Assessment of Changes in Rural and Urban Primary Care Workforce in the United States From 2009 to 2017 | Health Disparities | JAMA Network Open | JAMA Network
[Skip to Navigation]
Sign In
Figure 1.  Distribution of Primary Care Physicians and Nonphysician Clinicians Across US Counties, 2009 vs 2017
Distribution of Primary Care Physicians and Nonphysician Clinicians Across US Counties, 2009 vs 2017

The map contours conform to the default setting of the mapping platform ArcGIS (esri).

Figure 2.  Trends in the Density of Primary Care Clinicians in US Rural vs Urban Counties From 2009 to 2017
Trends in the Density of Primary Care Clinicians in US Rural vs Urban Counties From 2009 to 2017
Table 1.  Medians of the Density of the Different Primary Care Clinicians Stratified by County Rural and Urban Status, 2009-2017
Medians of the Density of the Different Primary Care Clinicians Stratified by County Rural and Urban Status, 2009-2017
Table 2.  Annual Mean Changes in the Density of the Different Primary Care Clinicians in US Rural and Urban Counties, 2009-2017
Annual Mean Changes in the Density of the Different Primary Care Clinicians in US Rural and Urban Counties, 2009-2017
Table 3.  Association Between Density of Primary Care Clinicians and Sociodemographic Characteristics in US Rural and Urban Counties, 2009-2017a
Association Between Density of Primary Care Clinicians and Sociodemographic Characteristics in US Rural and Urban Counties, 2009-2017a
1.
Meit  M, Knudson  A, Gilbert  T. The 2014 update of the rural-urban chartbook. Accessed September 30, 2020. https://ruralhealth.und.edu/projects/health-reform-policy-research-center/pdf/2014-rural-urban-chartbook-update.pdf
2.
Leira  EC, Hess  DC, Torner  JC, Adams  HP  Jr.  Rural-urban differences in acute stroke management practices: a modifiable disparity.   Arch Neurol. 2008;65(7):887-891. doi:10.1001/archneur.65.7.887 PubMedGoogle ScholarCrossref
3.
O’Connor  A, Wellenius  G.  Rural-urban disparities in the prevalence of diabetes and coronary heart disease.   Public Health. 2012;126(10):813-820. doi:10.1016/j.puhe.2012.05.029 PubMedGoogle ScholarCrossref
4.
Kulshreshtha  A, Goyal  A, Dabhadkar  K, Veledar  E, Vaccarino  V.  Urban-rural differences in coronary heart disease mortality in the United States: 1999-2009.   Public Health Rep. 2014;129(1):19-29. doi:10.1177/003335491412900105 PubMedGoogle ScholarCrossref
5.
Singh  GK, Siahpush  M.  Widening rural-urban disparities in life expectancy, U.S., 1969-2009.   Am J Prev Med. 2014;46(2):e19-e29. doi:10.1016/j.amepre.2013.10.017 PubMedGoogle ScholarCrossref
6.
Starfield  B, Shi  L, Macinko  J.  Contribution of primary care to health systems and health.   Milbank Q. 2005;83(3):457-502. doi:10.1111/j.1468-0009.2005.00409.x PubMedGoogle ScholarCrossref
7.
Ricketts  TC.  Workforce issues in rural areas: a focus on policy equity.   Am J Public Health. 2005;95(1):42-48. doi:10.2105/AJPH.2004.047597 PubMedGoogle ScholarCrossref
8.
Kirby  JB, Yabroff  KR.  Rural-urban differences in access to primary care: beyond the usual source of care provider.   Am J Prev Med. 2020;58(1):89-96. doi:10.1016/j.amepre.2019.08.026 PubMedGoogle ScholarCrossref
9.
MacDowell  M, Glasser  M, Fitts  M, Nielsen  K, Hunsaker  M.  A national view of rural health workforce issues in the USA.   Rural Remote Health. 2010;10(3):1531.PubMedGoogle Scholar
10.
Rosenblatt  RA, Hart  LG.  Physicians and rural America.   West J Med. 2000;173(5):348-351. doi:10.1136/ewjm.173.5.348 PubMedGoogle ScholarCrossref
11.
Bowman  RC.  Measuring primary care: the standard primary care year.   Rural Remote Health. 2008;8(3):1009.PubMedGoogle Scholar
12.
Friedberg  MW, Hussey  PS, Schneider  EC.  Primary care: a critical review of the evidence on quality and costs of health care.   Health Aff (Millwood). 2010;29(5):766-772. doi:10.1377/hlthaff.2010.0025 PubMedGoogle ScholarCrossref
13.
Hawks  L, Himmelstein  DU, Woolhandler  S, Bor  DH, Gaffney  A, McCormick  D.  Trends in unmet need for physician and preventive services in the United States, 1998-2017.   JAMA Intern Med. 2020;180(3):439-448. doi:10.1001/jamainternmed.2019.6538 PubMedGoogle ScholarCrossref
14.
Miller  ME, Zuckerman  S.  Comparing urban and rural physicians.   Health Aff (Millwood). 1991;10(4):243-253. doi:10.1377/hlthaff.10.4.243 PubMedGoogle ScholarCrossref
15.
Ingram  DD, Franco  SJ.  NCHS urban-rural classification scheme for counties.   Vital Health Stat 2. 2012;(154):1-65.PubMedGoogle Scholar
16.
Hall  SA, Kaufman  JS, Ricketts  TC.  Defining urban and rural areas in U.S. epidemiologic studies.   J Urban Health. 2006;83(2):162-175. doi:10.1007/s11524-005-9016-3 PubMedGoogle ScholarCrossref
17.
National Academies of Sciences, Engineering, and Medicine; Health and Medicine Division; Board on Population Health and Public Health Practice; Roundtable on the Promotion of Health Equity; Roundtable on Population Health Improvement.  Achieving Rural Health Equity and Well-Being: Proceedings of a Workshop. National Academies Press; 2018.
18.
Billi  JE, Pai  CW, Spahlinger  DA.  The effect of distance to primary care physician on health care utilization and disease burden.   Health Care Manage Rev. 2007;32(1):22-29. doi:10.1097/00004010-200701000-00004 PubMedGoogle ScholarCrossref
