primary care. About 25% of 471 annual graduates choose family medicine. Ross contributions are the result of US origin physicians that have not been accepted at United States schools that are also forced (to some degree) to choose primary care in order to re-enter the United States as physicians. The evidence of exclusion of primary care continues to build.
Osteopathic changes in family medicine have been predictable for some time. Osteopathic annual graduates entering family medicine have remained at about 500 to 600 for recent decades. The initial osteopathic focus of a century ago resulted in over 65% choice of general and family practice for decades. By the 1990s about 35% were choosing family medicine. The 2004 graduates indicated 22% planned to enter family medicine. Continued declines in the percentage entering family medicine predict no change in annual family medicine graduates even as osteopathic annual graduates double from 2004 to 2017. During the same period family medicine is set to decline below 15% of osteopathic graduates. Out of 1000 unfilled osteopathic graduate medical education positions, 31% were family medicine positions and 17% were internal medicine positions.
Bypassing Primary Care, Especially with Generic Expansion
It is important to understand that limited specialization opportunities also result in greater primary care. The top ranking allopathic private medical school graduates have top probability of a prestigious specialty followed by middle ranking schools with international and osteopathic graduates competing for remaining positions. If graduate medical education positions were expanded, one likely outcome is that fewer are likely to choose family medicine, womens health, psychiatry, and primary care residency positions that are less preferred at the current time.
One concern regarding an expansion of graduate medical education would be even fewer choosing (or forced to choose) family medicine each year. The 1990s peak family medicine level involved a period when medical students avoided hospital based careers fearful of managed care implementation with fewer specialists required. Expansions of graduate medical education positions would involve the opposite impact with family medicine choice declining below the current record low levels. Even with the forced family medicine choice of the 1990s policy era, 98% of the 1997 2003 family medicine residency graduates remained in family medicine as of 2005 data and there was no deterioration in this percentage across this 7 year period.36 Retention has not been the problem.
Flexible Primary Care Non-Physician Graduates
Nurse practitioners have reversed from 65% primary care to 65% specialty care from the 1980s to the current period. Physician assistants have reversed from 70% primary care when created to 70% specialty care by 2005. Fewer years in a career, more in part time or inactive mode, and lower volume all complicate primary care delivery in non-physicians. Also the non-physicians that do not remain in family practice for their entire careers also contribute less to primary care where needed. Family practice forms are 2 7 times more likely to be found in rural or underserved locations and family practice physician assistants are 30 times more likely than other PAs to be found in federally qualified rural health clinics. For 2010 to 2040 the United States will need efficient, effective, high volume primary care that delivers primary care predominantly in the broad generalist family practice mode. It is getting no increase in primary care and no increase in the family practice broad generalist mode that could help resolve health care deficits for the rural, underserved, and elderly populations left behind.
Non-physician training has moved away from needed design. Non-physician training has changed regarding the students attracted to training, the curricula, the training locations, the length of training, the preceptors, and the focus of training. Most of all, non-physicians have changed in one very important way. The non-physicians have gained wide acceptance across the wide range of specialty choices and practice locations. This is a positive factor as is increasing acceptance of minority physicians and osteopathic physicians, but the end result has been a loss of most needed health access. The barriers that have commonly forced primary care and underserved locations have been removed.
Much more than acceptance drives movement away from primary care. In addition the non-physician specialists enjoy the benefits of 10% or higher salaries.37, 38 The benefits of specialization are more team members and better facilities resulting from better support. Non-physicians as specialists also generate more revenue for employers, generate additional specialty visits as triage (first contact) specialists, capture market share, and funnel higher paid billing codes to specialty physicians. Each of these changes incent non-physicians steadily away from primary care. Studies regarding the addition of a nurse practitioner or a physician assistant to a cardiology practice results in $300,000 in additional revenue for the practice and an increase in revenue for the physician cardiologists in the practice as compared to the addition of a physician cardiologist that adds $700,000 in revenue but results in decreased revenue per physician in the practice. Non-physicians help maximize efficient practice and have lower costs of employment. The greatest gains and the greatest employment rates are found in the largest practices with the most specialty physicians. 39 Subspecialization of non-physicians clearly changes patterns of distribution from the most needed locations to locations with top concentrations of physicians. Studies indicate rapid increases in cardiology careers in NP and in PA studies with increases of 1 percentage point a year not uncommon and levels of 8% of NPs and 6% of PAs likely already or in the next few years. Nurse practitioners tend to move to internal medicine specialties where four major subspecialties have twice the percentage of nurse practitioners compared to physician percentages already. About 35% of physician assistants are evenly divided between surgical subspecialties, orthopedics, and emergency care. These are the careers that pay the most in income and the growth rate is highest.37, 38
Non-physician primary care numbers have stagnated while specialty numbers have continued to rise at a higher growth rate than the US population. The non-physicians changes are just like physician specialty workforce changes and for the same reasons.
