Taxonomy, Themes, and Theories Related to Experiential Place

Most Needed Health Access is the Result of Physician Origins, Training, Career Choice, and Policy Regarding Health Access

Robert C. Bowman, M.D.    rcbowman@atsu.edu

Common Sense Health Access Outline

Principles of Health Access Summary Points

Steps to Health Access

Basic Health Access Concepts To Review

Review of one or more of the previous links may be required prior to the following.

Logistic regression studies assist in the understanding of the most important health access outcomes, variables, and relationships. Multiple variables can be loaded. This is an improvement over the bivariate studies. The bivariate studies are able to fit within the 3000 or 5000 word limits of major journals, but have little utility for use in health access workforce as multiple factors must be considered simulataneously, just as in the design of health access workforce. Bivariate studies also cause misunderstandings as they are easily misinterpreted or abused when a single factor is loaded without proper controls. (see Why Physician Workforce Studies Need New Tools)

The Principles of Health Access are consistently illustrated by the following theme. Most needed health access workforce is the result of more normal and less exclusive

·         In physician origins

·         In career choice (broadest generalist)

·         In training

·         In policy distributions regarding health funding

More normal and less exclusive results in needed health access. Most exclusive in origin, in career choice, in training, and in policy is clearly related to least health access contributions.

Other Themes and Theories

·         The characteristics of physicians associated with increased probability of admission are consistently associated with decreased probability of health access careers.

·         Lower population density physician origins are associated with rural practice location.

·         Lower income physician origins are associated with underserved (high poverty, lowest physician concentration) practice locations.

·         Exclusive origins, training, and career choices are associated with exclusive career and location choices.

·         More normal origins, training, and career choices are associated with health access careers.

·         Physician origins are more than just simple bivariate relationships involving concentrations of income, people, physicians, or resources. Physician origins are represented by a number of variables. More commonly physicians gaining admission represent combinations of concentration across birth to admission life experiences. These enhance admission probability but result in lower probability of needed health access.

·         Origins associated with social organization and political organization are also related to increased probability of admission and decreased probability of needed health access.

·         Humanistic characteristics such as service orientation, empathy, and choice of people oriented careers are also related to more normal and less exclusive origins, training, and career choices.

Factors related to needed health access careers can be loaded as variables using secondary data. These studies involve complete populations of physicians, a major advantage of secondary data. The studies on physician workforce using primary data are also in agreement with these studies although these studies are limited to fewer graduates and are limited in interpretation by the lack of controls.  

Most needed health access workforce contributions can be found for graduates who are older age at graduation, lower and middle income origin, first generation to college, lower and middle population density origin (not just rural origin), family practice physicians (and non-physicians), and graduates training in schools and locations with less exclusive concentrations of physicians, specialists, and health resources

The following logistic regression equations involve complete populations of physicians as found in career and practice locations in the 2005 Masterfile.

Odds Ratios of Needed Health Access Origin for 316,792                Graduates of 1987 – 2000

·         Admit Rate per 100,000 per Class year – physicians can be compared by birth origins to a denominator population to gain a ratio of admission per 100,000 people (1970 birth county population). For this time period about 7 were admitted per 100,000 per class year from US origins and this increases to 8 when including the foreign origin graduates of US medical schools that were born in other nations but completed education in the United States. The lowest probability of admission is about 3 to 4 per 100,000 with top probability in counties with top concentrations at about 14 – 20 per 100,000 per class year. Admission rates can also be calculated by physician specialty. Admission rates are steady for family medicine across this birth county range from lowest to highest probability of admission from 3 to 20. Admission rates for those that become the most exclusive specialists increase the most across birth origins. As probability of admission increases, the probability of an exclusive school increases, the probability of an exclusive career increases, and the probability of an exclusive location increases.  Admission Ratios and Physician Origins

The following columns are odds ratio probabilities of needed health access careers and locations. Each was obtained with logistic regression with at least 5 variables loaded.

The Dependent Variables – Most Needed Health Access Careers and Practice Locations

·         65% of the Population with 23% of Physicians (Practice Location Outside of Physician Concentrations and Outside of the Current Design) –A best indicator of needed health access involves marginal (not high poverty) and underserved locations (high poverty)in rural and in urban locations. This is a practice location in one of 30,000 zip codes outside of current physician concentrations that are found in 4% of the land area in 3400 zip codes.1-5

·         All Underserved – About 7 – 8% of US physicians are found in zip code practice locations with 21% of the population (1 to 3) that have less than 75 physicians that also have greater than 19% of the local (or adjacent) zip code population in poverty or a major designation such as a Community Health Center, a National Health Service Corps site, or a whole county primary care shortage area. These are locations averaging 80 – 100 physicians per 100,000 as compared to the national average of 300 or the 400 – 5000 physicians per 100,000 found in zip codes with top concentrations.

·         Urban Underserved – About 4% of US physicians are found in urban locations that are underserved average about 60 physicians per 100,000. They are often nearby the nation’s top concentrations of physicians that also have 100 – 200 primary care physicians per 100,000. Urban underserved locations are dominated by primary care and family practice MD, DO, NP, and PA forms.

·         Needed Rural – About 7 – 8% of US physicians are found in practice locations that are marginal or underserved rural locations that have less than 75 physicians. Physicians found in all rural locations are about 9 – 10% but this includes rural zip codes with 75 or more physicians. Areas with more than 75 physicians are properly coded as one of the 3400 zip codes with top concentrations of physicians. Not only do they involve top concentrations, they also have low percentages of primary care and family medicine physicians, graduate medical education positions, and concentrations of specialists and health resources just like top urban concentrations. Needed rural location represents rural locations that have greater physician needs and are rural locations that are dominated by primary care and family medicine. These are locations with 100 – 120 physicians per 100,000, the typical 30 family physicians per 100,000 found across the nation, and 40 – 100% of total physicians are primary care physicians.