19.
Brundisini  F, Giacomini  M, DeJean  D, Vanstone  M, Winsor  S, Smith  A.  Chronic disease patients’ experiences with accessing health care in rural and remote areas: a systematic review and qualitative meta-synthesis.   Ont Health Technol Assess Ser. 2013;13(15):1-33.PubMedGoogle Scholar
20.
Buykx  P, Humphreys  J, Wakerman  J, Pashen  D.  Systematic review of effective retention incentives for health workers in rural and remote areas: towards evidence-based policy.   Aust J Rural Health. 2010;18(3):102-109. doi:10.1111/j.1440-1584.2010.01139.x PubMedGoogle ScholarCrossref
22.
US Census Bureau. Accessed September 30, 2020. https://www.census.gov/programs-surveys/acs/data.html
23.
von Elm  E, Altman  DG, Egger  M, Pocock  SJ, Gøtzsche  PC, Vandenbroucke  JP; STROBE Initiative.  The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies.   J Clin Epidemiol. 2008;61(4):344-349. doi:10.1016/j.jclinepi.2007.11.008 PubMedGoogle ScholarCrossref
25.
US Census Bureau Historical Income Tables. Tables H1 and H3 2019. Accessed September 30, 2020. https://www.census.gov/data/tables/time-series/demo/income-poverty/historical-income-households.html
26.
Clark  RA, Eckert  KA, Stewart  S,  et al.  Rural and urban differentials in primary care management of chronic heart failure: new data from the CASE study.   Med J Aust. 2007;186(9):441-445. doi:10.5694/j.1326-5377.2007.tb00993.x PubMedGoogle ScholarCrossref
27.
Basu  S, Berkowitz  SA, Phillips  RL, Bitton  A, Landon  BE, Phillips  RS.  Association of primary care physician supply with population mortality in the United States, 2005-2015.   JAMA Intern Med. 2019;179(4):506-514. doi:10.1001/jamainternmed.2018.7624 PubMedGoogle ScholarCrossref
28.
Aboagye  JK, Kaiser  HE, Hayanga  AJ.  Rural-urban differences in access to specialist providers of colorectal cancer care in the United States: a physician workforce issue.   JAMA Surg. 2014;149(6):537-543. doi:10.1001/jamasurg.2013.5062 PubMedGoogle ScholarCrossref
29.
Hao  Y, Landrine  H, Jemal  A,  et al.  Race, neighbourhood characteristics and disparities in chemotherapy for colorectal cancer.   J Epidemiol Community Health. 2011;65(3):211-217. doi:10.1136/jech.2009.096008 PubMedGoogle ScholarCrossref
30.
Roetzheim  RG, Pal  N, Gonzalez  EC, Ferrante  JM, Van Durme  DJ, Krischer  JP.  Effects of health insurance and race on colorectal cancer treatments and outcomes.   Am J Public Health. 2000;90(11):1746-1754. doi:10.2105/AJPH.90.11.1746 PubMedGoogle ScholarCrossref
31.
Xue  Y, Smith  JA, Spetz  J.  Primary care nurse practitioners and physicians in low-income and rural areas, 2010-2016.   JAMA. 2019;321(1):102-105. doi:10.1001/jama.2018.17944 PubMedGoogle ScholarCrossref
32.
Odisho  AY, Fradet  V, Cooperberg  MR, Ahmad  AE, Carroll  PR.  Geographic distribution of urologists throughout the United States using a county level approach.   J Urol. 2009;181(2):760-765. doi:10.1016/j.juro.2008.10.034 PubMedGoogle ScholarCrossref
33.
Steinwald  AB, Sloan  FA, Edmunds  M.  Geographic Adjustment in Medicare Payment: Phase II: Implications for Access, Quality, and Efficiency. National Academies Press; 2012.
34.
Hoadley  J, Alker  J, Homes  M. Health insurance coverage in small towns and rural America: the role of Medicaid expansion. Accessed September 30, 2020. https://ccf.georgetown.edu/wp-content/uploads/2018/09/FINALHealthInsuranceCoverage_Rural_2018.pdf
35.
Rajbhandari-Thapa  J, Zhang  D, MacLeod  KE, Thapa  K.  Impact of Medicaid expansion on insurance coverage rates among adult populations with low income and by obesity status.   Obesity (Silver Spring). 2020;28(7):1219-1223. doi:10.1002/oby.22793 PubMedGoogle ScholarCrossref
36.
Dill  MJ, Pankow  S, Erikson  C, Shipman  S.  Survey shows consumers open to a greater role for physician assistants and nurse practitioners.   Health Aff (Millwood). 2013;32(6):1135-1142. doi:10.1377/hlthaff.2012.1150 PubMedGoogle ScholarCrossref
38.
Kruse  CS, Bouffard  S, Dougherty  M, Parro  JS.  Telemedicine use in rural Native American communities in the era of the ACA: a systematic literature review.   J Med Syst. 2016;40(6):145. doi:10.1007/s10916-016-0503-8 PubMedGoogle ScholarCrossref
39.
Zhang  D, Wang  G, Zhu  W,  et al.  Expansion of telestroke services improves quality of care provided in super rural areas.   Health Aff (Millwood). 2018;37(12):2005-2013. doi:10.1377/hlthaff.2018.05089 PubMedGoogle ScholarCrossref
40.
Thompson  MJ, Hagopian  A, Fordyce  M, Hart  LG.  Do international medical graduates (IMGs) “fill the gap” in rural primary care in the United States? a national study.   J Rural Health. 2009;25(2):124-134. doi:10.1111/j.1748-0361.2009.00208.x PubMedGoogle ScholarCrossref
41.
Boulet  JR, Duvivier  RJ, Pinsky  WW.  Prevalence of international medical graduates from Muslim-majority nations in the US physician workforce from 2009 to 2019.   JAMA Netw Open. 2020;3(7):e209418. doi:10.1001/jamanetworkopen.2020.9418 PubMedGoogle Scholar
Limit 200 characters
Limit 25 characters
Conflicts of Interest Disclosure

Identify all potential conflicts of interest that might be relevant to your comment.

Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.

Err on the side of full disclosure.

If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.

Not all submitted comments are published. Please see our commenting policy for details.

Limit 140 characters
Limit 3600 characters or approximately 600 words
    2 Comments for this article
    EXPAND ALL
    Unacknowledged Data Limitations Likely Led to Overestimates of Increases in Primary Care Clinician Supply
    Davis G. Patterson, PhD, and C. Holly A. Andrilla, MS. | University of Washington School of Medicine; WWAMI Rural Health Research Center, Department of Family Medicine, University of Washington, Seattle, WA.
    The goal of this study is laudable, but there are at least three limitations to the NPI data that could significantly affect the findings.

    The first limitation is that we do not know whether NPI information is up to date, particularly in more recent years. The authors estimate fairly large increases in supply of all primary care clinicians in both rural and urban areas from 2009 to 2017. Given the expanding numbers of schools and graduates for all types of primary care clinicians, some increase may be expected. However, an important limitation of the NPI data file is that
    the Centers for Medicare and Medicaid Services, which operates this data system, does not require clinicians to update their NPI information.[1] NPIs were established in 2006, and thus one might expect that the information was reasonably current in 2009, but by 2017, this assumption becomes increasingly suspect. Clinicians could have retired or otherwise become clinically inactive and remain in the data file. This could account for some of the overall increase of clinicians in the NPI data file from 2009 to 2017, resulting in an overestimate of active clinicians in 2017.

    The second limitation is the lack of specialty information on physician assistants (PAs) in the NPI data. Only about one third of PAs practice primary care,[2,3] and the long-term trend among PAs has been away from primary care, much like the physicians they practice with.[3] Without knowing specialty information or appropriately discounting PA numbers in primary care, the authors’ counts of PAs in primary care are undoubtedly overestimates, perhaps three times the actual numbers.

    The third limitation is that the authors appear to have assumed likewise that all nurse practitioners (NPs) practice primary care. One estimate is that 81.6% of NPs practice primary care.[2] The NPI data do contain specialty information for NPs, and our own estimate of NPs in primary care at the WWAMI Rural Health Research Center, using 2015 and 2020 NPI data, is 87%, which may be high because NPs may further specialize over the course of their careers. If the authors did not restrict their NP analysis to primary care specialties, the numbers are likely to be overestimates.

    The inflated numbers of PAs and NPs are particularly concerning because these clinicians account for a large share of the reported total estimated increase in primary care clinician supply from 2009 to 2017. While the NPI data offer advantages for health workforce analysis,[1] it is important to understand the limitations of these data and proceed with caution.

    References

    [1] Bindman AB. Using the National Provider Identifier for health care workforce evaluation. Medicare Medicaid Res Rev. 2013;Jul 30;3(3):mmrr.003.03.b03. doi: 10.5600/mmrr.003.03.b03. eCollection 2013.

    [2] Hooker RS, Brock DM, Cook ML. Characteristics of nurse practitioners and physician assistants in the United States. Journal of the American Association of Nurse Practitioners. 2016;28:39-46. doi:10.1002/2327-6924.12293

    [3] Coplan B, Cawley J, Stoehr J. Physician assistants in primary care: trends and characteristics. Ann Fam Med. 2013;11(1):75-79. doi: https://doi.org/10.1370/afm.1432
    CONFLICT OF INTEREST: None Reported
    READ MORE
    The primary care workforce comprises distinct groups of clinicians
    William Phillips, MD, MPH | Professor Emeritus of Family Medicine, University of Washington, Seattle
    Zhang and colleagues recognize the importance of primary care to the health of individuals and populations in both rural and urban communities. They report an analysis of the primary care workforce that includes physicians, nurse practitioners and physician assistants. Unfortunately, lack of precision in the identification and classification of these clinicians limit the interpretation and policy relevance of their findings.

    The investigators aggregate three distinct groups of physicians together, general practitioners, family physicians and internists. They improperly refer to family physicians, over 80% of whom are board-certified specialists1 as “physicians who practiced family medicine,” but describe internists as
    those who “specialize in internal medicine.” They include general practitioners along with these two groups, even though they are very different in demographic and professional characteristics and most have no specialty training in primary care.1 Even though the study taxonomy provides for the identification of these groups, the authors do not analyze them separately or describe the clinicians in detail sufficient to justify combining the groups. This error is common and its impact upon research findings, conclusions and policy implications has been described.2

    Furthermore, the analysis combines two other groups of clinicians, even more distinct: nurse practitioners and physician assistants. The investigators go on to incorrectly presume that all clinicians in these two groups practice primary care.3 Most physician assistants and many nurse practitioners enter practice in limited specialties and those that enter primary care leave for other practice settings at rates greater than do family physicians. These are independent professional groups and deserve to be treated separately in analysis.