What Does the United States Need for Primary Care Workforce in 2040
The problem remains grossly insufficient support for primary care in so many dimensions. It is hard to get medical students to trust a permanent primary care choice when all evidence points to poor support with increasing costs and challenges. Also those entering primary care in all of the various forms fail to trust primary care as indicated by their departure after entry.
One solution is permanent primary care. The permanent primary care graduates required for the United States workforce are easy to calculate.
The United States population will double from 1980 to 2040. In addition the old, older, and oldest populations double in percentage. The change in age across age 45 to age 75 is a change requiring 6 times more primary care or an increase from 45 primary care physicians per 100,000 to 270 primary care physicians per 100,000.3 In addition the extremes of age are associated with loss of mobility and transportation ability, placing increasing emphasis on local zip code care or at least adjacent zip code care. Primary care is not the only health care out of position as stroke and heart attack centers are also malpositioned.2 Overall the aging changes will require an increase from 95 primary care physicians to about 110 for the United States population of 400,000,000. This translates to a need for 15% more primary care to be delivered. As might be expected it is not possible to deliver the primary care of 2040 with an aging population and a 1980 primary care design. The United States needs about 420,000 to 440,000 primary care physicians for sufficient primary care delivery in 2040.
Working backwards using the maximum primary care calculations, the requirement would be 14,000 annual graduates with 90% remaining in primary care and 90% remaining active in health care. The requirements would be 16,000 annual graduates with 85% remaining active and 85% remaining in primary care for 35 year careers. If the career length declines to 31 years, this requires 18,000 annual graduates.
The difficulty of reaching the required annual graduates is illustrated by the current permanent primary care level of 2500 in family medicine and 1500 in pediatrics. The United States has more than enough primary care graduates with 28,000 but the permanent primary care level is quite small.
Engineering the New Specialty Design
The 1990s reform interventions did not come with the additional funding as in the 1970s. The 1990s efforts forced medicine and medical education to support primary care at the cost of specialty care. The specialty forces rebounded with a response not long after the reforms began.
Even more organization state to state has led to hundreds of various publications regarding workforce needs. The primary care component has remained vague even for the General Accounting Office.40 Meanwhile new state and new specialty workforce reports continue to emerge.
Despite these various reports and campaigns it is difficult to determine just how much specialty workforce is needed. Studies of regional variation clearly indicate too much in many areas. Actually the major determinant of the specialist workforce is the thirst of the American people for specialty care at least for those that can access care at all. Another determinant is a desire not to die, but more hospital and specialty and cancer and heart disease and long term care and drugs and technologies cannot satisfy this desire. But the desire does result in more hospital and specialty care.
Frankly the United States can choose a lower, medium, or higher level of specialty support with the major issue being just how much the nation is willing to pay for overall health care. If primary care was funded separately, the specialty workforce decision would be a separate issue, but this is not the case. Primary care funding that is included in the same pool ends up compromised. Primary care is just not able to compete head to head with specialty care. People die of diseases, send money to disease fighting campaigns, and see those impacted by disease. They fail to see failure in health access and the countless millions that suffer needlessly in financial, morbidity, and mortality each year. When specialty care is increased, primary care is compromised. The kind of process needed to re-orient the nation to funding by priority was illustrated by Oregon, but even Oregon failed to maintain a superior effort to distribute care more equitably.
Since the design for specialty workforce has already been changed, all that remains is to graduate enough class years to achieve the new design. Since the percentage entering specialty workforce is increasing, the specialty workforce design is still changing. A design fixed at the 2005, the 2010, or the 2015 levels involve ever more specialists as a percentage of total graduates and even more as the result of physician expansion.
In the haste to graduate more specialists, few apparently understand that it is not possible to radically change a 1 million physician workforce in a short period of time. Does the United States really need more specialty care than can be provided by specialty care increasing faster than the United States population for decades with an increase to even greater levels due to higher percentages with an increase to even greater levels due to expansions of physicians with an increase to even greater levels due to conversions of non-physicians to specialty care.
Time and again in US workforce the lack of a consistent design has resulted in the consequences of too much and too little from time period to time period and from geographic location to geographic location. Already the lack of patience will result in a high probability of too much specialty workforce long before the new design is fully implemented.