·         Odds Ratio to Become FPGP – This final column is specific for the family practice or general practice career choice using the same origin and training factors. In this 1987 – 2000 cohort, very few general practitioners are involved. The same variables that result in greater probability of rural or underserved location also result in greater probabilities of family practice (older graduate, lower and middle income origin, not exclusive in origin or training, health access school).  

The odds ratios of most needed health access careers are comparable to the cardiovascular risk factors. Factors such as cholesterol, smoking, or hypertension multiply the risk of adverse  cardiac events by 1.2 to 2 times. For needed health access older age at graduation multiplies all needed health access careers by 1.3 times, choice of family medicine doubles or triples the “risk” of needed health access career choices, and origins shared with the 65% of the population left behind by health access doubles needed health access careers.

Since the logistic regression equations also have age of graduate, origin factors, and training factors, these are true representations of the actual doubling or tripling of needed health access. This is similar to the cardiovascular risk factors that must all be loaded simultaneously to prevent one or a few factors from being overrepresented.

The equations begin with a most needed health access outcome as the dependent variable and 4 or more independent variables are loaded - one or more origin variables, a physician age at graduation (younger, normal, slightly older, older), a career choice (family medicine, general, support, subspecialty), and a training factor (allopathic private, public, Caribbean, international, osteopathic, Top Ranking MCAT graduate, Health Access school grad).

 It is important to remember that each physician has origin, age, career choice, and practice location variables. The studies do not involve comparisons of some physicians to an outcome and other physicians to origin variables. Each physician is tracked from birth (approximately 1970) to medical school (1987 – 2000 graduate) and to career choice and practice location as found in the 2005 Masterfile.

Origins, Admission Rates, and Logistic Regression Equations for Most Needed Health Access Outcomes

Admit Rate per 100,000 per Class year

65% Pop with 23% Physi-cians

All Under-served

Urban Under-served

Needed Rural

Rural

Averages for 1987 – 2000 Grads

23 - 27%

7.5%

4.2%

7.0%

9.5%

14.2%

 

Experiential Place - Birth Origins Related to Concentrations of People

Origin % for n = 316,792

Admit Rate per 100,000 per Class year

Odds Ratios 65% Pop Outside of Concentrations

Odds Ratio All Under-served

Odds Ratio Urban Under-served

Odds Ratio Needed Rural

Odds Ratio Rural

Odds Ratio to BecomeFPGP

Birth 1000 PPSM

15.3%

7.94

0.997

0.761

0.869

0.981

0.987

1.081

Lower Population Density Origins less than 125 PPSM

15.8%

4.31

1.184

1.110

0.835

1.823

1.939

1.571

PPSM over 2500

55.7%

12.67

0.837

1.213

1.220

0.640

0.629

0.579

Predominantly Black County

3.3%

9.65

1.065

1.244

1.207

1.311

1.253

0.865

Birth Whole County PC Shortage

1.0%

2.32

1.560

1.843

1.273

1.959

1.763

1.568

Foreign Born US School 10%

10.4%

8.37

0.846

0.821

1.129

0.525

0.512

0.729

Rural Birth (RUCA)

8.4%

4.47

1.327

1.229

0.817

2.082

2.789

1.602

 

Experiential Place - Birth County Per Capita Income 1969

 

Top Income Quartile 51 million

20.4%

9.07

0.857

0.679

0.784

0.723

0.696

0.872

Second Quartile 51 million

17.8%

7.89

1.001

0.711

0.843

0.882

1.061

1.083

Third Quartile 51 million

14.6%

6.48

1.101

0.821

0.859

1.004

1.232

1.036

Bottom Quartile (60% rural origin)

9.6%

4.24

1.224

1.474

1.063

1.898

1.467

1.500

Admissions per 100,000 per class year including foreign born in US schools for the 1987 – 2000 grads

8.00

 

Experiential Place Related to Physician Career Choice in 2005

Origin % for n = 316,792

Admit Rate per 100,000 per Class year

Needed Health Access Outside of Concentrations

All Under-served

Urban Under-served

Needed Rural

Rural

Odds Ratio to BecomeFPGP

Subspecialty (Card, GI, NS, TS)

5.2%

0.490

0.566

0.679

0.382

0.551

Hospital Based Specialty

9.4%

0.697

0.600

0.700

0.564

0.652

Office General Pediatrics

5.7%

1.225

1.225

1.414

0.945

0.939

Office General Internal Medicine

10.8%

1.196

1.290

1.219

1.233

1.183

Family Practice

14.2%

3.111

2.392

1.762

3.490

2.715

Office Family Practice

11.0%

3.217

2.337

1.725

3.649

2.941

 

Experiential Place Related to Concentrations of Physicians

 

Birth in a Medical School County

 

Yes

55.6%

0.854

0.830

0.872

0.645

0.653

0.779

Not Known

19.4%

0.867

1.026

1.147

0.943

0.923

1.254

No

24.9%

1.295

1.205

1.063

1.370

1.736

1.574

 

Experiential Place - Life and Health Experiences Prior to Medical School - Age at Medical School Graduation

Origin % for n = 316,792

Admit Rate per 100,000 per Class year

Needed Health Access Outside of Concentrations

All Under-served

Urban Under-served

Needed Rural

Rural

Odds Ratio to BecomeFPGP

Less than 28 yrs (no delay, no life experience prior to medical school)

62.1%

0.813

0.821

0.860

0.754

0.752

0.582

Less than 26 yrs (early admission)

18.3%

0.785

0.937

0.940

0.680

0.698

0.481

Age 26 to 29 (normal age)

60.6%

0.937

0.843

0.875

0.762

0.886

0.900

Older than 27 yrs (any older)

37.8%

1.299

1.220

1.133

1.227

1.281

1.784

Older than 29 yrs (non-traditional)