    Clinicians in these distinct groups – even those actively providing primary care – have different demographic and professional characteristics, scope of practice, breadth of services, and patterns of practice location.1 Analysis of workforce trends among these groups is complicated and interpretation requires more clarity and detail. Primary care researchers have emphasized the need to identify clinicians by profession, specialty, and training; to keep groups separate in the analysis; and to carefully justify any aggregation in reporting.2

    1. Phillips WR. Dai M, Frey J, Peterson LE. General practitioners in US medical practice compared to family physicians. Ann Fam Med 2020;18:127-130. https://doi.org/10.1370/afm.2503. https://www.annfammed.org/content/18/2/127

    2. Diep AM, Thoppe HS, Yang A, Agnani AS, Phillips WR. Accuracy of reporting primary care specialty status in medical research. J Am Board Fam Med 2019 November, 32(6): 941-943. doi: 10.3122/jabfm.2019.06.190141. https://www.jabfm.org/content/32/6/94

    3. Patterson DG, Andrilla HA. Unacknowledged data limitations likely led to overestimates of increases in primary care clinician supply. Comment (epublished Nov. 9, 2020) on, Zhang D, Son H, Shen Y, et al. Assessment of changes in rural and urban primary care workforce in the United States from 2009 to 2017. JAMA Network Open. 2020;3(10):e2022914. doi:10.1001/jamanetworkopen.2020.22914
    CONFLICT OF INTEREST: None Reported
    READ MORE
    Views 7,169
    Citations 0
    Original Investigation
    Health Policy
    October 28, 2020

    Assessment of Changes in Rural and Urban Primary Care Workforce in the United States From 2009 to 2017

    Author Affiliations
    • 1Department of Health Policy and Management, College of Public Health, University of Georgia, Athens
    • 2Department of Epidemiology & Biostatistics, College of Public Health, University of Georgia, Athens
    • 3Department of Population Health Science and Policy, Icahn School of Medicine at Mount Sinai, New York, New York
    • 4Department of Obstetrics, Gynecology, and Reproductive Science, Icahn School of Medicine at Mount Sinai, New York, New York
    • 5Department of Geography, University of Georgia, Athens
    • 6School of Medicine and Health Management, Tongji Medical College, Huazhong University of Science and Technology, Wuhan, Hubei, China
    • 7Department of Public Health Policy and Management, School of Global Public Health, New York University, New York
    JAMA Netw Open. 2020;3(10):e2022914. doi:10.1001/jamanetworkopen.2020.22914
    Key Points

    Question  What are the changes in the rural-urban distribution of the primary care workforce in the US from 2009 to 2017?

    Findings  In this cross-sectional study of 3143 US counties (1167 urban and 1976 rural) using county-level data, the density of primary care clinicians increased significantly in both rural and urban counties from 2009 to 2017. The increase in primary care clinician density was more pronounced in urban counties compared with rural counties.

    Meaning  In this study, the density of primary care clinicians increased overall, yet rural-urban disparities in the primary care workforce are increasing in the US.

    Abstract

    Importance  Access to primary care clinicians, including primary care physicians and nonphysician clinicians (nurse practitioners and physician assistants) is necessary to improving population health. However, rural-urban trends in primary care access in the US are not well studied.

    Objective  To assess the rural-urban trends in the primary care workforce from 2009 to 2017 across all counties in the US.

    Design, Setting, and Participants  In this cross-sectional study of US counties, county rural-urban status was defined according to the national rural-urban classification scheme for counties used by the National Center for Health Statistics at the Centers for Disease Control and Prevention. Trends in the county-level distribution of primary care clinicians from 2009 to 2017 were examined. Data were analyzed from November 12, 2019, to February 10, 2020.

    Main Outcomes and Measures  Density of primary care clinicians measured as the number of primary care physicians, nurse practitioners, and physician assistants per 3500 population in each county. The average annual percentage change (APC) of the means of the density of primary care clinicians over time was calculated, and generalized estimating equations were used to adjust for county-level sociodemographic variables obtained from the American Community Survey.

    Results  The study included data from 3143 US counties (1167 [37%] urban and 1976 [63%] rural). The number of primary care clinicians per 3500 people increased significantly in rural counties (2009 median density: 2.04; interquartile range [IQR], 1.43-2.76; and 2017 median density: 2.29; IQR, 1.57-3.23; P < .001) and urban counties (2009 median density: 2.26; IQR. 1.52-3.23; and 2017 median density: 2.66; IQR, 1.72-4.02; P < .001). The APC of the mean density of primary care physicians in rural counties was 1.70% (95% CI, 0.84%-2.57%), nurse practitioners was 8.37% (95% CI, 7.11%-9.63%), and physician assistants was 5.14% (95% CI, 3.91%-6.37%); the APC of the mean density of primary care physicians in urban counties was 2.40% (95% CI, 1.19%-3.61%), nurse practitioners was 8.64% (95% CI, 7.72%-9.55%), and physician assistants was 6.42% (95% CI, 5.34%-7.50%). Results from the generalized estimating equations model showed that the density of primary care clinicians in urban counties increased faster than in rural counties (β = 0.04; 95% CI, 0.03 to 0.05; P < .001).

    Conclusions and Relevance  Although the density of primary care clinicians increased in both rural and urban counties during the 2009-2017 period, the increase was more pronounced in urban than in rural counties. Closing rural-urban gaps in access to primary care clinicians may require increasingly intensive efforts targeting rural areas.

    Introduction

    In recent decades, disparities in population health between rural and urban residents have widened in the US.1 One notable disparity is the higher prevalence of coronary heart disease, stroke, and all-cause mortality among rural vs urban populations.2-5 Primary care workforce supply is a key factor in access to primary care clinicians,6 which is necessary for improving quality of care and population health.7 However, rural-urban disparities occur in access to primary care clinicians.8 Despite efforts to address these disparities, the shortages of rural primary care clinicians have persisted.7,9,10 Lack of an adequate number of rural physicians is a major barrier preventing individuals living in rural and underserved areas from accessing coordinated and integrated care that is essential to healthy and productive lives.11

    In the US health care system, the main primary care workforce includes primary care physicians such as internists, family physicians, and general practitioners, as well as nonphysician clinicians such as nurse practitioners and physician assistants.12 Historically, people living in rural areas of the US have experienced limited access to primary care clinicians compared with those living in urban areas.13,14 More than one-third of rural US residents live in federally designated health professional shortage areas15,16 and approximately 82% of rural counties are classified as medically underserved regions.17 The shortage of primary care physicians in rural areas is associated with longer travel distance to accessing services, which may mitigate the prevention and management of the prevalent chronic diseases.18,19 While this shortage has been described and incentives to attract more physicians to rural areas have been adopted, few analyses have characterized the potential changes in urban-rural physician distributions over the past 10 years.20

    This study examines and compares the primary care workforce and its growth between urban and rural counties or county-equivalents from 2009 to 2017 in the US. Rural clinician shortages may be associated with the widening gap in population health outcomes between rural and urban residents.

