Objectives. To determine current geographic distribution of pediatricians in the United States, to assess the changes in the geographic distribution of pediatricians between 1982 and 1992, and to identify factors associated with the distribution of pediatricians among the 50 states.
Methods. A data set was constructed using several published data sources including the American Medical Association Physician Masterfile as the principal source for physician information. The pediatrician-to-child population ratio (PCPR, the number of pediatricians per 100 000 people under 18 years of age) was calculated to compare the distribution of pediatricians among states and the distributional changes between 1982 and 1992. Lorenz curves and Gini indices were used to describe distributions and to compare distributions across time periods. Linear regression analysis was performed to assess the relationship between PCPR (dependent variable) with 9 predictor variables.
Results. Between 1982 and 1992, there was a 5.4% increase in the United States (US) child population and a 46.1% increase in the number of pediatricians in patient care. During that time period, the PCPR increased by 38.6% from 35.1 per 100 000 to 48.6 per 100 000. There was a more than 4-fold difference in the PCPRs of the highest state (Maryland, 84.3) and the lowest state (Idaho, 18.5) in 1992. The PCPR increased in all 50 states, but varied from a 4.1% increase in Wyoming to a 63.4% increase in Massachusetts. The Lorenz curve showed that pediatricians were less evenly distributed than all physicians, but more evenly distributed than pediatric cardiologists. Between 1982 to 1992 the Gini index decreased 9.8% for all physicians and 10.2% for pediatric cardiologists, but only 1.9% for pediatricians. Since a decrease in the Gini index signifies better overall distribution, these changes are relatively modest for pediatricians as a whole, especially when compared to other physicians. Regression analysis showed that a higher PCPR was associated with a greater number of residency positions per 100 000 children and with the per capita income of the state (R = .93).
Conclusions. The distribution of pediatricians does not parallel the distribution of the child population in the US, nor has this distribution changed substantially in spite of a 38.6% increase in the PCPR. Pediatricians tend to concentrate in states with high per capita income and in states with a larger number of residency training positions. The failure of market forces to improve the geographic distribution may require manpower policy changes designed to improve distribution in underrepresented states. The uncertain impact of market changes due to increased use of managed care could affect distributional requirements of pediatricians in the future.
Pediatrics is a primary care specialty dedicated to the care of children. However, the geographic distribution of pediatricians does not match the geographic distribution of the US child population.1 Despite the steady increase in the total number of pediatricians, persistent areas of relative undersupply continue. In the era of health care reform, debates on pediatric workforce issues have focused primarily on the overall supply of pediatricians and pediatric subspecialists and the relationship of physician supply to health care costs.2-6 The geographic distribution of pediatricians remains a relatively neglected issue.
Up to the present, the most comprehensive study on the geographic distribution of physicians was conducted by the Graduate Medical Education National Advisory Committee (GMENAC) in the late 1970s.7 The 1980 report of the GMENAC has influenced public policy on physician workforce for more than a decade. The assumptions, findings, and recommendations of the GMENAC report also elicited criticism from many quarters, and several investigations have challenged the methodology and findings. Since the time of the GMENAC report, the health care delivery system has undergone substantial changes, most recently with the rapid expansion of managed care. Several of the assumptions that were used in the GMENAC report may not accurately reflect present market realities. Moreover, the number of pediatricians has increased substantially over the past decade and the demographic composition of the pediatric workforce also continues to change.
Questions raised by the GMENAC report about the appropriate geographic distribution of physicians have been considered by other investigators.8-10 In a 1982 Rand report, Newhouse and colleagues suggested that the uneven distribution of physicians and relative undersupply in certain geographic areas is sensitive to market forces.11 They predicted that the relative uneven distribution of physicians would improve as a result of increasing supply.
In 1986, Congress authorized the formation of the Council on Graduate Medical Education (COGME) to provide an ongoing assessment of physician supply trends and to recommend appropriate policies to better match the supply of physicians to the health needs of the country.12In their third report, issued in 1993, COGME highlighted how poor geographic distribution would serve as a barrier to improved quality and access for all Americans.13 Recent estimates by COGME suggest that given current rates of production, the number of specialist physicians will greatly exceed need (by 125 000) in the year 2000, whereas there will be a modest shortage (by 20 000) of generalist physicians at that time.12 While pediatricians are included as part of the generalist physician workforce estimate, the specific projection for pediatrician supply and their distribution has not been examined.
