Shaping a Nation: Physicians Who Serve

Robert C. Bowman M.D., John Brandon M.D., Matthew L. Hunsaker M.D., Katherine Margo M.D. 

Handout for STFM Predoctoral Meetings, Memphis, January 2007

One hundred years ago medical education shaped the nation in new directions. New academic standards forced students, parents, education, high schools, colleges, advisors, and medical school leaders to make important and needed changes. These changes have resulted in the present focus on academics, centralization, technology, subspecialization, and science-orientation. Training locations moved to universities, urban areas, and major medical centers.

The blueprint clearly laid the foundation for medical schools, admissions, and current health care, but erosions in the cornerstone threaten the entire structure. Medicine is more than academics. Medicine without a top priority on people is unfulfilling for physicians at best and dangerous for people at worst. Diseases and disabilities are increasingly recognized as related to socioeconomics and origins. Those in charge of our society are those who have the greatest access to education, top colleges, and professional training. The increasing focus on academics has come at some cost of development in other critical areas. Parents and societies who develop their children academically while neglecting people development face grave dangers. Without some contact with a broad range of people, without empathy for all peoples, without awareness of the broad range of needs of all peoples, and without sufficient numbers of service oriented professionals to form essential infrastructure for all of the people, a society faces greater potential for poor leadership, inefficiency, neglect, increasing hopelessness, and disruption.

In an age of increasing diversity, divisions, and instant information it is impossible to hide those who have every opportunity for life, liberty, and pursuit of happiness from those who border on hopelessness every day of their lives. EMultiple growing problem areas involving education and child development and opportunity A nation that allows market forces to concentrate wealth, education, and opportunity to excess cannot help but concentrate poverty and hopelessness, extending inequitably to its own cities and states and moving beyond borders to take the most important human resources of other nations. Weaker nations and states losing these most educated and professional people will only grow weaker and unstable, especially in health and education infrastructures. A strong state surrounded by weakening neighbors or a strong nation surrounded by increasing chaos in other nations is no longer possible. Efforts to build walls or increase security funding eventually fail. Nobel prize winners in peace (Nash, others) and economics (microcredit, Muhammad Yunus) have realized that the leaders and professionals who advise the leaders must consider not just the individual benefit of a person, business, or government, but also the impact on groups, entire markets, and the world community. .

Other than a few brief periods of health policy change, health care has lost sight of broader societal goals. The evidence is overwhelming in all types of health statistics and the related education and societal outcomes. For physicians the most dramatic changes have involved the medical school classes graduating one hundred years ago (Flexner’s Impact on American Medicine), those graduating from 1965 – 1978 (Medicare, Medicaid, federal medical education support, expansion, FP, new schools), and the classes of 1990 – 1997 (doubling of Medicaid, increased primary care reimbursement, limits for hospital careers).Five Periods of Health Policy and Physician Career Choice ¹

Changes have improved health care costs, quality, and access by improving individual physicians, their career choices, the way they relate to others, and how they interact with society. Improvements in the physicians themselves are a most important consideration. Improvements in the physicians, their career choices, and their distribution have much to do with family medicine. The nation divides into:

• Highest cost, lowest quality, poorest distribution – major medical center location
• Lowest cost, highest quality, best distribution – any other location

It is not a surprise that family medicine is related to the above since family physicians are least likely to be found in major medical centers and dominate physician concentrations in all other locations. Those most associated with major medical centers before, during, and after training contribute to higher costs directly and also indirectly in the resources taken away from lower and middle income populations. The children of highest income parents have the most advanced child development, ² have the best public or private education, http://www2.edtrust.org/NR/rdonlyres/30B3C1B3-3DA6-4809-AFB9-2DAACF11CF88/0/funding2004.pdf  ³ and dominate the top 146 colleges with 74% from the top income quartile. http://www.tcf.org/Publications/Education/leftbehindrc.pdf ⁴ Top income origin medical students represent only 30% of the population and claim over 70% of medical school positions. ⁵ Those at the top levels are youngest at admission, most urban in origin, born in cities or counties with medical schools, have professional parents, and are closely associated by family, education, and geographic proximity with major medical centers for their first 30 years.

