PURPOSE:
To examine the geographic distribution of physician and pharmacist workforce specialized in oncology in the United States.
METHODS:
Using the National Provider Identifier data, we identified two types of oncology workforce via the healthcare provider taxonomy codes. Oncologists were physicians self-identified as providing oncologic care to patients. Oncology pharmacists were pharmacists with an oncology subspecialty. We calculated the geographic density of physician and pharmacist oncology workforce and used county-level cancer crude rates to quantify the demand for oncology workforce. We used spatial data to plot the density of oncology workforces relative to county-level cancer rates and compared the county-level density of oncologists and oncology pharmacists.
RESULTS:
Of the 30,553 members of the oncology workforce in 2019, 28,681 were oncologists and 1,090 were oncology pharmacists. The mean county-level density of oncologists was 2.94 (SD = 7.32) per 100,000 persons. Sixty-four percent of counties had no oncologists with primary practice location in that county and 12% had no oncologists in the local and adjacent counties. Counties in the top quartile of cancer rates had the highest percentage without any oncologists with primary practice location in the county (75%) and with no oncologists in the local as well as adjacent counties (16%).
CONCLUSION:
Geographically mismatched demand and supply characterized the current oncology workforce. Wide discrepancies in the supply of oncologists across geographic regions highlight the importance of developing core competencies for health professions not specialized in oncology to deliver quality cancer care in areas with unmet need for oncology care.
INTRODUCTION
Health workforce research estimates the availability of a specific health profession practicing in the current healthcare marketplace, projects the changes in the supply of workforce by considering the market entrance from new graduates and exit of retired workforce, and studies the future demand based on trends of demographics and disease epidemiology. Many professional societies, such as the Association of American Medical Colleges, American Academy of Otolaryngology-Head and Neck Surgery, American Medical Association, and American Academy of Pediatrics,1-3 have relied on workforce studies to guide the planning of educational programs (eg, the number of admissions or postgraduate training positions) as well as retention and recruitment strategies.
The ASCO formed a special task force to study the oncology workforce in 2007.4 A report from this task force sounded the alarm of a shortage of oncologists; this 2007 report predicted a 14% increase in the number of oncologists between 2005 and 2020 despite a 48% increase in the demand for oncology services.5 To better monitor the supply of and demand for oncologists in the healthcare marketplace, ASCO created Workforce Information System (WIS).6 A workforce report from the WIS has been included in ASCO’s State of Cancer Care in America series since 2014. More recent analysis continued to project a shortage of oncologists, with the estimated shortage of 2,250 oncologists by 2025.7
The oncology workforce monitored by the ASCO WIS includes physicians specialized in oncology, such as hematologists/oncologists, medical oncologists, radiation oncologists, and pediatric hematology-oncology among others. The complexity of cancer care has motivated healthcare workers in nonphysician health professions, such as pharmacists, nurses, and allied health professions, to subspecialize in oncology so as to better serve patients with cancer. These oncology-focused, nonphysician health professions expand the capacity of the oncology workforce and offer an opportunity to alleviate shortage of oncologists in rural communities.8 To date, only one study explored advanced practice providers (APPs), including nurse practitioners and physician assistants, as expanded oncology workforce.9 No oncology workforce study has examined the geographic codependency between oncologists and nonphysician health professions, such as pharmacists, with an oncology subspecialty. In addition, oncology workforce studies have focused on the number of oncologists within a county but have not considered oncologists available in the adjacent counties.6 This study examined the availability of physicians and oncology pharmacists and their geographic distributions in the United States.
METHODS
Data Sources
We obtained information on oncology workforce from the Full Replacement Monthly National Provider Identifier (NPI) data downloaded from the Centers for Medicare and Medicaid Services (CMS) National Plan and Provider Enumeration (NPPES) Downloadable File.10 The CMS started disclosing NPPES healthcare provider data in September 2007; these data include all healthcare providers with an NPI, which is required for providers to submit HIPAA-compliant claims to all payers.11 For each NPI, the data collect information on sex, entity type (individual v organization), solo proprietor status (yes, no, and unknown), geographic information of the practice location (country, state, postal code, and address), provider enumeration date, and healthcare provider taxonomy codes. The NPI files are updated weekly, and the data used in this study were downloaded on October 1, 2019, covering providers active in practice as of September 8, 2019. When constructing the study cohort, we also used the monthly NPI deactivation file to remove providers no longer in practice. We limited the study cohort to individuals practicing in the 50 states of the United States and excluded those in the US territories.
To understand oncology workforce in the context of demand for oncology services, we obtained the 2019 submission of cancer crude rates at the county level from the Centers for Disease Control and Prevention’s United States Cancer Statistics for counties except for Minnesota and Kansas.12 For Minnesota, we manually extracted the number of cancer cases for each county from the Minnesota Department of Health’s Cancer Query website,13 and calculated county-level cancer crude rates as the total number of cancer cases per 100,000 persons. We went through similar data-generating process for Kansas using the cancer registry query website accessible through University of Kansas Medical Center.14 The total population in each county in 2019 was obtained from the US Census Bureau.15
Density of Oncology Workforce at County Level
We identified oncology workforce from the healthcare provider taxonomy code and categorized the workforce into two groups: Oncologists versus nonphysician health professions subspecialized in oncology (oncology nonphysicians). Oncologists included physicians with a taxonomy code indicating hematology/oncology, medical oncology, gynecologic oncology, pediatric hematology-oncology, radiation oncology, and surgical oncology. Oncology nonphysicians captured nonphysician health professions, such as nurses, pharmacists, or allied health professions, with an oncology subspecialty, and included oncology nurse, pediatric oncology nurse, oncology pharmacist, oncology clinical nurse, and pediatric oncology clinical nurse. The numbers of oncology nonphysicians other than oncology pharmacists total to only 793 individuals (Table 1); as this clearly represents substantial underreporting, we limited our analysis to oncology pharmacists.