To determine how the evolution of the health care marketplace, increase in pediatrician supply, and changes in the composition of pediatric workforce over the past decade have influenced the geographic distribution of pediatricians, we conducted the present study. Specifically this study focuses on the following questions: 1) What is the current geographic distribution of pediatricians in the US? 2) As the number of pediatricians has increased, how has the distribution of pediatricians changed over time? 3) What factors are associated with or have possibly influenced the distribution of pediatricians among the 50 states? 4) What are the policy implications of these findings for future manpower considerations?
A multistep process was used to construct the data base for this study. In this study, the geographic units of analysis were the 50 states of the US. Data on physicians were obtained from the American Medical Association (AMA) Physician Masterfile. The AMA Physician Masterfile was established by the AMA in 1906 and is updated periodically. The Masterfile contains nearly all physicians with MD degrees in the US, including graduates of international medical schools who are in the US and meet educational standards for primary recognition as physicians. Both members and nonmembers of the AMA, and federal and nonfederal physicians are included in the Masterfile. Data used in this study refer to nonfederal physicians only. The Masterfile classifies physicians according to their self-reported major professional activity into two categories: patient care and nonpatient care. As has been done in other studies, the present study includes only physicians whose major professional activity is in patient care and this does not include pediatricians whose primary activity is research, administration, or teaching. In addition, physicians in training (eg, residents and clinical fellows) were excluded. Therefore, physicians in this study refer to office-based physicians and hospital staff physicians. There is no differentiation between part-time and full-time physicians in the Masterfile. Physicians are classified by specialties according to the self-designated practice specialty reported by the physicians themselves. Pediatricians thus include board certified pediatricians and noncertified pediatricians. The discipline of pediatrics, as defined in the AMA Specialty Profiles and used in the present study, included general pediatrics, adolescent medicine, neonatal-perinatal medicine, pediatric endocrinology, pediatric hematology-oncology, pediatric nephrology, pediatric cardiology, and pediatric allergy and immunology. Because the primary purpose of this study was to investigate the overall distributional changes of pediatricians, we elected not to analyze differences among subspecialties since data in 1982 and 1992 were somewhat different. We selected pediatric cardiology as a representative subspecialty to compare the differences and changes in distribution between pediatric subspecialists and all pediatricians. We chose pediatric cardiologists since, among pediatric subspecialties, cardiologists spend a large portion of their time in direct patient care of their subspecialty patients. The Physician Characteristics and Distribution in the US, published yearly by the AMA, summarizes the data compiled in the Masterfile.14,15
We obtained information on the US population and child population from the Statistical Abstract of the United States published yearly by the US Department of Commerce, Bureau of Census.16,17 Data on the proportion of child population of each state living in a metropolitan area were calculated from1994 County and City Extra.18 Metropolitan areas refer to 327 metropolitan areas (including 240 metropolitan statistical areas) defined by the US Office of Management and Budget. The numbers of pediatric residency positions of each state were listed in theNRMP Directory 1992 Match, published by National Resident Matching Program.19 The NRMP Directory lists the pediatric residency programs and positions at all hospitals participating in the 1992 match. Data on pediatric beds were obtained from the American Hospital Association (AHA).20 The data include hospitals and units as reported by hospitals in the 1992 AHA survey. Data on per capita income in 1992 of each state was from the US Department of Commerce, Bureau of Economic Analysis.21 Data on percent of persons not covered by health insurance in 1992 was from the US Bureau of the Census.22 Data on the percent of population enrolled in health maintenance organizations (HMOs) in 1992 was from Group Health Association of America.23 Data on percent of population enrolled in preferred provider organizations (PPOs) in 1992 was from American Managed Care and Review Association.24Data on PPOs were not calculated in Arizona, Delaware, Hawaii, Nebraska, and Vermont due to reporting inconsistencies. The data from these various sources were merged to form a unified data set.