"One of the greatest problems of history is that the concepts of love and power are usually contrasted as polar opposites. Love is identified with a resignation of power and power with a denial of love....What is needed is a realization that power without love is reckless and abusive and that love without power is sentimental and anemic. Power at its best is love implementing the demands of justice. Justice at its best is love correcting everything that stands against love." ~ Rev. Dr. Martin Luther King, Jr.

Contrasting Admissions and Distributions of Physicians

There is one way to admit future physicians who concentrate in major medical centers at the highest levels: focus on the top status indicators involving parents, origins, scores, colleges, academics, and medical school locations.

There are a number of ways to admit future family physicians, rural physicians, underserved physicians, and primary care physicians that will serve a much broader variety of people, improve access to health, and help shape costs and quality.

• Admissions of those who are most different from the medical students admitted in the highest probabilities in age or a number of socioeconomic, geographic, cultural, and social dimensions, those that share origins with 70% of middle and lower income Americans most in need of care
• Admissions of those who have overcome the most obstacles and persisted
• Locating medical education in places unknown to those most likely to gain admission or where the most exclusive types would never live
• Admissions of those who have the strongest associations to family medicine by life experiences, desire, and plans
• Admissions of those with the most connections to direct patient and people aspects of medicine in skills, relationships, interests, and service orientation

Different peoples, different locations, different life experiences, the different career of family medicine, different distribution, and different relationships involving people are important.

Measures of Service Orientation

• Service Oriented Students - older, female, married, and those who serve with a business-like quality (O’Connor)  www.sba.muohio.edu/management/mwAcademy/2000/38c.pdf ⁶
• Service Orientation – primary care choice, broad prep, past careers of service, older age graduate, leader of extracurricular activities, service statement, past serving career (Madison) ⁷  see Service Orientation
• Urban Underserved Interest – 40% interest level for Black, Mexican American, Native American medical students ⁸ - this level of services is found in those choosing FP, a choice that doubles distribution (Bowman, Ethnicity and Distribution in 1997-2003 FP Grads) Ethnicity Gender and Rural Practice Choice
• Rural Underserved Interest - 60% interest level for rural interested senior medical students, a group that was older, 68% choice of family medicine, married, even mix of rural and urban origin, rural spouse prominent Rural Interested Senior Medical Students 1995 ⁹
• Accelerated Family Medicine Grads – 132 in 12 medical schools/residency programs, 50% rural choice (5 times normal, double the FP usual), 18% underserved choice (3 times underserved location, 60% greater than the most comparable white FPs), 56% were older grads (twice the FP level of 28%, three times level of elite medical schools at 18%) Accelerated Family Medicine Training Programs ¹⁰
• Older grads have greater choice FP, rural, underserved; are more diverse in origins by geography, income, culture, language. Older age at admission is a direct and linear  reflection of barriers that delay admission and that translate to greater distribution (Bowman, Older Graduates)  Older grads tie all of these service-oriented types together.Divisions in Physician Career and Location Choice Related to Age at Graduation⁵
• Serving patients directly in primary care, mental health, women’s health rather than as a secondary or tertiary provider; not passing the buck; continuing to expand skills and expertise

Those without service orientation before medical school are the most resistant to service oriented training. (O’Connor) ⁶ It is likely that areas such as professionalism, awareness, empathy, and similar personal qualities will also be resistant to training or will cost so much to enforce that the value is negated.

Physician distribution is similarly related to origins. Those found in distributional locations are those with origins most similar. Choice of family medicine allows maximum expression of distributional tendencies. Any other career choice forces a location with greater concentrations of patients, technology, patients, and staff. This gives the physician a more and more narrow range of locations.