18.4%

1.327

1.330

1.173

1.623

1.466

1.840

Older than 32 years (most life and health experience prior

10.0%

1.531

1.430

1.248

1.792

1.596

2.105

 

Experiential Place - Impacts of Selection and Training Not Supportive of Health Access

 

Top 10 MCAT

5.7%

0.565

0.620

0.730

0.482

0.492

0.319

Top 20 MCAT

17.4%

0.611

0.678

0.778

0.555

0.666

0.533

Allopathic Private US School

25.6%

0.661

0.735

0.805

0.623

0.671

0.535

Allopathic Private Exclusive

20.8%

0.677

0.595

0.709

0.536

0.576

0.415

Medical School India or China (top concentrations of population)

5.6%

0.796

0.941

0.940

0.779

0.669

0.440

Caribbean (70% US origin)

2.0%

1.065

0.850

0.955

0.770

0.794

1.221

 

Experiential Place – Impacts of Selection and Training Supportive of Health Access

Origin % for n = 316,792

Admit Rate per 100,000 per Class year

Needed Health Access Outside of Concentrations

All Under-served

Urban Under-served

Needed Rural

Rural

Odds Ratio to BecomeFPGP

Allopathic Private Not Exclusive

4.6%

1.161

1.309

1.526

1.011

0.952

1.226

Osteopathic (1990s grads 35 - 40% FPGP, more rural, older, few from most urban origins)

7.8%

1.526

1.050

0.840

1.728

1.588

2.944

All Allopathic Public

43.6%

1.184

1.061

1.020

1.065

1.459

1.206

Lowest MCAT (not Puerto Rican, Early Admission, Osteopathic, Military, Historically Black,)

31.0%

1.472

1.465

1.388

1.599

1.699

1.380

Health Access US (selected allopathic and osteopathic with mission, admission, training)

4.0%

1.500

1.514

1.237

1.813

2.123

2.226

U A Guadalajara, Philippines

3.0%

1.305

2.262

2.581

1.418

0.951

1.438

Historically Black

0.8%

 

1.430

1.932

2.328

0.859

0.825

1.423

 

 

 

 

 

 

 

 

 

 

Birth Origins Categorized by Concentrations

·         Birth 1000 PPSM – This is a neutral origin marker involving counties with average concentrations of people, income, and physicians. The logistic regression equations reflect this with 0.9 to 1.1 odds ratios – also neutral.

·         Lower Population Density Physician Birth Origins in Counties that had less than 125 PPSM – This 1970 birth origin is more closely related to lower concentrations of people, income, physicians, hospitals, and health resources. It is negatively associated (-0.835) with urban underserved workforce but is related to all other most needed health access careers. Many of the same counties are included in this group as compared to Bottom Income Quartile Counties, but the bottom income quartile counties do demonstrate some contribution to urban underserved workforce at 1.063.

Other variables illustrate the theme or consistent relationship of lower population density physician origins to physician rural location and lower income physician origins to physician underserved practice locations. Levels of income or poverty have more to do with underserved practice levels while population density reflects rural outcomes.

·         PPSM over 2500 – Only 33 counties in the United States had such top concentrations of people per square mile in 1970 and all have medical schools, concentrations of physicians, and concentrations of income.

·         Predominantly Black County – Counties with a majority of the population comprised of African Americans. These are generally found in the eastern half of the United States and in urban and rural counties. They include the least organized and least concentrated in people and income to the most powerful and most densely populated. The contributions of African American physicians have been significant to urban underserved locations although the rate is declining with changes in admission (highest income origins replacing lower and middle income as in all races and ethnicities). Rural counties with majority black populations demonstrate experiential place connections related to most needed health access. About 25% of the physicians born in predominantly black rural counties (birth about 1970) are found practicing in a predominantly black county in 2005. Origins associated with predominantly black counties in rural locations have the lowest probability of admission in the nation. Once again the theme of lowest probability of admission results in highest probability of needed health access. Also the 25% found in predominantly black counties doubles to 50% when the physician chooses family medicine. This confirms the impact of origins plus family medicine choice.

·         Birth Whole County Primary Care Shortage Area – Admission ratios are also lowest for those born in the counties with consistently the lowest concentrations in a number of dimensions. Those managing to overcome considerable barriers of income, education, and parents to gain admission also have top return to practice to the counties in greatest need of health access. They return to all needed rural and underserved locations in the nation at nearly 2 times higher levels. The urban underserved probability is 1.3 times but this is still 30% increased even though the physicians arising from such locations are predominantly rural in origin.

·         Foreign Born US School – About 10% of physicians entering the workforce from 1987 – 2000 graduates were born in other nations but attended US medical schools. In this group this was predominantly allopathic schools although an increase above 15% for US allopathic and osteopathic schools can be seen in more recent years. Those first generation to American and managing to gain admission to US medical schools are consistently higher income and most urban in origin in census and in medical association reports. They also tend to be children of physicians or professionals. About half are Asian and Asian populations in America are 90% foreign born or have a parent who is. About 50% of the Asian population can be found in 1% of the land area (51 counties) with 20% of the total population and 17% of the white population. About 32% of the Hispanic population and 22% of the African American population is concentrated in 1% of the land area in 51 counties. Asian populations regardless of origin (US or foreign) have 3 to 8 times greater probability of admission and have lowest contributions to rural locations also consistent for those of top urban origins. They also have lowest rates of primary care with family practice at lowest contributions. The urban underserved contributions are slightly increased as are all urban location rates. Concentrations in practice in 4% of the land area are the highest for Asian and foreign born physicians just as in all physicians with origins in most urban and highest income counties.