Pediatrician-to-Child Population Ratio (PCPR)
The physician to population ratio has been a commonly accepted method to reflect the physician resources available to the population. In this study, the PCPR of each state was calculated as the number of pediatricians per 100 000 people age 17 years or younger. The 50 states and the District of Columbia were ranked according to the calculated value of PCPR. The PCPRs were used to compare pediatrician manpower among the 50 states and to assess changes of pediatrician manpower between 1982 and 1992. To determine whether a low initial PCPR in 1982 was associated with a greater increase over this time period, we regressed the percent increase in PCPR with the level of PCPR in 1982.
Lorenz Curve and Gini Index
The Lorenz curve and Gini index are traditionally used to analyze the distribution of income and wealth.25 The Lorenz curve is a curve fitted to the percentile shares of income and distribution. Several researchers have used this method to analyze physician distributions.26-28 In this study, thex axis of the Lorenz curve represents the cumulative percentage of the child population as ranked by PCPRs. The yaxis represents the cumulative percentages of total physicians. The Gini index is derived from the Lorenz curve.29 The straight line between the points of origin and the maximum on the Lorenz curve graph represents a perfectly even distribution. The area between the straight line and the curved line is divided by the total area beneath the straight line to obtain the Gini index. Therefore, Gini index represents a value from 0 (perfect evenness) to 1 (maximum possible unevenness) and provides a standardized value to reflect the relative unevenness of distribution. In this study, the Gini indices are used to compare the unevenness of distribution among physicians in different specialties. In addition, Gini values in themselves have a linear relationship. Therefore, the increase or decrease in Gini indices can be used for comparison of the degree of changes in the unevenness of distribution over a period of time.
We compared the Lorenz curves and Gini indices to determine the changes of the unevenness of pediatrician distribution among the states between 1982 and 1992. They were also used to compare the differences in distribution among pediatricians, general practitioners and family practitioners, and all physicians (all MDs in the US). Pediatric cardiology was selected as a representative pediatric subspecialty to compare with the distribution of pediatricians.
To assess potential determinants of the geographic distribution of pediatricians among the states, linear regression was used. The dependent variable was the PCPR of each state. The independent variables included: the number of general practitioners and family practitioners per 100 000 children; number of pediatric residency positions per 100 000 children; number of pediatric beds per 100 000 children; percent of pediatricians who were female; percent of pediatricians who were international medical graduates; percent of population with no health insurance; per capita income (in $1000); percent of population enrolled in PPOs or HMOs, and proportion of child population living in a metropolitan area. All data used were from 1992 except for the estimates of the metropolitan population that came from the 1990 census. Because the District of Columbia was an outlier in PCPR, it was excluded in the regression analysis. The mean value was substituted for the 5 states with missing data on the proportion of populations enrolled in PPOs. All dependent and independent variables were weighted by the number of children of the state during the regression process. Variables that created excessive multicolinearity, as measured by tolerance, were excluded. Statistical significance was determined at P < .05. Standard weighted linear regression was conducted using the software package SPSS 6.0 for Windows (SPSS Inc, Chicago, IL).
Between 1982 and 1992, there was a 5.4% increase in the US child population and a 46.1% increase in the number of pediatricians in patient care. The national mean PCPR increased by 38.6% from 35.1 pediatricians per 100 000 population 17 years or younger in 1982 to 48.6 in 1992. States with high PCPR's were generally in the New England or Middle Atlantic regions. States with low PCPR's were mainly in the Mountain, West North Central or West South Central regions.
Table 1 lists the PCPRs of the 50 states in 1982 and 1992, the rankings of the states according to the PCPRs, and the percentage of changes in the PCPRs between 1982 and 1992. Of the 50 states, Maryland was the state with the highest PCPR in both 1982 and 1992. Its PCPR increased from 59.2 in 1982 to 84.3 in 1992. Idaho had the lowest PCPR in 1992 with 18.5 which showed little change from a value of 16.5 in 1982. The difference between the state with the highest and the state with the lowest PCPRs had become larger (from 3.6-fold in 1982 to 4.6-fold in 1992) over the past decade. The PCPR for the District of Columbia had increased from 108.0 in 1982 to 177.8 in 1992. The PCPRs had increased in all 50 states. The magnitude of the PCPR increase varied among the states. There were 11 states and the District of Columbia with PCPR increases that exceeded 50% between 1982 and 1992. Among the 50 states, Massachusetts had the greatest increase in PCPR with 63.4% and Wyoming had the lowest increase with 4.1%. A regression analysis indicated that the level of PCPR of a state in 1982 has no significant correlation with the percentage of PCPR increase between 1982 and 1992.