    Methods

    Data on all registered primary care clinicians who had an active National Provider Identifier record in the National Plan and Provider Enumeration System from 2009 to 2017 were obtained from the Centers for Medicare & Medicaid Services.21 Data were analyzed from November 12, 2019, to February 10, 2020. We used the National Uniform Claim Committee Health Care Provider Taxonomy code in the National Plan and Provider Enumeration System to identify the main primary care workforce. The workforce included general practitioners, physicians who practiced family medicine and those who specialized in internal medicine only (without other subspecialty), nurse practitioners, and physician assistants.12 All the data on health care clinicians were deidentified and linked with the American Community Survey database collected from the US Census Bureau.22 Per Common Rule, this study was exempt from institutional review board review because all data were deidentified and publicly available and, therefore, did not qualify as human subjects research. We followed the Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) reporting guideline for cross-sectional studies to report our findings.23

    Primary outcome measures included the density of primary care physicians, nurse practitioners, and physician assistants in each county derived from the number of each type of primary care clinician per 3500 residents in each county. This is the definition used by the Health Resources and Services Administration to identify primary care health professional shortage areas.24 According to the Urban-Rural Classification Scheme for Counties defined by the Centers for Disease Control and Prevention National Center for Health Statistics in 2013,15 we classified counties as rural or urban, assuming that their rural or urban status did not change during the study period.15 Counties with an urban-rural score of 1 to 4 were classified as metropolitan (urban [≥250 000 inhabitants]) and those with a score of 5 to 6 were classified as nonmetropolitan (rural [<250 000 inhabitants]).15

    We tested for linear trends in the density of primary care clinicians (primary care physicians, nurse practitioners, and physician assistants) using the Jonckheere-Terpstra test for rural and urban counties. A nonparametric method (ie, the Jonckheere-Terpstra test) was used because the outcome measure (density of primary care clinicians) was not normally distributed over the study period. In addition, we presented the descriptive statistics of the mean densities of the 3 types of clinicians from 2009 to 2017 across rural and urban counties and calculated the average annual percentage change (APC) in the means.

    To estimate the changes in annual rates of the densities of each type of primary care clinicians by rural and urban counties, we fitted generalized estimating equations adjusting for county-level sociodemographic variables. From the generalized estimating equations analysis, we compared the changes in the densities of primary care clinicians between rural and urban counties over time using the Wald χ2 test. County-level sociodemographic variables included the proportion of women, White individuals, Black individuals, Asian individuals, population aged 65 years and over, born outside the US, and population with a 4-year college degree residing in the county. The yearly county-level median household income was categorized according to the published household income quintiles from the Census Bureau Historical Income Tables and was included in the generalized estimating equations analysis.25 A 2-tailed P < .05 was regarded as statistically significant. Data management and analysis were performed with R, version 3.6.2 (R Foundation for Statistical Computing), and SAS, version 9.4 (SAS Institute Inc).

    Results

    The study included yearly data from 3143 US counties (1167 urban and 1976 rural counties). The trend analysis comparing the distribution of primary care clinicians per 3500 people in 2009 and 2017 reported in Figure 1 depicts a geographic distribution in the density of all 3 types of primary care clinicians across the US. The overall density of primary care clinicians has been increasing from 2009 to 2017 (ie, the 2017 maps in the right panel have higher numbers of darker shade counties than the 2009 maps in the left panel).

    Figure 2 shows the trend in the average densities for primary care clinicians in the period from 2009 to 2017. The density of primary care physicians in urban counties increased by 20.8%, from 2.69 in 2009 to 3.25 in 2017. In rural counties, the density increased by 14.3%, from 2.22 in 2009 to 2.54 in 2017. Likewise, the density of nurse practitioners in urban counties increased by 93.6% from 1.65 to 3.20, while in rural counties, it increased by 90.1% from 1.31 to 2.49. The density of physician assistants in urban counties increased by 64.9% from 0.73 to 1.20, and it increased by 49.3% from 0.65 to 0.97 in rural counties.

    Table 1 reports the trend analysis and results from the Jonckheere-Terpstra test. The density of primary care physicians increased substantially from 2009 to 2017 in rural counties (2.04 [interquartile range {IQR}, 1.43-2.76] for 2009 vs 2.29 [IQR, 1.57-3.23] for 2017; P < .001), for nurse practitioners (1.16 [IQR, 0.61-1.80] for 2009 vs 2.24 [IQR, 1.39-3.30] for 2017; P < .001), and for physician assistants (0.43 [IQR, 0.00-0.96] for 2009 vs 0.69 [IQR, 0.16-1.37] for 2017; P < .001). Similar increasing trends were observed in urban counties for primary care physicians (2.26 [IQR, 1.52-3.23] for 2009 vs 2.66 [IQR, 1.72-4.02] for 2017; P < .001), nurse practitioners (1.28 [IQR, 0.72-2.11] for 2009 vs 2.52 [IQR, 1.61-3.99] for 2017; P < .001), and physician assistants (0.56 [IQR, 0.22-0.95] for 2009 vs 0.90 [IQR, 0.42-1.55] for 2017; P < .001).

    Table 2 presents the mean densities of primary care clinicians and the APCs from 2009 to 2017. In urban counties, the APC in the mean density of primary care physicians was 2.40% (95% CI, 1.19%-3.61%), for nurse practitioners was 8.64% (95% CI, 7.72%-9.55%), and for physician assistants was 6.42% (95% CI, 5.34%-7.50%). In rural counties, the APC for the density of primary care physicians was 1.70% (95% CI, 0.84%-2.57%), for nurse practitioners was 8.37% (95% CI, 7.11%-9.63%), and for physician assistants was 5.14% (95% CI, 3.91%-6.37%). The APC in the mean density of primary care clinicians is more prominent in urban than in rural counties.