·         Rural Birth (RUCA) – When city and state of birth allows, more specific linkages can be made to zip code origins as compared to less specific county origins. Rural origins remain associated with half the probability of admission and twice the distribution to rural locations. Rural origin physicians have 18% rural location rates as do urban origin family physicians. With rural origin and family medicine career choice, the rural rate increases to 43%. With the addition of health access training and older age, the rate increases above 50%. The same is found for Native American origin family physicians that tend to have rural origins, older age, lower income origins, and allopathic public school training that all boost most needed health access. There is one impediment; however, family medicine training programs recruit Native American female family physicians as well as minority African American and Hispanic females. This takes the graduates with top probability of needed underserved location. Rural origin males have the highest probability of health access when choosing family medicine but also are older and this tends to result in military careers. African American and Hispanic males are also found in military careers at higher levels. Consistently lower and middle income populations serve in the military. Those older tend to have families and the military package is one of the only support packages that meet their specific needs, especially with the move from scholarships to loan repayment by state and federal programs. Loan repayments are too late for some types of medical students and also fail to impact the one career choice involved in top levels of primary care and most needed health access – family medicine. Once students pass up family medicine, they pass up decades of primary care service in the current policy and they also pass up any reasonable probability of needed health access location.

Experiential Place - Birth County Per Capita Income 1969

The counties were arranged by per capita income levels and population was summed from highest income county toward lowest until 51 million people were counted. This became the top income quartile counties. The bottom 51 million became the lowest income quartile counties.

·         Top Income Quartile 51 million – Higher probability of admission is again found for those born in the top income quartile counties as well as lower probability of needed health access at 0.7 odds ratios. The admission ratios are highest for the top status populations within these counties. There are also populations within the same counties in highest poverty with lowest rates of admission and with lowest health care outcomes.

·         Second and Third Income Quartile Counties are relatively neutral in health access contributions. It is likely that the higher status children within such counties gain admission and negate the generic secondary coding as second and third quartile in birth county income.

·         Bottom Quartile – The lowest income counties yield physicians that have lowest county income origins but 60% also have rural origin. With lowest income county origins multiple barriers of income, education, parents, and opportunity exist. Graduates entering medical school are older, also indicating additional years of delay. The probability of admission is for origins associated with lowest income quartile counties is lowest and the probability of health access contributions is highest. When states have lower investments in higher education and medical school positions, it is very difficult for those with multiple barriers to gain admission.

Experiential Place Related to Physician Career Choice in 2005

·         Subspecialty (Cardiology, Gastroenterology, Neurosurgery, Thoracic Surgery) – These subspecialty careers with more graduates were combined for a subspecialty cohort. This group is most consistently concentrated in super center practice locations with over 200 physicians at a zip code. These are typically medical school zip codes as well. Distribution is most limited to rural and underserved locations as well as locations outside of physician concentrations with less than 75 physicians at a zip code, 65% of the population, and only 23% of physicians.

·         Hospital Based Specialties include Radiology, Anesthesiology, and Pathology – These are also physicians that concentrate in the top concentrations of physicians, although not at the peak concentrations found in subspecialists.

·         Office General Pediatricians have modest contributions to urban populations in need of services and are slightly lower probability in rural locations. About 60 – 70% of pediatricians were office based in this group but fewer remain in this basic primary care indicator. The office based forms have top distribution outside of concentrations. Expansions of pediatricians fail to improve needed health access as pediatricians remain in top concentrations.

·         Office General Internal Medicine physicians – Expansions of internal medicine physicians also fail to improve needed health access due to practice locations in top concentrations, but there are also other reasons that internal medicine fails. About 90% of internal medicine graduates depart primary care for specialty careers. These departures from generalist office based care result in departures from most needed health access. The primary care losses have been most dramatic in new graduates with declines from over 50% remaining in primary care 15 years ago to 10% in the most recent class years. Failure in most needed health access also involves shorter duration of services for over 40% of internal medicine graduates that are foreign origin international medical school graduates (FIMG IM). These FIMG IM graduates lose 8 years of a career due to late entry and 30% deliver no health care in the US after graduation. This results in loss of half of the workforce compared to a US origin graduate. Another failure in more recent graduates involves changes in obligation programs. FIMG IM graduates were driven to primary care and rural underserved locations by the J-1 Visa obligations that were required until recent years. New graduates can bypass the J-1 Visa. Also after obligations the internal medicine graduates distribute much like other internal medicine graduates. Internal medicine clearly has top concentration of any physician specialty in counties with over 10,000 people per square mile.

·         Family Practice – The family practice career choice contributes at the highest levels to most needed health access. This is also intensified by top volume of primary care, top activity levels (least part time or inactive), top retention in primary care, and most years in a career delivered by family physicians. Family physicians are least likely of all physicians to be found in top concentrations of physicians and health resources and are most likely to be found in needed health access locations. Actually the current paradigm misrepresents family physicians. It is other physicians that change in distribution. Other physicians melt away as concentrations of people, income, physicians, and health resources decrease. Family physicians are found at a constant rate of 30 – 40 per 100,000 across all locations although rates can be lower in urban underserved areas where the lowest physician concentrations at 60 – 80 per 100,000 are found. Family physicians are documented in their equitable distribution to all populations in need of health access. Family physicians have 0.3 odds ratios of practice location in zip codes with 200 physicians, 2 times odds ratios for needed urban locations, three times odds ratios for rural locations, and 4 times odds ratios for isolated rural, lower income rural, or whole county primary care shortage locations. The stellar distribution is possible mainly because low percentages are found in top concentrations. Studies confirm only 4% of full time faculty are family physicians, about one-third of the national level.6 The origin, age, and training factors that contribute to rural or to underserved outcomes also contribute to increased choice of family medicine. This is why most needed health access outcomes are rural, underserved, and family medicine. Using birth county origins the rate of family medicine is 1 per 100,000 per class year across the most rural and lowest income to the most urban and highest income counties. The rate of underserved physicians is 0.5 per 100,000 per class year and the rural rate is 0.7 per 100,000 per class year. Family medicine, rural, and underserved careers are the only careers that arise at a constant level from all birth origins. This results in variation in such careers according to the probability of admission to medical school. As the probability of admission increases from the most rural and lowest income counties to the highest income and most urban, the probability of most needed health access decreases. In the most rural and lowest income counties about 25% are found in family practice or 1 admitted per 100,000 per class year compared to 4 admitted per 100,000 per class year. In the most urban and highest income counties once again about 1 per 100,000 per class year become a family physician but 14 – 20 per 100,000 are admitted for a rate of 5 – 6%. Rural location probability decreases from 20% to less than 5% for physicians with top concentration origins. Underserved physician probability decreases from over 15% to less than 5%. These “constant or set point” rates were set by policy impacting 1987 – 2000 graduates. With change from somewhat supportive policy to the current policy destructive for health access, the set point has changed. Family practice is not supported at the same rate as in the 1990s, and family practice, rural, and underserved choices are all declining in all origins. The new constants for family medicine, rural, and underserved choices are reduced by at least 40% which is a major reason why health access has become so difficult.7