The overall distribution patterns and changes between 1982 and 1992 are shown by the Lorenz curves (Figs 1 and2). Among pediatricians, general practitioners and family practitioners, and all physicians, the distribution of general practitioners and family practitioners most closely paralleled the distribution of children throughout the country. The distribution of general practitioners and family practitioners was more even than the distribution of all physicians, and the distribution of all physicians was more even than the distribution of pediatricians. Pediatric cardiologists were less evenly distributed than pediatricians relative to the child population according to the 1992 Lorenz curves. A comparison of the Lorenz curves for pediatrician distribution in 1982 and 1992 in Fig 2 shows no obvious changes in the unevenness of pediatrician distribution during this period.
Table 2 lists the Gini indices calculated from the Lorenz curves for pediatricians, pediatric cardiologists, general practitioners and family practitioners, and all physicians in 1982 and 1992. The Gini index provides a standardized scale (from 0 to 1) that can be used to compare the distribution among physicians of different specialties at one point in time and the difference of a particular specialty over a period of time. Because of the linear relationship within Gini values, a decrease in the value of the Gini index (percentage) indicates the degree of improvement of distribution toward maximum evenness. The Gini index for all physicians had decreased by 9.8% between 1982 and 1992, and the index for general practitioners and family practitioners had increased by .5% during the same period. The index for pediatric cardiologists had also decreased by 10.2%. However, the Gini index for pediatricians had only decreased by 1.9%. The results indicate that, when compared with all physicians and pediatric cardiologists, pediatricians had much less improvement in developing a more even distribution over the past decade.
The linear regression analysis (Table 3) shows that the number of residency positions per 100 000 children (β = .45,P < .0001), and per capita income (β = .62,P < .0001) are independently associated with the PCPR in any given state (R = .93, P< .0001). The regression model suggests that the PCPR would increase by 5.1 for every additional residency position per 100 000 children and increase by 3.7 with an increase of per capita income by $1000 of the state.
The number of pediatricians is increasing much more rapidly than the number of children in the US. As a result, the ratio of pediatrician-to-child population had increased remarkably in the past decade with significant variation among the states. Although the number of pediatricians and the ratio of pediatrician-to-child population had increased in virtually all 50 states, the distribution of pediatricians has not significantly changed over this period. After controlling for several potential confounders, the factors associated with the distribution of pediatricians among states indicate that the number of residency positions and per capita income are significant predictors of pediatrician to child population ratio in any given state.
In 1980, the GMENAC report indicated that an ideal “child health physician” per child population ratio is 49.2 per 100 000 or approximately one child health physician per 2033 children. In calculating ideal physician-to-population ratios, GMENAC incorporated a variety of assumptions about the percentage of children cared for by different child health physicians (which include pediatricians, family practitioners, and nurse practitioners), the number and workloads of female physicians, and the amount of time allocated to direct patient care. These estimates were also based on a profile of the health needs of the child population and the estimated number of pediatricians to meet these needs. Not only have the needs of the child population changed over the intervening years, but training of pediatricians has changed, as has the make up and time allocation of the child health workforce. During the decade between 1982 to 1992 utilization of pediatricians increased especially for adolescents, and professional organizations like the American Academy of Pediatrics (AAP) believe that the demands for pediatricians will continue to increase due to growth of new morbidities including AIDS, behavioral and learning problems, and the need of other vulnerable populations that pediatricians are specifically trained to care for (AAP 1993).
By 1992, the national mean PCPR (48.6) was approximately what GMENAC predicted in 1979 would be an ideal ratio for all child health physicians in 1990. The GMENAC estimate included in the “child health physician” designation general and family practitioners, nurse practitioners and internists, as well as pediatricians. Thus, the relative comparison between the GMENAC projections of adequate supply and the current PCPR suggests that according to GMENAC criteria the US as a whole, and many states in particular, meet if not exceed, GMENAC criteria for adequacy. It should be noted that while many states are well above this projected ideal ratio, a substantial proportion of states are still well below.