    Results from the generalized estimating equations analysis adjusting for county-level sociodemographic characteristics are shown in Table 3. The adjusted estimates show that the densities of primary care physicians (β = 0.04; 95% CI, 0.03-0.05; P < .001), nurse practitioners (β = 0.16; 95% CI, 0.15-0.16; P < .001), and physician assistants (β = 0.04; 95% CI, 0.04-0.05; P < .001) have been increasing during the study period, although at faster rates in urban counties. The difference in annual changes of the primary care physician density between urban and rural counties was significant (β = 0.03; 95% CI, 0.02 to 0.04; P < .001). Similarly, the density of nurse practitioners was also increasing faster in urban counties than in rural counties (β = 0.05; 95% CI, 0.03-0.06; P < .001). The same discrepancy between urban and rural areas was observed in changes of the physician assistant density (β = 0.02; 95% CI, 0.01-0.02; P < .001). The density of primary care physicians was positively associated with the median household income categories, the proportion of Asian residents (β = 7.04; 95% CI, 2.37-11.71; P = .003), and residents with a college degree (β = 1.37; 95% CI, 0.04-2.69; P = .04). But there was no significant association between the median household income quintiles and the density of nurse practitioners. Counties with a higher proportion of Black (β = 1.01; 95% CI, 0.33-1.70; P = .004) and Asian residents (β = 4.45; 95% CI, 1.46-7.44; P = .004) were more likely to have a higher density of nurse practitioners. In contrast, an inverse association was observed between the proportion of Black individuals (β = −1.24; 95% CI, –1.63 to –0.85 P < .001) and the density of physician assistants.

    Discussion

    This cross-sectional study examined trends in the density of primary care physicians, nurse practitioners, and physician assistants across rural and urban counties in the US from 2009 to 2017. Overall, there were increasing trends in the density of each type of primary care clinicians across all counties. While there is an overall increase, our study revealed significant geographic disparities in the density of primary care clinicians (primary care physicians, nurse practitioners, and physician assistants) between rural and urban areas. Rural-urban disparities widened during the study period. Moreover, the rural-urban differences continued to be substantial and significant after adjusting for county-level sociodemographic variables.

    Prevalence of chronic disease is 1 of the major contributors to rising health care costs in the US. For rural residents who are at a higher risk of developing coronary heart disease and stroke, greater access to primary care clinicians would improve the coordination of care and health outcomes.2,3,26 Furthermore, the shortage of primary care clinicians may be related to worse health outcomes in US rural counties.7,9,27 Basu et al27 found an association between primary care physician supply and mortality in the US. These outcomes have been analyzed with socioeconomic status, health care coverage, and travel distance to health care facilities, in addition to access to primary care clinicians.13,27-32

    To our knowledge, this study is the first to analyze trends in density of primary care clinicians in rural and urban counties adjusting for sociodemographic factors that may be associated with the characteristics of each county. We found a positive linear correlation between the number of primary care clinicians per 3500 people and median household income quintile categories, and the proportion of some race/ethnicity categories.

    The findings have implications for policy. For example, programs and policies to improve access to and quality of care in rural areas such as geographically based adjustments in Medicare payment,33 or programs that encourage clinical practice in rural areas may be necessary to reduce the gaps. State actions related to the expansion of Medicaid have led to hospital closures in rural areas,34 which might act as a disincentive for physicians to practice in rural areas. The extensive availability of Medicaid coverage provides the needed linkages and payment infrastructure for primary care clinicians to serve patients in need of care but with limited resources.35 In addition, studies have shown that primary care provided by nonphysician clinicians such as nurse practitioners and physician assistants are increasingly accepted by patients.36 However, variation in practice regulations across states and communities exists, and some regions may lag behind in the national trend of growing nonphysician clinicians.37 Furthermore, the expansion of telemedicine has the potential to improve access, use, and outcomes among rural residents, partially mitigating health disparities due to rural clinician shortages.38,39

    Limitations

    This study has 3 key limitations. First, because it is difficult to dissect the complex employment-related decision-making process of individual clinicians, including dual careers, childcare, cultural background and geographic attachment, our analysis is not immune to omitted variable biases. Second, many factors may affect the supply of primary care clinicians. The number of explanatory variables for a county in this study was limited by the availability of existing data sources and, thus, our conclusions may be difficult to generalize. For example, international medical graduates contribute to rural health in some states by helping reduce rural health care workforce gaps.40,41 The National Plan and Provider Enumeration System does not contain individual-level information such as age, gender, race/ethnicity and place of birth, thus, we were unable to assess the association between clinicians born outside the US and the size and change over time of the primary care workforce in rural and urban counties, particularly with regard to health professional shortage areas. Third, we did not use a causal approach to assessing the factors contributing to changes over time of the primary care workforce in our generalized estimating equations analysis given data limitations.

    Conclusions

    While the density of primary care clinicians increased in both rural and urban counties, the pace of growth of primary care clinicians in urban counties has been faster than that in rural counties from 2009 to 2017. Reducing urban-rural disparities in access to primary care clinicians requires policy and intervention efforts that can meaningfully increase the supply of primary care clinicians in rural areas.

    Back to top
    Article Information

    Accepted for Publication: August 25, 2020.

    Published: October 28, 2020. doi:10.1001/jamanetworkopen.2020.22914

    Open Access: This is an open access article distributed under the terms of the CC-BY License. © 2020 Zhang D et al. JAMA Network Open.

    Corresponding Author: Donglan Zhang, PhD, Department of Health Policy and Management, College of Public Health, University of Georgia, 100 Foster Rd, Wright Hall 205D, Athens, GA 30602 (dzhang@uga.edu).

    Author Contributions: Drs Zhang and Son had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis. Dr Zhang and Mr Heejung Son contributed equally as co–first authors.

    Concept and design: Zhang, Chen, Rajbhandari-Thapa, Y. Li, Mu.