·         Office Family Practice – Office family practice increases to 95% of family physicians in rural areas and is 70% or lower in the areas with greater concentrations where teaching, research, and administration take greater percentages. The physicians that designate themselves as office based have the optimal distribution and family physicians illustrate this principle. Family practice in the 1987 – 2000 graduates has been combined with general practice as both have similar origins and outcomes and the largest source of general practice physicians is actually osteopathic family physicians that have not been properly coded in the allopathic Masterfile data source. Also over the past one hundred years, family medicine and general practice have remained the broadest scope generalists – the most important consideration for most needed health access. Levels of FPGP have remained steady or declined over the past 100 years while total physicians, and specialty (non-primary care) physicians have rapidly increased. Departures from health access are actually tracked across multiple decades with only the 1970s and 1990s decades representing improvements in primary care and health access. Once again the database captures the 1987 – 2000 graduates as of 2005 careers and locations, leaving some time for this allopathic data sources to catch up in areas known to be behind including osteopathic physicians, international physicians, Puerto Rican graduates, and graduates who are minorities.  

Experiential Place Related to Concentrations of Physicians

·         Birth in a Medical School County – This group includes US and foreign origin graduates that have a city or county of birth listed that has a medical school. This was the result of an exhaustive search to match city origins to city locations with medical schools in other nations. It is interesting that 70% of US origin physicians and 70% of foreign born physicians that had origins listed in the Masterfile had origin noted as a medical school county or a city with a medical school. Birth in a medical school county is consistently associated with lowest family medicine, primary care, rural, and underserved outcomes. For each birth origin, birth in a medical school county slightly increases the probability of medical school admission.

·         Not Known – This is a mixed group of physicians that includes physicians with missing birth origin data including about 5% of Allopathic US graduates, 30% of osteopathic graduates, and 50% of international graduates. The osteopathic graduates in the group contribute to the increased family medicine choice.

·         Not Born in a Medical School County – The physicians with origins separate from concentrations of medical school physicians have consistently greater distribution to most needed health access and choose family medicine at higher levels. Those not born in a county with a medical school also have consistently lower rates of medical school admission that also fit the pattern or lower admission ratio and greater contribution to rural, underserved, primary care, and family practice careers.

·         Studies coding counties with top concentrations of physicians also note that physician origins associated with top concentrations of physicians also are associated with lower health access contributions.

Experiential Place - Life and Health Experiences Prior to Medical School - Age at Medical School Graduation

The physicians with any delay in entry to medical school graduate are typically age 28 or older at graduation from medical school and have 1.2 or greater odds ratios of rural or underserved location. The odds ratios of needed health access careers increase with physician age at graduation. Older graduates also have greater odds ratios of family medicine choice even when controlling for birth origins. It is important to control for birth origins and career choice as physicians of lower and middle income origins (rural or urban) are more likely to be older at graduation and are also more likely to choose family medicine.

·         Less than 28 yrs at medical school graduation (no delay, least life and health experiences prior to medical school) – The peak graduation ages are age 26 and 27 in United States medical schools. This is a group that is admitted to medical school directly after college. This allows no time for additional life experiences to dilute a pure academic focus. There is no time for an additional health career as with graduates delayed 4 or more years. Younger students have only experienced a dependent life first on parents and then dependent upon the academic life. Their admission is more about parent influences such as college preparation, college selection, income, and education. They have had few if any setbacks in life and few that their parents could not modify before, during, or after the setback.

·         Less than 26 yrs (early admission) – These are students that have the least life experience. They have often skipped grades or entered grades or college earlier. Such a feat requires income, education, professional parents, or top career advisors to negotiate complex steps in a shorter period of time.

·         Age 26 to 29 (the predominant graduation ages) – The normal age graduates include those graduating at age 26, 27, 28, and 29 with entry at age 22 – 25. Reduced health access contributions are found.

·         Older than 27 yrs (any older) – At least some delay in medical school entry for these students and 1.2 times odds ratios of needed health access.

·         Older than 29 yrs – 4 or more years older is considered a non-traditional medical student but life and health experiences and marriage and financial decisions are made.

·         Older than 32 years – these are mature students that have made a number of important life and health decisions prior to medical school. In many cases they have established a life and then integrate a medical career into their life and family. A great deal of sacrifice can be involved. They are also more likely to be on their own rather than dependent on parents and significant financial risks are involved in application, admission, and training. Many have applied more than a few times to medical school. Odds ratios of most needed health access careers are highest for the oldest graduates. As with other types of graduates with most needed health access, fewer are admitted.

Experiential Place - Impacts of Selection and Training

Previous factors have been about origins and career choices by the graduates. The following factors are about the school of graduation. Medical College Admission Test scores fail to indicate a higher quality physician but they do indicate increased selectivity to a narrow range of origins.  In addition the most exclusive schools with regard to MCAT scores also train in the most exclusive concentrations of physicians, specialists, and health resources. The end point regarding physician workforce is lowest levels of most needed health access careers even when controlling for multiple origin factors, career choice, and age at graduation.