In addition the new overall United States workforce design for specialty care has absorbed every bit of the physician and non-physician expansion since primary care has remained stagnant. Also specialty workforce has increased beyond the design due to conversions of existing primary care.
Graphic Past and Present Physician and Non-Physician Numbers with Future Estimated Primary Care and Non-Primary Care Graduates From Colwill and COGME and the AMA Masterfile, projections by the author
Physicians and non-primary care physicians continue to increase at a rate of growth faster than the United States population. Primary care physicians and non-physicians remain stagnant and departures of internal medicine from primary care result in declines in primary care delivery. This is unfortunate as the nation needs adult and elderly primary care more than ever.
Family physicians remain most connected to ambulatory care of the elderly with 63% seeing a family physician in 2004. Family physicians are also the only primary care form with equitable distribution to rural, poor, near poor, and other populations in most need of health access.41 Family physicians continue to provide multiple times greater health care delivery in Community Health Centers compared to other forms.42 Family physicians are the only physicians that maintain a relatively constant rate of 30 per 100,000 across the United States serving the 65% of the population that has marginal or underserved primary care levels. Choice of family medicine doubles underserved location and triples rural location.1 non physician family practice forms have comparable distribution patterns but non-physicians are departing family practice modes resulting in declines in primary care, rural, and underserved workforce.
Physician Workforce Numbers Indicated by HRSA Compared to Major Journals
What HRSA Projects
· Total Physician Increase
· NonPrimary Care Increase
· Primary Care Increase
· Internal Med PC Increase
· Family Med PC Increase
· Pediatric PC Increase
What the Major Journals Indicate along with Maximal Primary Care Calculations
· Total Physician Increase
· NonPrimary Care Marked Increase
· Primary Care Decrease
· Internal Med PC Marked Decrease
· Family Medicine PC No Change
· Pediatric PC No Change
The major difference between the HRSA estimates and estimates based on major journal articles is internal medicine. The problem of projections is also documented by the General Accounting Office with great uncertainty using the usual methods.40 There really is not uncertainty in internal medicine. It is not possible for the United States to have an increase in internal medicine primary care from 110,000 to 155,000 from 2000 to 2020 as projected by HRSA.3 Documentation includes
· Surveys of internal medicine residents and future residents indicating 20% and then 10% planning to remain in primary care; 15, 16
· Departures of 30,000 to the newly created hospitalist workforce as indicated by 75% of hospitalists from internal medicine; 4
· About 45% of internal medicine residency graduates that are foreign origin international medical graduates that deliver the least primary care of any form due to shortest career length, fewest remaining active in the United States workforce, and lowest volume of primary care for physicians; and
· With rapid aging of the remaining primary care internal medicine physicians and studies indicating older age physician activity overestimated by 10% 43
Only 800 internal medicine residents can be counted upon to enter primary care each year and the steady decline in entry indicates that this will be a maximum entry for some time. The net effect is two or more thousand more internal medicine physicians departing primary care practice compared to those entering in each calendar year.
Studies continue to document fewer and fewer internal medicine entering primary care and steady departures from primary care in articles in major journals, in secondary databases, in annual surveys, and using a variety of measuring tools. 5, 8, 9, 15-17, 44-47
WARNING TO PHYSICIANS
A serious problem should be obvious for physicians - too much specialty care plus non-physicians as a increasing low cost source of specialty care source and out of control health care cost issues. Add on as concerns
· Out of control health care costs compromising national, state, corporate, and local budgets with even school districts forced to cut teachers and education support to meet health care costs.
· Promotions of internal medicine as a primary care solution even though internal medicine is 90% not a primary care solution.
· Continued support of a design that results in 80 85% of health funding going to only 3400 zip codes clustered together that already have 75% of physicians
· Continued support of a design that results in only 23% of physicians found in zip codes with 65% of Americans and 70% of the elderly. 20
· Continued support of a design that results in six states (actually only a few dozen zip codes in six states) receiving half of the economic impact arising from medical education and graduate medical education with 25 states receiving only 10% and this same study is promoted as an economic solution for the United States.48
· Physician origins that are more and more exclusive in higher income, most urban, and parent origins and less and less like the 65% of the American people left behind (likely impact on health care cost, quality, and access).
· Declining levels of service orientation, choice of direct patient contact careers, empathy, and awareness of the needs of the underserved
· Failure to admit medical students based on known areas related to physician quality such as communication skills.49
· At least forty medical schools (and increasing) that contribute less than 10% of graduates to primary care. These medical schools receive the most federal funding for graduate medical education, research, corporate, and clinical funding. The same medical schools have levels of 80% of graduates that remain in 3400 zip codes in 4% of the land area and contribute the least to family medicine (less than 3%), rural workforce (less than 4%), and underserved workforce (less than 4%) or less than half of the national averages for each. The same medical schools provide the least opportunity for lower and middle income children through admission and also admit the fewest born in the United States. They do have the highest prestige rankings, Medical College Admission Test scores, exclusive training, research graduates, and subspecialty graduates.