Over the past decade, the health care delivery system has also undergone dramatic changes, most notably by the rapid growth of managed care and prospective payment mechanisms. With the significant changes in the marketplace, the assumptions and methodology adopted by the GMENAC to determine appropriate numbers of pediatricians and other providers may not be applicable to the evolving health care system of the 1990s. As managed care continues to spread and markets consolidate and become more fully integrated, the physician workforce needs and geographic distribution of physicians are also likely to be transformed. Because there is not one managed care model, different types of managed care organizations will utilize different staffing patterns. Some managed care organizations are increasing the use of family practitioners or nonphysician providers for providing primary care to children. For example, 18% of pediatric ambulatory visits were provided by nurse practitioners and physician assistants in Kaiser Permanente Portland in 1992.30 However, a recent analysis of available HMO data on staffing patterns indicates that the pediatrician-to-population ratios actually being utilized by some HMOs exceed the 1:2033 child health physicians to children ratios recommended as optimal by the GMENAC (unpublished data, AAP).
The increasing enrollment of children in managed care organizations and the staffing pattern within these organizations raise important policy considerations. At the present time what constitutes an appropriate ratio is neither well defined conceptually nor empirically (in terms of what is actually being done). If the use of family practitioners and nonphysician providers for children's care increases, pediatricians' role in some managed care organizations may resemble the British system where pediatricians take on the role of child health consultants to primary care providers. Additional research on the relative value of different staffing patterns in different managed care environments is warranted. The impact of managed care market changes on geographic distribution of pediatricians remains to be investigated, especially in relationship to child health outcomes.
In the latest pediatric workforce statement, the AAP identified the “geographic distribution of pediatricians” as one of the key factors influencing pediatric workforce.3 The AAP emphasized the importance of financial incentives at both state and national levels to retain pediatricians in underserved areas. In the present study, we have found the average per capita income of the state residents is one of the major predictors for the distribution of pediatricians among the states. The fact that physicians choose to practice in affluent areas had been studied and documented in other studies as well.31 While pediatrics is among the lowest paid medical specialties, the variations in the earnings of pediatricians tend to be not as wide as other specialties.32 Whereas the role of the economic factors in physicians' location choice is not clear, pediatricians tendency to concentrate in more affluent states might reflect other socioeconomic factors including the overall quality of life and other amenities, as well as their ability to achieve a certain target income associated with high per capita income in the areas they choose to practice in. It also should be noted that until the advent of managed care and other recent changes in health care financing, that most indemnity type insurance plans did not cover many routine health care maintenance services such as immunizations.33 Therefore since a substantive amount of pediatric ambulatory care was accounted for by out-of-pocket expenditures, pediatricians in the premanaged care era may have had more incentives to locate where individuals could afford such expenses. Again the transition to managed care is changing the mode and patterns of physician reimbursement dramatically.
Previous studies have also shown a strong propensity for new physicians, especially generalist physicians, to remain in states where they have finished their medical school or residency training.34-36 Many physicians choose their residency training in the state to gain helpful contacts for joining or establishing a practice upon completion of residencies. In our regression model, the number of residency positions in a state was the most significant predictor of the pediatrician-to-children ratio in that state. Our regression model indicates that an increase of 1 residency training position per 100 000 children is associated with an increase of 5.1 pediatricians per 100 000 children. This “amplifying effect” might be due to the tendency of pediatricians to stay in the same state where they completed their residency, and the cumulative effect over time. The GMENAC report recommended decentralizing medical education programs to improve distribution. The Washington-Alaska-Montana-Idaho (WAMI) Medical Education Program has provided evidence that decentralization with extensive exposure to medical practice in rural areas plays an important role in decisions about practice location.37 Adkins and associates reported that 23% of the graduates of the WAMI program end up practicing in nonmetropolitan areas whereas only 13% of US physicians locate in such areas.38 Manpower policies like the WAMI program could be considered in other areas where the PCPR is low and incentives are necessary to encourage better geographic coverage.