Origins, Career Choice, and Maximal Distribution

It is not possible to separate origins (any within the 70% in middle and lower income), lower standardized test scores, the broadest range of diversity in America, lack of opportunity, older age graduation, family medicine career choice, direct service to people, and lower probability of admissions, and higher probability of distribution. Those most socially, geographically, and numerically (age, scores) different from the typical medical student admission are the least likely to be admitted, they are the most likely to choose primary care and family medicine, and they are the most likely to distribute where physicians are most needed. There are good indications that they are also the most:

• Service oriented – increased in those most different (see Service Orientation)
• Empathetic – potentially lowest in highest status or highest scoring student types although not fully studied, levels decrease with increased contact with major medical centers and training (Hojat) ^{11, 12}
• Satisfied and Satisfying – studies demonstrate that certain specialties have the lowest satisfaction levels (Leigh). ¹³ These are physicians defined by the youngest age at medical school admission, the highest board scores, and preference for the least direct contact with patients. Research categories fail to consider this in direct studies. Also consistently across patient quality studies the physicians most like their patients are rated highest in quality. It is not a stretch to add class and origins to the race, ethnicity, and gender findings.
• Awareness of the broadest range of American peoples – barriers to care and responsibilities for serving the poor are least understood by those raised in the most exclusive environments (AAMC Minorities in Medicine) Awareness and Future Physician Leaders

One can also argue that those who have overcome obstacles and persevered and have gained admission, are also people of character. Those of privileged origins are not devoid of character, but they are likely to have fewer challenges. They also have parents more than ready to move in and modify the experience. Only elite origin students with unfortunate life events or those who have lives structured differently by parents, locations, or personal preference are likely to be different.

Infrastructures, Opportunity, Professionals, and Distribution

There are other important considerations for societies. Those most different, diverse, and distributional depend upon intact infrastructures from birth to child development to secure neighborhoods to adequate housing to preschool to early education to college access to professional school admissions. This is not small group. This is 70% of the middle and lower income group in America. States and nations that limit infrastructure will concentrate professionals, income, and influence in the hands of a few as in the top 30% for top colleges and professional schools in America. States and nations that build strong infrastructures extend opportunity to a broader segment. The most complex societies require a population that are not only highly educated, but include a broader segment of highly educated. Societal efficiency and effectiveness depend on leaving the fewest behind and getting contributions from all segments of society. Where achievement scores are lowest and where the fewest complete a meaningful high school education, the challenges will be the greatest.

States that distribute resources also graduate more family physicians. States, counties, and areas with increased high school grad rates graduate more family physicians. When no segment of society is left out of education, no segment is likely to be left out of distribution of professionals. Areas with the greatest divisions related to income, education, and professions graduate fewer professionals, a greater percentage of elite professionals, and lower percentages of family physicians and all serving professionals. Income quintile ratios are considered measures of inequality that correlate highly with Gini Indexes. The most inequality in the nation is a 30 to 1 ($210,000 to $7,000) income quintile ratio in Washington DC followed by 20 to 1 in New York (worse in NYC), and 14-16 to 1 in the Southern States. The states with the broadest distributions of 7 to 1 in Utah and 8 or 9 to 1 in the upper Midwest distribute physicians the best; have the fewest physicians in major medical centers; graduate the most family physicians; supply teachers and nurses to most other states; have the lowest health care, prison, social, and education costs; and have the top outcomes in education and health. States that concentrate resources graduate the fewest family physicians and have the fewest family physicians in practice. They also concentrate the most physicians within major medical centers. (Bowman, Bright Future Rankings) Washington DC physicians are 94% major medical center physicians. This is far above the 78% national average.

Family physicians are only 7 - 9% of the physicians in major medical center locations. This is a location that contains 50% of the nations family physicians and 80% of all other types of physicians. Family physicians are 12% in rural major medical centers, 16% in urban served and underserved areas, and 30 – 50% of physicians in rural served and rural underserved locations. Major Medical Centers ¹⁴

Demand drives workforce. Prior to the establishment of a career that preferred locations outside of major medical centers and a permanent primary care career choice, the nation could not understand the current workforce equilibrium. Without health policy that established secure financing for health care and health facilities in rural and underserved areas, physicians could not consider distribution. Adequate funding in Medicare, Medicaid, and primary care along with the creation of family medicine changed the face of health care in the nation.