· Foreign origin international medical graduates that deliver half of the workforce of a US origin graduate for the same cost of training, that take entering US physician positions instead of children born in the United States, and that have greater rates of physician discipline likely due to the difficulties adjusting to the United States. The brain drain departures encouraged by the United States design is a major concern regarding weakening these nations.
1. Bowman RC. They really do go. Rural Remote Health. Jul-Sep 2008;8(3):1035.
2. Perrotta BL, Perrotta AL. Access to state-of-the-art healthcare: a missing dynamic in consumer selection of a retirement community. J Am Osteopath Assoc. Jun 2008;108(6):297-305.
3. Health Resources and Services Administration. The Physician Workforce: Projections and Research into Current Issues Affecting Supply and Demand; 2008.
4. McMahon LF, Jr. The hospitalist movement--time to move on. N Engl J Med. Dec 20 2007;357(25):2627-2629.
5. Rauner T. Recent Primary Care Changes in Nebraska: Nebraska Office of Rural Health and the Health Professions Tracking Center; 2007.
6. Bodenheimer T. Primary care--will it survive? N Engl J Med. Aug 31 2006;355(9):861-864.
7. Bodenheimer T, Berenson RA, Rudolf P. The primary care-specialty income gap: why it matters. Ann Intern Med. Feb 20 2007;146(4):301-306.
8. Bowman RC. Measuring Primary Care: The Standard Primary Care Year. Rural Remote Health. Jul-Sep 2008;8(3).
9. Sox HC. Leaving (internal) medicine. Ann Intern Med. Jan 3 2006;144(1):57-58.
10. American Academy of Physician Assistants. Data and Statistics. http://www.aapa.org/research/index.html, 2009.
11. Freed GL, Stockman JA. Oversimplifying primary care supply and shortages. JAMA. May 13 2009;301(18):1920-1922.
12. Medical Marketing Service. AMA Physician Masterfile. (Online) Available: www.mmslists.com (Accessed 1 July, 2005) 2005.
13. Freed GL, Dunham KM, Jones MD, Jr., McGuinness GA, Althouse L. General pediatrics resident perspectives on training decisions and career choice. Pediatrics. Jan 2009;123 Suppl 1:S26-30.
14. Goolsby M. 2004 National NP Sample Survey Comparisons Over 15-Year Period. http://www.aanp.org/NR/rdonlyres/ewz24bs6jt72aeldxgvk3woyo4dhasuc5hvwpt65bs2iyej2edd3723ri3ggbwiptvoym2x7o37rwridsnb2tf3gfxh/2004NatlNPSampleSurveyWeb.pdf. Accessed February 22, 2007.
15. Garibaldi RA, Popkave C, Bylsma W. Career plans for trainees in internal medicine residency programs. Academic Medicine. May 2005;80(5):507-512.
16. Hauer KE, Durning SJ, Kernan WN, et al. Factors associated with medical students' career choices regarding internal medicine. JAMA. Sep 10 2008;300(10):1154-1164.
17. Teitelbaum HS. Osteopathic Medical Education in the United States: Improving the Future of Medicine. A report jointly sponsored by the American Association of Colleges of Osteopathic Medicine and the American Osteopathic Association. . Washington, D.C. 2005.
18. Bowman RC. Physician Distribution By Concentration. Primary Care Research Methods and Statistics Conference. San Antonio, Texas; 2007.
19. Bowman RC. Basic Health Access Web Site. www.basichealthaccess.org.
20. Bowman RC. The Physician Distribution By Concentration Coding System. 2008. http://www.ruralmedicaleducation.org/basichealthaccess/pdccoding.htm. Published Last Modified Date|. Accessed Dated Accessed|.
21. UNICEF. Child Poverty in Perspective: An overview of child well-being in rich countries. Innocenti Report Card 7 ed:
UNICEF Innocenti Research Centre, Florence.; 2007.
22. Kirsch I, Braun H, Yamamoto K, Sum A. America's Perfect Storm: Three Forces Changing Our Nation's Future. http://www.ets.org/Media/Education_Topics/pdf/AmericasPerfectStorm.pdf.
23. Murray CJ, Kulkarni S, Ezzati M. Eight Americas: new perspectives on U.S. health disparities. Am J Prev Med. Dec 2005;29(5 Suppl 1):4-10.