The last decade has also witnessed major changes in the composition of pediatric workforce. The total number of pediatricians has increased by more than 40%, and the proportion of female pediatricians has increased from 31% to 40%. The number of pediatric subspecialties and the number of pediatricians in subspecialties have grown.5,6 In spite of these changes, there was no remarkable change in their uneven distribution. Fritz and Lantos reported that female pediatricians are more likely to work part-time and those who work full-time tend to work fewer hours than their male counterparts.39 It is possible that the higher proportion of female pediatricians reduces the effective workforce therefore requiring more pediatricians per child population. The trend of increasing female pediatricians must be taken into account on assessing or planning for the geographic distribution of pediatricians. The increasing pediatric subspecialization and localization of pediatric subspecialists in urban areas will also be one of the main key factors in determining the geographic distribution of pediatricians in the future.5,6
Data used in this study are aggregate data at the state level. There are obviously tremendous variations of many of the variables examined within a particular state. Therefore, these results are not necessarily applicable to smaller geographic units such as counties or cities. While our study provides some insights regarding the uneven distribution of pediatricians among the states, we did not explore the differences in rural/urban locations or towns with various sizes of population. Further investigations on small area analysis are required to understand the factors influencing distribution of pediatricians at these lower levels of geographic aggregation.
The problems with multicollinearity in physician location studies have been addressed elsewhere.31 High intercorrelations between explanatory variables inflate estimates of standard errors and may cause statistical significance of certain variables to be understated. We believe that the tolerance for correlations set in our regression procedure, the fit of the model base on the significant predictors (adjusted R2 = .86), and the strength of the associations mitigate against erroneous conclusions.
In developing our predictive model we selected the factors suggested by previous studies to be most relevant to the distribution of pediatricians among the states. However, there might be other factors associated with the distribution of pediatricians that are not included in our regression model. Other factors such as increasing pediatric subspecialization might be influencing the persistent uneven distribution of pediatricians, yet we did not account for this in our analysis.
At the time of the GMENAC report, Newhouse and colleagues suggested that the uneven distribution of physicians would be sensitive to marketplace forces, and as the physician supply increased, there would be increased diffusion into previously underserved areas.11 Our study suggests that while the number of pediatricians has increased significantly, the distribution has not changed that much, especially in relation to other physician groups.
If a better geographic distribution of pediatricians is to be attained, several issues must be considered.
Marketplace: The changing health care marketplace, the accelerating impact of managed care and the corporate organization of health services will undoubtedly have a dramatic impact on the need, allocation and geographic distribution of pediatricians. Depending on how a managed care organization decides to staff its primary care services, pediatricians could become more or less essential. In certain geographic areas (inner cities and rural areas), managed care penetration and the organization of services may vary. Moreover, different types of managed care organizations may choose different staffing models based on both effectiveness and efficiency concerns.
Training: This study indicates that presence and productivity of pediatric training programs have important effects on geographic distribution. The changing marketplace, legislative imperatives to train more primary care physicians at both state and federal levels, changes in the allocation of Medicare Graduate Medical Education dollars, and the potential reallocation of Title VII funds, will all impact on future training goals and funding. If state policy makers and managed care organizations perceive a maldistribution and desire to take steps to mitigate this problem, changes in local and state training capacity are an alternative that can be considered.
Better information: Given the changes that have transpired in the past 15 years, additional information is needed and new simulation models should be constructed to address these distributions and system capacity questions. The changing threat to children's health, the need for additional and different types of services, the changing demographics of the workforce, and the overwhelming impact that vertically integrated managed care market will demand a fresh look at this essential set of policy questions. Studies addressing the relationship of pediatric staffing patterns to child health outcomes are essential.
The results of this study indicate that the distribution of pediatricians does not parallel the distribution of child population, that this distribution has not dramatically changed in spite of the significant growth in the number of pediatricians, and that marked difference in the distribution of pediatrician relative to child population continues among states. Irrespective of significant increases in the number of pediatricians in the workforce, there has been no remarkable change in the uneven distribution of pediatricians over the past 10 years, although improvement is seen in the distribution of all physicians. Given drastic marketplace changes, changes in the health needs of children and adolescents, and changes in the demographic profile of trainees and the number receiving subspecialty training, the issue of the appropriate geographic distribution of pediatrician demands further attention. Smaller geographic area studies and studies that focus on the outcomes and effectiveness of different organizational arrangements would help guide the development of appropriate policies.
This work was supported in part by a grant from the Maternal and Child Health Bureau of the US Department of Health and Human Services.