·         Family medicine choice enables any physician of any origin to distribute. Loss of family medicine means lower distribution for US MD Grads. Decreases in family medicine choice are related to enough support relative to the costs of delivering care. Loss of family medicine is a most serious problem because it is a dependable source of primary care with preferred distribution outside of major medical centers.

·         Rural physicians have urban origins and this means that many would prefer urban locations. Nearly 70% of rural physicians have urban origins because 90% of physicians have urban origins. The 10% of US MD Grads is small, but becomes 30% of rural physicians. Health policy and choice of family medicine are the major factors that can magnify the distribution of both groups. 

·         Health policy that increases funding for primary care and for rural and underserved locations allows maximal choice of family medicine (increasing FP choice for 1965 – 1978 and from 1990 to 1997), maximal primary care capacity (retention of primary care, FP that stays in primary care), and maximal distributional location (maximal ever in 1995 – 1997 US MD Grads).Five Periods of Health Policy and Physician Career Choice¹

·         Major medical centers with 78% of physicians and 50% of family physicians are the primary determinant of physician workforce. Health policy can shift major medical center opinions, but the centers and the urban served locations still self determine the composition of local physician and non-physician workforce and thus access to primary care. They also make the call regarding training composition by supporting different types of physicians, physician assistant programs, and nurse practitioner programs.

·         Black Box Financing - Few areas of the nation understand the concept of maximizing investments as successfully as major medical centers. For any concern from indigent care to the most elite subspecialty, major medical centers can take the appropriate steps. They can work through state and federal regulations to manage a Community Health Center, obtain cost-based capitations, maximize reimbursement (from prospective payment to the current scheme), establish a rural health clinic just across shortage lines, receive Medicare bonus pay for visiting subspecialists, recruit their own National Health Service Corps positions, or a fill a difficult subspecialty faculty position through J-1Visa recruitment. Even if they receive funding for one area such as indigent care, research, or training they can shift funding (indirect, barter with departments or states) to the areas that they see as most profitable. The excesses of a few major medical centers have also forced major regulations and inefficiencies on all of health care with greater impact on those smaller with fewer lines of revenue. Cost shifting is rampant and beyond the comprehension of even major medical centers.

·         Current Health Policy Impacts on Major Medical Centers - Under current health policy, primary care and mental health and care of the underserved are loss leaders with the lowest rates of reimbursement. This is a major reason why major medical centers have determined that only 7 – 9% of physicians in major medical centers are family physicians. If MMCs do not hire family physicians, the demand is decreased as is family medicine career choice. MGMA studies also demonstrate no advantage for physician assistants and nurse practitioners compared to family physicians. However NPs and PAs have significant advantages when working in major medical centers and not in primary care. Again under current health policy NPs and PAs are hired by MMCs away from primary care to do emergency care, hospitalist care, and subspecialty care where they generate more revenues and save major costs compared to more expensive physicians. The result is that primary care workforce levels and health access indicators plummet, especially for those with the least health policy support. PAs are leaving pediatric practices and family physicians to improve salary levels by 10% or more. In the past ten years PAs working with FPs have decreased from 40% to 28.5% of all PAs and rural PAs have declined from 25% to 16% (AAPA data). ¹⁵ PAs, NPs, and FPs, especially females, are in great demand in major medical centers and urban served locations. Since these are the areas where they are trained and where they have lived for several years and where their spouses also work, these are areas that have preferential recruitment. Underserved areas and rural areas face great difficulties, especially when the major medical center and urban served areas are not fully supplied. When there is increased saturation of major medical center and urban served “receptors,” the various primary care types are more likely to be attracted and held in place by urban underserved and rural “receptors.” Rural and underserved origin types and family physicians also have greater affinity and bonding capacity.