The United States requires a biomedical research effort and some level of biomedical focus for medical education, subspecialty careers, and hospital based physicians. But it also requires a solid foundation for all of health care delivery with a primary care and health access component. In the current design, the top ranking schools capture the most lines of revenue and the highest levels of revenue in each line. It is not a surprise that nearly all schools are attempting to emulate the top MCAT schools that do best financially, as presented in the media, and as represented by medical associations.

·         Top 10 MCAT – The schools with lowest contributions to health access at half of the probability controlling for origins, career choice, and age of graduate.

·         Top 20 MCAT – Still lowest contributions to needed health access. Contributions to family medicine were low for the top MCAT schools and have virtually disappeared in more recent years along with primary care contributions now down to 1 – 2 Standard Primary Care years per graduate. A health access school such as Duluth requires 6 graduates to equal the entire primary care output of 100 or more graduates of an exclusive school. The rural primary care contribution of a Duluth graduate is 64 times the contribution of a top ranking MCAT graduate.

·         Allopathic Private US School – A group also with lowest health access contributions. Allopathic private schools as a group are the most exclusive, but there are allopathic private medical schools that have made significant contributions to health access. Removal of the three Historically Black schools, Loma Linda, Creighton, and Oral Roberts (closed) from this group reduces the allopathic private contribution to the lower levels seen in top 10 and top 20 MCAT schools. In the 40 zip codes with 50% of Massachusetts physicians, allopathic private graduates are over 60% of the total. Family practice is only 2% and primary care levels are 16% in this Boston area dominated by allopathic private graduates and allopathic private schools. Health reform is unlikely when these schools and graduates do better by following US policy that rewards hospital and specialty careers. Origins, training, career choice, and policy all align for lowest health access contributions. Family physicians can be tracked moving away from such concentrations throughout their lives from birth to medical school to training to practice location to subsequent practice location. Locations with top concentrations are not hospitable locations for those that become broad generalists before, during, and after training.

·         Allopathic Private Exclusive – Lowest health access contributions are found when the few allopathic private schools that address health access are removed.

·         Allopathic Private Not Exclusive – Howard, Meharry, Morehouse, Loma Linda, Creighton, and Oral Roberts meet or exceed national averages in underserved location. Oral Roberts fits the pattern of closure of a health access school. Howard underserved contributions are likely to be less than the national average given changes over the class years. Meharry contributions are lower. Morehouse, Loma Linda, and Creighton contributions have remained stable although declining family medicine choice will decrease health access contributions even though training influence remains supportive for health access.

·         Caribbean (70% US origin) – Caribbean schools are an option for some dedicated lower and middle income and rural origin students. But the exclusive origins and the exclusive practice locations of Caribbean graduates indicate that Caribbean schools are mostly a way for exclusive origin students to gain admission that did not manage to find a US school to accept them. Even the increased choice of family medicine is also about a return to the United States for graduate training and this often requires an open residency position. Ross University dominates this group with 469 graduates a year and 115 of these graduates were family physicians in the 2009 class year. Ross is the number one source of primary care and health access for the United States contributing one of the highest levels of primary care per graduate and by far the most graduates.

·         Medical School India or China (top concentrations of population) – Schools with students from the ultimate most urban origins in the world are found in the most exclusive practice locations in top concentrations.

·         Osteopathic Schools – Most osteopathic schools make significant health access contributions, particularly in middle and lower population density counties of the United States. Unlike allopathic schools where the MCAT score appears to determine health access contributions, osteopathic schools with the oldest graduates contribute the most to needed health access. Family medicine choice follows the same pattern. In these 1990s graduates osteopathic schools averaged 35 - 40% that chose family practice careers. This was actually a decline from previous decades with 60% family practice graduates. Osteopathic graduates have also continued to change in origins to become much like allopathic graduates with fewer from rural, lower income, and middle income origins and more from the most urban and highest income origins. Graduates of the osteopathic schools in West Virginia, Missouri, Texas, Ohio, and Iowa have continued with top health access contributions. West Virginia School of Osteopathic Medicine manages to have 22% of graduates in underserved rural locations even though West Virginia physicians have only 16% rural underserved location. It is an incredible accomplishment to exceed state workforce parameters and WVSOM does so with graduates found in the most needed Appalachian rural locations. In addition, WVSOM medical students have earned top honors in the humanistic dimension in their second COMLEX board testing. Side effects of health access selection and training appear to be service orientation and people orientation. Not all osteopathic schools make health access contributions. Graduates of New York, Chicago, New Jersey, and Philadelphia osteopathic schools have underserved contributions that are below the national average of 7 – 8%. Not surprisingly admission and training involves top concentrations in these most metropolitan locations. Selection, training, and limited family practice contributions combine for lower health access contributions. Rapid declines in osteopathic choice of family medicine, changes toward more exclusive admission, changes in training, and graduates moving to allopathic graduate training will limit osteopathic health access contributions in the future.  Some indication of the cumulative change in admission can be provided by changes in MCAT scores. Osteopathic matriculants have been increasing in MCAT scores over the past 7 years at a rate 30% higher than allopathic US graduates. Some have expressed concerns about osteopathic data in allopathic databases. To reduce inaccuracies due to delays in data gathering, the most recent 5 class years from 2001 to 2005 were not used. The 1998 – 2000 osteopathic graduates do have some missing graduates and some inaccuracies in career and location choice due to delays found in the allopathic focused Masterfile, but the 1987 – 1997 graduates or 11 of 14 class years are comparable.

·         All Allopathic Public Graduates – Controlling for origin and other factors, allopathic graduates do have greater odds ratios of most needed health access, especially as compared to allopathic private schools with 0.5 – 0.6 odds ratios of needed health access careers. The average allopathic public school graduate had enhanced health access choices as noted in the 1990s graduates. It is only the top ranking allopathic public schools with exclusive selection and training that fail in health access contributions. They also have even lower contributions due to more exclusive admission and training. This is important to understand as the lower MCAT schools admit the most older, rural origin students, lower income, and middle income origin students and graduate the most family physicians. Specific training influence is still noted beyond these factors, but the combined impact of selection and training is significantly higher.