    Acquisition, analysis, or interpretation of data: Zhang, Son, Shen, Rajbhandari-Thapa, Y. Li, Eom, Bu, Mu, G. Li, Pagán.

    Drafting of the manuscript: Son, Shen, Chen, Rajbhandari-Thapa, G. Li.

    Critical revision of the manuscript for important intellectual content: Zhang, Son, Shen, Chen, Rajbhandari-Thapa, Y. Li, Eom, Bu, Mu, Pagán.

    Statistical analysis: Son, Shen, Chen, Rajbhandari-Thapa.

    Obtained funding: Zhang, Rajbhandari-Thapa, Y. Li, Mu.

    Administrative, technical, or material support: Zhang, Eom, Mu, Pagán.

    Supervision: Zhang, Shen, G. Li.

    Conflict of Interest Disclosures: None reported.

    Funding/Support: This study was funded by National Institute on Minority Health and Health Disparities grant 1R01MD013886-01 (to Dr Zhang [principal investigator]).

    Role of the Funder/Sponsor: The funders had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

    References
    1.
    Meit  M, Knudson  A, Gilbert  T. The 2014 update of the rural-urban chartbook. Accessed September 30, 2020. https://ruralhealth.und.edu/projects/health-reform-policy-research-center/pdf/2014-rural-urban-chartbook-update.pdf
    2.
    Leira  EC, Hess  DC, Torner  JC, Adams  HP  Jr.  Rural-urban differences in acute stroke management practices: a modifiable disparity.   Arch Neurol. 2008;65(7):887-891. doi:10.1001/archneur.65.7.887 PubMedGoogle ScholarCrossref
    3.
    O’Connor  A, Wellenius  G.  Rural-urban disparities in the prevalence of diabetes and coronary heart disease.   Public Health. 2012;126(10):813-820. doi:10.1016/j.puhe.2012.05.029 PubMedGoogle ScholarCrossref
    4.
    Kulshreshtha  A, Goyal  A, Dabhadkar  K, Veledar  E, Vaccarino  V.  Urban-rural differences in coronary heart disease mortality in the United States: 1999-2009.   Public Health Rep. 2014;129(1):19-29. doi:10.1177/003335491412900105 PubMedGoogle ScholarCrossref
    5.
    Singh  GK, Siahpush  M.  Widening rural-urban disparities in life expectancy, U.S., 1969-2009.   Am J Prev Med. 2014;46(2):e19-e29. doi:10.1016/j.amepre.2013.10.017 PubMedGoogle ScholarCrossref
    6.
    Starfield  B, Shi  L, Macinko  J.  Contribution of primary care to health systems and health.   Milbank Q. 2005;83(3):457-502. doi:10.1111/j.1468-0009.2005.00409.x PubMedGoogle ScholarCrossref
    7.
    Ricketts  TC.  Workforce issues in rural areas: a focus on policy equity.   Am J Public Health. 2005;95(1):42-48. doi:10.2105/AJPH.2004.047597 PubMedGoogle ScholarCrossref
    8.
    Kirby  JB, Yabroff  KR.  Rural-urban differences in access to primary care: beyond the usual source of care provider.   Am J Prev Med. 2020;58(1):89-96. doi:10.1016/j.amepre.2019.08.026 PubMedGoogle ScholarCrossref
    9.
    MacDowell  M, Glasser  M, Fitts  M, Nielsen  K, Hunsaker  M.  A national view of rural health workforce issues in the USA.   Rural Remote Health. 2010;10(3):1531.PubMedGoogle Scholar
    10.
    Rosenblatt  RA, Hart  LG.  Physicians and rural America.   West J Med. 2000;173(5):348-351. doi:10.1136/ewjm.173.5.348 PubMedGoogle ScholarCrossref
    11.
    Bowman  RC.  Measuring primary care: the standard primary care year.   Rural Remote Health. 2008;8(3):1009.PubMedGoogle Scholar
    12.
    Friedberg  MW, Hussey  PS, Schneider  EC.  Primary care: a critical review of the evidence on quality and costs of health care.   Health Aff (Millwood). 2010;29(5):766-772. doi:10.1377/hlthaff.2010.0025 PubMedGoogle ScholarCrossref
    13.
    Hawks  L, Himmelstein  DU, Woolhandler  S, Bor  DH, Gaffney  A, McCormick  D.  Trends in unmet need for physician and preventive services in the United States, 1998-2017.   JAMA Intern Med. 2020;180(3):439-448. doi:10.1001/jamainternmed.2019.6538 PubMedGoogle ScholarCrossref
    14.
    Miller  ME, Zuckerman  S.  Comparing urban and rural physicians.   Health Aff (Millwood). 1991;10(4):243-253. doi:10.1377/hlthaff.10.4.243 PubMedGoogle ScholarCrossref
    15.
    Ingram  DD, Franco  SJ.  NCHS urban-rural classification scheme for counties.   Vital Health Stat 2. 2012;(154):1-65.PubMedGoogle Scholar
    16.
    Hall  SA, Kaufman  JS, Ricketts  TC.  Defining urban and rural areas in U.S. epidemiologic studies.   J Urban Health. 2006;83(2):162-175. doi:10.1007/s11524-005-9016-3 PubMedGoogle ScholarCrossref
    17.
    National Academies of Sciences, Engineering, and Medicine; Health and Medicine Division; Board on Population Health and Public Health Practice; Roundtable on the Promotion of Health Equity; Roundtable on Population Health Improvement.  Achieving Rural Health Equity and Well-Being: Proceedings of a Workshop. National Academies Press; 2018.
    18.
    Billi  JE, Pai  CW, Spahlinger  DA.  The effect of distance to primary care physician on health care utilization and disease burden.   Health Care Manage Rev. 2007;32(1):22-29. doi:10.1097/00004010-200701000-00004 PubMedGoogle ScholarCrossref
    19.