The question remains, how much of physician distribution is birth origins and how much involves specialty choice?

Physician workforce outcomes such as rural location or underserved location can be examined by logistic regression. The database was the birth origins Masterfile (MMS, Bowman). Only graduates of United States Allopathic Medical Schools (US MD Grads) were used. This was further restricted to those that had a birth origin listed in a city and state or a foreign nation and those with a zip code for practice that was not military or outside of the 50 states. The restrictions do not limit the results since 93% of the US MD Grads are still included. This is 189,742 US MD Grads for the 1987 – 1999 classes.

The outcomes were practice zip codes as listed in the OfficeMax version of the 2005 AMA Masterfile. Family medicine career choice, type of medical school by MCAT score average, public medical school, physician birth in a county or city with a medical school, birth in a rural location (RUCA), age over 29 years at medical school graduation, age younger than 26 years at medical school graduation, birth in an area with over 1 million people, birth outside of the United States, birth in a county in the bottom income quartile in 1969 per capita income. The analysis involved logistic regression using SPSS 14.0. A separate logistic regression equation was determined for each location holding the factors constant.

Previous work categorized the nation’s physicians into 8 basic locations. The International and Military locations do not impact the 50 states. The other six locations include Urban Major Medical Centers (73.5%), Rural Major Medical Centers (3.5%), Urban Served (12.3%), Rural Served (5.2%), Urban Underserved (3.0%), and Rural Underserved (2.3%). The United States location distributions are shown in parentheses. Both underserved categories divide equally into zip codes with a federal designation (NHSC, CHC, whole county shortage) and those with high poverty (20% or more in poverty). This suggests that some areas and populations are organized and gain designations and resources while other areas, equally needy, do not.

For the first table the two Major Medical Center locations were grouped together, the two Underserved locations were grouped, and the three rural locations were grouped into a rural location. The first row lists the distribution of physicians in each location for the 1987 – 1999 US MD Grads. These are preliminary studies to guide more detailed work. The relationships share significant interactions in similar socioeconomics, geographic, prestige, and other dimensions. The findings are consistent with current distributions. Those more familiar with match determinations, first time practice locations, or single program or medical school outcomes may not be familiar with studies of all graduates in multiple class years distributed at near equilibrium conditions. In such studies the types of physicians who stay in primary care and in distributional locations are more important than those in temporary obligations. Even in the 1987 – 1999 US MD Grads, the temporary obligation group is over-represented.

Odds Ratios, Practice Locations, 189,742 Recent US MD Grads

Those found in MMC locations are top MCAT school grads (1.6 times) younger than 26 at MS graduation (1.1) , and top income county born physicians (1.1). FPs are 0.29 times as likely to be found in major medical centers. Only half of FPs are in major medical centers. This leaves half of family physicians to distribute to locations only 20% of all other types of physicians. In all other rural and urban areas, the FP choice factor represents over 2 times greater odds ratio to have urban served and underserved locations and this increases to 3 times for rural served locations and to 3.5 times for rural underserved locations. For rural underserved location, bottom quartile income birth (born in lowest income county) has 1.8 times location rate, older graduates (over 29 years at MS graduation) have 1.4 times odds ratios, and rural instate birth is 1.4 times. For all rural locations, FP contributes a 3 times odds ratio and rural instate birth is 2 times.

FP choice is greater than lower income birth, rural birth, or older age factors regarding all forms of physician distribution outside of major medical centers.

It is likely that these concentration factors would be stronger if this secondary data was replaced by primary data on the individual physicians: MCAT scores, Board Scores, parents who are professionals, and parent income. For example there are Black, Mexican American, and White US MD Grads within the most urban areas that have lower or middle incomes. The county birth origin coding gives them the highest income levels by secondary assignment. There are also rural born, lower income origin, and older US MD Grads that graduate from top MCAT schools and fail to distribute. Children of professional parents may the most humble origins by birth county, but the lower levels of distribution. The elite student types attending a Duluth or Mercer also have nearly the same 30% rural choice and 40 – 50% choice of family medicine. There are diverse groups within the most urban areas and the highest income counties and the highest MCAT schools that do not share these major medical center concentrations.