·         Lowest MCAT Allopathic US Graduates – These are not graduates of Puerto Rican, Early Admission, Osteopathic, Military, or Historically Black medical schools. The lower MCAT graduates have greater odds ratios of most needed health access. Sadly many of these schools graduate only half of the family physicians of a decade ago and this significantly limits their health access contributions. In the case of Mercer the fall has been dramatic from top health access contributions to average contributions with a decline from 32% family medicine to less than 8% indicating the changes in admission, training, and policy influences.

Schools Making Contributions To Health Access

The US MD schools were divided into different groups based on admission and training focus as well as MCAT scores of matriculants. This is a ranking based on selective admissions. The MCAT groups listed did not include Puerto Rican, Early Admission, Osteopathic, Military, International, or Historically Black medical schools that are different in admission and training, The schools with MCAT Grouping were ranked by the MCAT averages of 2000 – 2003 matriculants.

·         Health Access US school graduates - Allopathic and osteopathic schools with a focus on health access were selected from schools spread across the nation. These were schools with mission, admission, and training focus on health access. These are schools that increase needed health access by 1.5 to 2 times above origins, age, and career choice. The health access contributions can be rural and urban or specific for rural or urban underserved. Because health access schools select students with different origins and emphasize family medicine career choice, the effect of a health access school is maximal such as 16 SPC years per graduate at Duluth with 6 rural SPC years per graduate. Health Access US schools were selected as a national sample from across the nation for rural and urban health access contributions. Many were created in the 1970s during the last period of significant federal and state support for medical education - UC Davis, Mercer (optimal then, minimal now), Morehouse, Southern Illinois, Duluth, Kirksville College of Osteopathic Medicine, University of Nebraska Medical Center, Brody East Carolina, Wright State, East TN State, U N Texas TCOM, Marshall, and West Virginia School of Osteopathic Medicine. Once again the contributions of health access schools are confirmed using controls for these variables and remain even when loading multiple origin variables (population density origins, birth in a county with a medical school or not, county income origins).

·         U A Guadalajara, Philippines – Top health access contributions have been provided by these schools in underserved dimensions, particularly urban underserved populations. There is no Historically Hispanic medical school that focuses on the needs of Hispanic populations. In fact the Hispanic population in need of care is the fastest growing and Hispanic medical students gain admission at the lowest probability along with others with the most cumulative barriers to admission that also result in top health access contributions for those gaining admission. The contributions of UAG were intensified by the Fifth Pathway that also emphasized family medicine choice, Spanish language fluency, and a fifth pathway to entry to the US workforce. This pathway like other pathways (family medicine residency programs, the accelerated family medicine model) has been terminated. Philippines graduates are unique among all other medical schools examined in that graduates choosing family medicine have average rural and underserved contributions rather than the higher level found in all other schools and origins. The Hispanic population in need of care is the fastest growing and Hispanic medical students gain admission at the lowest probability. In addition the Hispanic medical students gaining admission are increasing fastest in parent income level.

·         Historically Black Graduates – Graduates of Historically Black schools have contributed to most needed health access for over a century. Howard and Meharry were the only schools to survive Flexnerian reforms. Drew and Morehouse were added. The contributions continue but declines can be seen in underserved location rates across the last few decades of class years. The rates have declined from 30% underserved location rates to 15%. This is still twice the 7 – 8% level of the nation’s physicians but further declines are expected with changes in admission, in training, and 80% decline in choice of family medicine in the past decade. Changes in admission, training, and career choice coupled with US health policy devastates most needed health access.

·         The only way to insure failure in health access is to move steadily toward more exclusive in origin, more exclusive in career choice, more exclusive in training, and more exclusive with health policy concentrating 90% of the funding related to physicians in 4% of the land area in top concentrations of physicians in 3400 zip codes. This is exactly what the United States health care design has implemented.

·         Also workforce studies imply that only the extremes in origin distribute where needed such as rural origin to rural location or lower income graduates to underserved locations. Actually it is the extremes regarding most exclusive that fail to distribute. A movement away from most exclusive is associated with increased health access contributions.  

·         Consequences of exclusive in origin, training, career choice, and policy are more than just destruction of physician health access. The consequences include non-physician workforce as physician assistants and nurse practitioners continue to depart primary care with each class year and year after graduation.  Currently 4 or 5 primary care sources graduate less than a majority into primary care with the lowest primary care of all in internal medicine with 10% primary care retention. This source of the most primary care only a few decades ago now contributes the least primary care despite the most graduates per year at over 8000. Nurse practitioners and physician assistants have increased over the past 40 years to 7000 graduates from each source. If the initial primary care and family practice contributions remained intact, the United States would have resolved many underserved locations. Unfortunately only 30 – 40% remain in primary care. It is indeed a poor health access design where the most dependable primary care sources that remain at the highest levels in primary care contribute the fewest annual graduates. Family medicine graduates have decreased from 4000 to 3000 per year as fewer medical students can trust a permanent primary care choice.  Pediatric graduates have increased to 3000 per year but pediatric primary care contributions have not increased as primary care retention levels have declined from 80% to the 50% level. In some states most in need of health access, pediatric programs contribute no new graduates remaining in primary care for 1 or 2 consecutive class years. Much the same is true in family medicine choice. The top ranking MCAT medical schools often had 0 to 1 family medicine graduate a year and more schools slightly lower on the MCAT ladder are added to this 0 – 1 group each year. With internal medicine and pediatric sources less reliable with each passing year and without family medicine contributions, more US schools are added each class year to the exclusive category of minimal health access contribution.