    Brundisini  F, Giacomini  M, DeJean  D, Vanstone  M, Winsor  S, Smith  A.  Chronic disease patients’ experiences with accessing health care in rural and remote areas: a systematic review and qualitative meta-synthesis.   Ont Health Technol Assess Ser. 2013;13(15):1-33.PubMedGoogle Scholar
    20.
    Buykx  P, Humphreys  J, Wakerman  J, Pashen  D.  Systematic review of effective retention incentives for health workers in rural and remote areas: towards evidence-based policy.   Aust J Rural Health. 2010;18(3):102-109. doi:10.1111/j.1440-1584.2010.01139.x PubMedGoogle ScholarCrossref
    22.
    US Census Bureau. Accessed September 30, 2020. https://www.census.gov/programs-surveys/acs/data.html
    23.
    von Elm  E, Altman  DG, Egger  M, Pocock  SJ, Gøtzsche  PC, Vandenbroucke  JP; STROBE Initiative.  The Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) statement: guidelines for reporting observational studies.   J Clin Epidemiol. 2008;61(4):344-349. doi:10.1016/j.jclinepi.2007.11.008 PubMedGoogle ScholarCrossref
    25.
    US Census Bureau Historical Income Tables. Tables H1 and H3 2019. Accessed September 30, 2020. https://www.census.gov/data/tables/time-series/demo/income-poverty/historical-income-households.html
    26.
    Clark  RA, Eckert  KA, Stewart  S,  et al.  Rural and urban differentials in primary care management of chronic heart failure: new data from the CASE study.   Med J Aust. 2007;186(9):441-445. doi:10.5694/j.1326-5377.2007.tb00993.x PubMedGoogle ScholarCrossref
    27.
    Basu  S, Berkowitz  SA, Phillips  RL, Bitton  A, Landon  BE, Phillips  RS.  Association of primary care physician supply with population mortality in the United States, 2005-2015.   JAMA Intern Med. 2019;179(4):506-514. doi:10.1001/jamainternmed.2018.7624 PubMedGoogle ScholarCrossref
    28.
    Aboagye  JK, Kaiser  HE, Hayanga  AJ.  Rural-urban differences in access to specialist providers of colorectal cancer care in the United States: a physician workforce issue.   JAMA Surg. 2014;149(6):537-543. doi:10.1001/jamasurg.2013.5062 PubMedGoogle ScholarCrossref
    29.
    Hao  Y, Landrine  H, Jemal  A,  et al.  Race, neighbourhood characteristics and disparities in chemotherapy for colorectal cancer.   J Epidemiol Community Health. 2011;65(3):211-217. doi:10.1136/jech.2009.096008 PubMedGoogle ScholarCrossref
    30.
    Roetzheim  RG, Pal  N, Gonzalez  EC, Ferrante  JM, Van Durme  DJ, Krischer  JP.  Effects of health insurance and race on colorectal cancer treatments and outcomes.   Am J Public Health. 2000;90(11):1746-1754. doi:10.2105/AJPH.90.11.1746 PubMedGoogle ScholarCrossref
    31.
    Xue  Y, Smith  JA, Spetz  J.  Primary care nurse practitioners and physicians in low-income and rural areas, 2010-2016.   JAMA. 2019;321(1):102-105. doi:10.1001/jama.2018.17944 PubMedGoogle ScholarCrossref
    32.
    Odisho  AY, Fradet  V, Cooperberg  MR, Ahmad  AE, Carroll  PR.  Geographic distribution of urologists throughout the United States using a county level approach.   J Urol. 2009;181(2):760-765. doi:10.1016/j.juro.2008.10.034 PubMedGoogle ScholarCrossref
    33.
    Steinwald  AB, Sloan  FA, Edmunds  M.  Geographic Adjustment in Medicare Payment: Phase II: Implications for Access, Quality, and Efficiency. National Academies Press; 2012.
    34.
    Hoadley  J, Alker  J, Homes  M. Health insurance coverage in small towns and rural America: the role of Medicaid expansion. Accessed September 30, 2020. https://ccf.georgetown.edu/wp-content/uploads/2018/09/FINALHealthInsuranceCoverage_Rural_2018.pdf
    35.
    Rajbhandari-Thapa  J, Zhang  D, MacLeod  KE, Thapa  K.  Impact of Medicaid expansion on insurance coverage rates among adult populations with low income and by obesity status.   Obesity (Silver Spring). 2020;28(7):1219-1223. doi:10.1002/oby.22793 PubMedGoogle ScholarCrossref
    36.
    Dill  MJ, Pankow  S, Erikson  C, Shipman  S.  Survey shows consumers open to a greater role for physician assistants and nurse practitioners.   Health Aff (Millwood). 2013;32(6):1135-1142. doi:10.1377/hlthaff.2012.1150 PubMedGoogle ScholarCrossref
    38.
    Kruse  CS, Bouffard  S, Dougherty  M, Parro  JS.  Telemedicine use in rural Native American communities in the era of the ACA: a systematic literature review.   J Med Syst. 2016;40(6):145. doi:10.1007/s10916-016-0503-8 PubMedGoogle ScholarCrossref
    39.
    Zhang  D, Wang  G, Zhu  W,  et al.  Expansion of telestroke services improves quality of care provided in super rural areas.   Health Aff (Millwood). 2018;37(12):2005-2013. doi:10.1377/hlthaff.2018.05089 PubMedGoogle ScholarCrossref
    40.
    Thompson  MJ, Hagopian  A, Fordyce  M, Hart  LG.  Do international medical graduates (IMGs) “fill the gap” in rural primary care in the United States? a national study.   J Rural Health. 2009;25(2):124-134. doi:10.1111/j.1748-0361.2009.00208.x PubMedGoogle ScholarCrossref
    41.
    Boulet  JR, Duvivier  RJ, Pinsky  WW.  Prevalence of international medical graduates from Muslim-majority nations in the US physician workforce from 2009 to 2019.   JAMA Netw Open. 2020;3(7):e209418. doi:10.1001/jamanetworkopen.2020.9418 PubMedGoogle Scholar
    ×