Does Training Facilitate Distribution?  

1987 – 1999 US MD Grads can be categorized by age at graduation, birth origins, and birth county income levels.

Family Medicine: Facilitating Physician Distribution Beyond Birth Origins

The shaded cells represent distributions above 5.4% underserved and 10% rural, the national averages for this distributional locations.

Only two types of physicians distribute beyond the national averages: distributional types of students and those choosing family medicine.  Choice of family medicine doubles distribution across birth origins, schools, and other categories.

Family Medicine, More Than Just Origins   Rural origin family physicians distribute rural at 43%. According to origins and family medicine choice, the rural maximum should be 43%. A mix of younger, urban, and higher scoring types should result in lower levels of distribution. Rural graduate program outcomes extend beyond 45% rural location for integrated types and rural located programs, beyond 60% for rural training tracks, and beyond 90% in a few individual programs.

The following data includes gender, ethnicity, and race data that was available only on 1997 – 2003 Family Medicine Residency Graduates. Underserved distribution for Black or Mexican American family physicians is 20 – 25% or 4 to 5 times national averages. Underserved levels increase beyond 30% for inner city FP residency program graduates and for Native American FPs. Age appears to be a factor in facilitating distribution and is often found together with humble origins in the physicians found in the most complex locations. Older age and diverse origins combined with family medicine would be even more effective but those most likely to distribute are also most likely to be found in military or academic careers.

Perhaps the most interesting distributional type of program was the accelerated residency training program. Of 132 grads from 12 accelerated programs spread across the nation, 50% were found in rural locations and 18% were found in underserved locations. These levels are 5 times the rural distribution mean and 3 times the underserved distribution mean. This was 60% higher than the underserved distributions of white family physicians. Accelerated graduates were not selected for rural or underserved interest and did not train rural or underserved (with the exception of one program that had no different rural distribution). There are really only a few explanations for this distribution. The first involves an exclusive selection process where graduates desired family medicine and were willing to commit 1 year early and the program directors desired them based on perceptions of commitment, maturity, and readiness for a challenge). The unique origins factor in this group is age. Accelerated graduates were 56% older than 29 at graduation or nearly three times as old as US MD Grads (22%) and twice as old as typical family physicians (28%).

Racial and ethnic origins also interact with family medicine choice to facilitate urban underserved distribution (Black, Mexican American) and rural distributions (rural Whites or Natives). Unfortunately urban low income white physician distributions cannot be determined without specific data on individual incomes. This is a major missing piece since 69% of those in poverty are white and an even greater percentage of the middle income population is white. Admitting more rural, minority, lower income, or distributional types of students presents challenges since the barriers extend back to birth. Admitting more who will be found in family medicine is also difficult since this is a group defined by those with the greatest barriers of education and income. Some might despair, given the challenges of efforts requiring decades for impact. Much less is required to accomplish desired outcomes. The nation only needs a few percent more family physicians, a few percent more service oriented professionals, 1 more percent rural or underserved physicians, or 0.5% to 1% more GDP (double or triple the current 0.5%) spent on age 0 – 6 child development. Little change or decline in these areas is not compatible with health, education, economics, efficiency, or effectiveness outcomes for lower and middle income areas in most need.

Challenges Regarding Distributional Admissions

When considering the county origins of US MD Grads, correlations can indicate those most likely to gain admission. Using retrospective studies, admissions correlations can be compared for all physicians, those not found in family medicine, and for those found in family medicine. It is important to remember that the family physician group tends to be more of a combination of the various groups least likely to gain admission, older, lower in income, different, and more likely to distribute.

County Level Probability of Admissions 1987 – 1996 US MD Grads