·         With the departure of policy supporting primary care and health access and continued poor support for the US lower and middle income populations in most need of health access, what remains for most needed health access is physician origins, medical education training, and family practice choice. Only family medicine remains a permanent primary care form with 85% retention in primary care. The unique US primary care design with 22,000 of 28,000 primary care graduates arising from flexible primary care training has uniquely failed in health access. The United States graduates half enough primary care per class year with 8000 internal medicine graduates and 3000 family medicine graduates. If the nation graduated 8000 family medicine graduates, the United States would begin to address needed health access although another 20 – 30 years would be required for recovery.

·         Matters are more complex for non-physicians. Nurse practitioner and physician assistant most needed health access vanishes as the NP and PA graduates move away from the broad generalist family practice mode during training, at graduation, and each year after graduation. With departures from family practice and primary care, even the massive doubling of PA graduates from 1998 to 2008 is not able to boost primary care as primary care retention declined from 54% to less than 28%.

·         The same multiplication of needed health access is found for all forms of family practice, (MD, DO, NP, and PA) that have 50 – 60% of family practice forms found in zip codes with 65% of the population and only 23% of US physicians and non-physicians. But the departure of NP and PA from family practice destroys not only primary care contributions but also rural and underserved contributions.

·         To understand most needed health access, one must understand actual primary care delivery (not just claims), distributions of physicians according to concentrations of physicians (not just geographic or income related), the impact of health policy regarding primary care and the populations most in need of health care that most depend upon real primary care, and the need to plan health access decades in advance.

·          There is never a failure with policies that produce more primary care than needed as this adds workforce versatility and the ability to respond within a few years to any number of public health, mental health, procedural, hospital (hospitalist), emergent (emergency medicine), and other needs.

·         The major failure in workforce design is a failure in needed health access as this is the workforce that takes the most time, coordination, and planning. For over 40 years the successful state and medical school efforts have taken decades to result in significant health access workforce. This is because the United States has never established a basic health access foundation and has always attempted with poor result to produce health access with the same admission, training, and policy process that emphasizes hospital and specialty careers.

·         Specialty workforce has been able to arise rapidly from other specialties or primary care for over 50 years.

 

These studies are supported by 400 pages of text and over 80 tables and graphics.

The US design also is responsible for limitations in needed health access.

·         Origins – physicians with top concentration origins have dozens of advantages that aid in admission throughout the birth to admission process

·         Family medicine and primary care reimbursement levels are the lowest with hospital and specialty careers rewarded at the highest levels.

·         Optimal health access training programs and schools have been terminated. Schools and programs making top contributions graduate lower numbers at least partly due to poor support for their health access mission. The exclusive schools have the most lines of revenues while health access schools have the least support.

Other data sources include associations8-13and other publications on complete populations.14-17

 

1.            Birth Origins Database from American Medical Association Masterfile Using OfficeMax Practice Locations. Medical Marketing Service; 2005. Updated Last Updated Date.

2.            Bowman RC. Physician Distribution By Concentration. Primary Care Research Methods and Statistics Conference. San Antonio, Texas; 2007.

3.            Bowman RC. Logistic Regression and Rural Practice Location. In: Proceedings, Association of American Medical Colleges 2007 Workforce Conference; 2 May; Washington DC, 2007.

4.            Bowman RC. They really do go. Rural Remote Health. Jul-Sep 2008;8(3):1035.

5.            Bowman RC. The illusion of minority status. Rural Remote Health. Jul-Sep 2008;8(3):1036.

6.            Barzansky B, Etzel SI. Medical schools in the United States, 2007-2008. JAMA. Sep 10 2008;300(10):1221-1227.

7.            Bowman RC. Five Periods of Health Policy and Physician Career Choice Submission.

8.            American Academy of Physician Assistants. Data and Statistics.  http://www.aapa.org/research/index.html, 2009.

9.            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.

10.          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.

11.          Goolsby M. 2004 National NP Sample Survey Comparisons Over 15-Year Period.  http://www.aanp.org/AANPCMS2/ResearchEducation/Research/AANP+Research+Archive.htm

12.          Goolsby MJ. AANP Survey Report 2002.  http://www.aanp.org/AANPCMS2/ResearchEducation/Research/AANP+Research+Archive.htm

13.          Goolsby MJ. 2004 AANP National Nurse Practitioner Sample Survey, part III: NP income and benefits. J Am Acad Nurse Pract. Jan 2006;18(1):2-5.

14.          Bowman RC. Measuring Primary Care: The Standard Primary Care Year. Rural Remote Health. Jul-Sep 2008;8(3).

15.          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/.

16.          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.

17.          Larson EH, Hart LG. Growth and change in the physician assistant workforce in the United States, 1967-2000. J Allied Health. Fall 2007;36(3):121-130.

 

Featured Graphic

Past and Present Physician and Non-Physician Numbers with Estimated Primary Care and Non-Primary Care Graduates   The United States has had a steady population growth pattern from 1960 and this will continue for a few more decades before slowing. Growth of specialist physicians, specialist nurse practitioners, and specialist physician assistants has increased more rapidly than population growth. Growth of primary care physicians, family physicians, primary care physician assistants, and primary care nurse practitioners has been stagnant for decades and will continue to remain flat while the US population increases and the most complex populations needed the most primary care such as the elderly double.

Principles of Health Access Summary Points

Steps to Health Access

Basic Health Access Concepts To Review

The Basic Table - Taxonomy, Themes, and Theories Related to Experiential Place and the Principles of Health Access

Experiential Place and Health Access Considerations

The Counterproductive and Untrue Perspective of the Impossibility of Health Access

Nebraska: A Practical Application of Experiential Place and Workforce

Why Physician Workforce Needs New Tools (and a health access perspective)

Admission Probability and Experiential Place: Admission Ratios and Physician Origins

www.basichealthaccess.org

Physician Workforce Studies

www.ruralmedicaleducation.org