The authors would like to thank Sarah Brotherton, PhD of the American Academy of Pediatrics for providing data sources and advice for analysis. The authors would also like to thank Arleen Leibowitz, PhD and Jeffrey Stoddard, MD for critically reviewing this manuscript.
- Received February 2, 1996.
- Accepted May 13, 1996.
Reprint requests to (N.H.) Child and Family Health Program, UCLA School of Public Health, 10833 Le Conte Ave, Los Angeles, CA 90095–1772.
- US =
- United States •
- GMENAC =
- Graduate Medical Education National Advisory Committee •
- COGME =
- Council on Graduate Medical Education •
- AMA =
- American Medical Association •
- AHA =
- American Hospital Association •
- HMO =
- health maintenance organization •
- PPO =
- preferred provider organization •
- PCPR =
- pediatrician-to-child population ratio •
- AAP =
- American Academy of Pediatrics •
- WAMI =
- Washington-Alaska-Montana-Idaho Medical Education Program
- Federation of Pediatric Organizations, Task Force on Graduate Medical Education Reform
- American Academy of Pediatrics
- ↵American Academy of Pediatrics. Pediatric Manpower Recommendations. Elk Grove Village, IL: American Academy of Pediatrics; 1987
- Brotherton SE
- McCrindle BW,
- Starfield B,
- DeAngelis C
- ↵Graduate Medical Education National Advisory Committee. Report to the Secretary, Department of Health and Human Services, Geographic Distribution Technical Panel, Volume III. (DHHS Publication No. HRA 81–653). Washington, DC: Government Printing Office; 1980
- Morgan BC
- ↵Newhouse JP, Williams AP, Schwartz WB, Bennett BW. The Geographic Distribution of Physicians: Is the Conventional Wisdom Correct? Santa Monica, CA: RAND; 1982
- ↵American Medical Association, Department of Data Release Services, Division of Survey and Data Resources. Physician Characteristics and Distribution in the United States. 1993 ed. Chicago, IL: American Medical Association; 1994
- ↵American Medical Association, Department of Data Release Services, Division of Survey and Data Resources. Physician Characteristics and Distribution in the United States. 1983 ed. Chicago, IL: American Medical Association; 1984
- ↵United States Department of Commerce, Bureau of the Census. Statistical Abstract of the United States, 1993. 113th ed. Washington, DC: Government Printing Office; 1993
- ↵United States Department of Commerce, Bureau of the Census. Statistical Abstract of the United States, 1984. 104th ed. Washington, DC: Government Printing Office; 1984
- ↵Slater CM, Hall GE. 1994 County and City Extra. Annual Metro, City and County Data Book. Lanham, MD: Bernan Press; 1994
- ↵National Resident Matching Program. NRMP Directory 1992 Match. Washington, DC: National Resident Matching Program; 1991
- ↵American Hospital Association. 1993 AHA Hospital Statistics, 1992 Data. Chicago, IL: American Hospital Association; 1993
- ↵US Department of Commerce, Bureau of Economic Analysis. Survey of Current Business. September 1993:7
- ↵US Bureau of the Census. Money Income of Households, Families, and Persons in the United States: 1992. September 1993:60–184
- ↵Group Health Association of America. National Directory of HMOs—1993. Washington, DC: Group Health Association of America; 1993
- ↵American Managed Care and Review Association. AMCRA Foundation Managed Care Database
- ↵Todaro MP. Economic Development of the Third World. New York, NY: Longman; 1989:143–186
- ↵Ernst RL, Yett D. Econometric and statistical studies of the geographic distribution of physicians. In: Physician Location and Specialty Choice. Ann Arbor, MI: Health Administration Press; 1985
- ↵Goldberg JH. Pediatricians' earnings take a nosedive. Med Econ. Pediatric edition. 1994;13:44–49
- Gabel J,
- DiCarlo S,
- Sullivan C,
- Rice T
- Watson CJ
- ↵Wilensky GR. Retention of medical school graduate: a case study of Michigan. In: Scheffler RM, ed. Research in Health Economics. Greenwich, CT: JAI Press; 1979
- Fritz NE,
- and Lantos JD
- Copyright © 1997 American Academy of Pediatrics