24. Hart B, and Risley, T. . Meaningful Differences in the Everyday Experience of Young Children. Baltimore: Paul H. Brookes; 1995.
25. Larson E, Hart LG. Historical Trends in Physician Assistant Education and their Contribution to Primary Health Care for Rural and Underserved Populations in the U.S. . http://www.ruralhealthresearch.org/projects/100002096/.
26. Larson E, Hart LG. Geographic and Demographic Dimensions of the Adoption of a Health Workforce Innovation: Physician Assistants in the United States, 1967-2000 Working Paper #105 http://depts.washington.edu/uwrhrc/uploads/CHWSWP105.pdf. Accessed October 2007.
27. Shipman SA, Lurie JD, Goodman DC. The general pediatrician: projecting future workforce supply and requirements. Pediatrics. Mar 2004;113(3 Pt 1):435-442.
28. Committee on Pediatric Workforce. Pediatrician workforce statement. Pediatrics. Jul 2005;116(1):263-269.
29. Randolph GD, Pathman DE. Trends in the rural-urban distribution of general pediatricians. Pediatrics. Feb 2001;107(2):E18.
30. Althouse LA, Stockman JA, 3rd. Pediatric workforce: a look at general pediatrics data from the American Board of Pediatrics. J Pediatr. Feb 2006;148(2):166-169.
31. Alaska Primary Care Programs. 2009.
32. Association of American Medical Colleges. Minority Students in Medical Education: Facts and Figures XI Available at https://services.aamc.org/Publications/showfile.cfm?file=version12.pdf&prd_id=89&prvid=87 Accessed April, 2003. Washington DC 1998.
33. Association of American Medical Colleges. Minority Students in Medical Education: Facts and Figures XIII Available at https://services.aamc.org/Publications/showfile.cfm?file=version53.pdf&prd_id=133&prv_id=154&pdf_id=53, Accessed July 2006. Washington DC 2005.
34. Bowman RC. Five Periods of Health Policy and Physician Career Choice. http://www.ruralmedicaleducation.org/five_periods_of_health_policy.htm.
35. McGaha AL, Schmittling GT, DeVilbiss AD, Pugno PA. Entry of US medical school graduates into family medicine residencies: 2008-2009 and 3-year summary. Fam Med. Sep 2009;41(8):555-566.
36. Bowman RC. Retention Within the Specialty of Family Practice. http://www.unmc.edu/Community/ruralmeded/retention_within_fp.htm. Accessed October 26, 2006.
37. Physician Assistant Income: Changes in Inflation-adjusted Total Annual Income from Primary Employer 2003-2004 and 2004-2005. AAPA; 2005. http://www.aapa.org/research/index.html. Updated Last Updated Date. Accessed November 2006.
38. Health Resources and Services Administration. The Registered Nurse Population: Findings from the 2004 National Sample Survey of Registered Nurses; 2004.
39. The Lewin Group PFHaEB. Value of Mid-Level Practitioners in the Subspecialty Practice: The Case of Cardiology: AAMC; 2009.
40. General Accounting Office. Primary Care Professionals Recent Supply Trends, Projections, and Valuation of Services; 2008.
41. Ferrer RL. Pursuing equity: contact with primary care and specialist clinicians by demographics, insurance, and health status. Ann Fam Med. Nov-Dec 2007;5(6):492-502.
42. Rosenblatt RA, Andrilla CH, Curtin T, Hart LG. Shortages of medical personnel at community health centers: implications for planned expansion. Jama. Mar 1 2006;295(9):1042-1049.
43. Staiger DO, Auerbach DI, Buerhaus PI. Comparison of physician workforce estimates and supply projections. JAMA. Oct 21 2009;302(15):1674-1680.
44. Sox H. Career Changes in Medicine: Part II. Ann Intern Med. Nov 21 2007;145(10):782-783.
45. Bowman RC. The Standard Primary Care Year Web Site. http://www.ruralmedicaleducation.org/basichealthaccess/The_Standard_Primary_Care_Year.htm.
46. National Resident Matching Program. National Resident Matching Program Results and Data 19752009. . 2009.
47. International Medical Graduate Section of the American Medical Association. Report on International Medical Graduates. Chicago 2007.
48. Association of American Medical Colleges. The Economic Impact of AAMC-Member Medical Schools and Teaching Hospitals. Washington DC 2008.
49. Tamblyn R, Abrahamowicz M, Dauphinee D, et al. Physician scores on a national clinical skills examination as predictors of complaints to medical regulatory authorities. Jama. Sep 5 2007;298(9):993-1001.