TABLE 1.
Characteristics of Study Cohort
Because the NPI data provide zip code but not county information, we derived county code from providers’ address using OpenCage Geocoder.16a Based on the state and county information, we then calculated the total number of oncologists for each county and quantified the density as the number of oncologists per 100,000 persons. The NPI data report one practice location and providers may practice in multiple locations; therefore, we also measured the supply of oncologists by considering the availability of oncologists in adjacent counties as providers practicing in multiple locations are likely to stay within this geographic boundary. We obtained County Adjacency File from the US Census Bureau and linked the file to the counties in the NPI.16
Analysis
We first presented the descriptive statistics of oncologists and oncology pharmacists. Next, we summarized the supply of oncologists at county level by the mean density, the proportion of counties without any oncologists whose primary practice site located in that county, and the proportion of counties without any oncologists in the local as well as adjacent counties. We then categorized counties into quartiles by cancer crude rates, color-coded counties by cancer rate quartiles, and plotted the density of oncologists as dot maps in spatial data on a US map, with larger dots representing higher densities. We created another dot map to mark counties with no oncologists in local as well as adjacent counties. To examine the relationship between the geographic distribution of oncologists and that of oncology pharmacists, we represented both densities as dot maps and overlaid the density of oncology pharmacists on top of the density of oncologists on a US map. These overlapping dot maps offer an appealing visual presentation to comprehend geographic codependency of two health professions.
The clustering of pediatric hematologists/oncologists in specialized pediatric oncology medical centers in selected geographic regions raised a concern that combining oncologists who treat adult patients with cancer and those who treat pediatric patients could potentially skew the geographic distribution of oncologists.17 We added a sensitivity analysis that separated NPIs with self-reported provider taxonomy indicating pediatric hematology/oncology from other oncology NPIs and created the density plot of each group of oncologists. Visual presentation of spatial data was conducted using spmap command in STATA 15.1 (StataCorp LLC, College Station, TX).
RESULTS
Data from the NPPES reported that there were 30,553 healthcare providers in the oncology workforce in 2019, including 28,681 oncologists and 1,872 oncology nonphysicians. Table 1 shows that a larger proportion of oncologists were male (66.0%), not solo proprietors (83.7%), and started their practice before 2011 (87.8%). More than 65% of oncologists were located in four of the nine Census divisions: South Atlantic (19.4%), Middle Atlantic (16.4%), Pacific (15.3%), and East North Central (14.5%). Medical oncology, and hematology and oncology accounted for more than 55% of the oncologists, followed by radiation oncology (20.6%), and pediatric hematology-oncology (10.5%). Table 1 also shows that a larger proportion of oncology nonphysician health professionals identified from provider taxonomy codes were oncology pharmacists (58.2%).
At the county level, the mean density of oncologists was 2.94 per 100,000 (SD = 7.32). Figure 1 shows that 64% of counties in the United States had no oncologists with primary practice location in that county, whereas 12% had no oncologists within the county as well as in the adjacent counties. When stratified by the cancer rates, we found a negative association between the availability of oncology workforce and the cancer rates (corr. = −0.085, P < .01). Figure 1A shows counties in the top quartile of cancer rates included the highest percentage of counties without any oncologists with primary practice location in the local county (75%) or without any oncologists in the local and adjacent counties (16%), compared with counties in the lowest quartile (52% and 11%, respectively). Figure 1B illustrates wide geographic variation in the availability of oncologists, with patients with cancer in the West North Central and West South Central divisions at the highest risk of experiencing shortage of oncologists. More than 80% of counties in these two Census divisions had no oncologists with primary practice location in the local counties. Even after including oncologists available in the adjacent counties, 28% and 17% of counties in West North Central and West South Central divisions had no oncologists within their geographic proximity.
Fig 1.
Availability of oncology workforce at county level (local and adjacent), by county-level cancer rate quartiles, and by census division.
Figure 2A illustrates the density of oncologists as dot map and categorized US counties into four quartiles based on its cancer rate, with darker colors indicating counties with higher cancer rates. It reveals two patterns. First, many counties did not have any oncologists, confirming the information shown in Figure 1A. Second, higher density of oncologists (ie, larger blue circles) tended not to be in counties in the highest quartile of cancer rates (ie, shaded in dark gray). Density plots that separated pediatric from nonpediatric oncologists show that although higher densities of pediatric hematologists/oncologists were found only in a few states, only one state (Wyoming) had no pediatric hematologists/oncologists (Data Supplement, online only). The pattern of geographic distribution of nonpediatric oncologists (Data Supplement) was similar to the pattern observed in all oncologists (Fig 2A). Red dots on the map in Figure 2B mark counties with no oncologists in the local county as well as adjacent counties. It shows the clustering of these counties in the west central region of the United States. Of the 376 counties without oncologists in the local and adjacent counties, 253 (67%) counties were located in West North Central and West South Central Census divisions.
Fig 2.
(A) Density of oncologists and cancer rates (in quartiles) by US counties, (B) US counties without oncologists in the home and adjacent counties and cancer rates (in quartiles), and (C) density of oncologists versus density of oncology pharmacists by US counties.
Figure 2C depicts the spatial relation between oncologists and oncology pharmacists. Oncology pharmacists are largely concentrated in counties with presence of oncologists. Of the 273 counties with one or more oncology pharmacists, only 12 counties (4.4%) had no oncologists. Visual inspection of the density of oncology pharmacists (size of orange dots) and the density of oncologists (ie, size of blue dots) suggested that oncology pharmacists were more available in counties with higher density of oncologists.
DISCUSSION
The current landscape of the oncology workforce indicated that 64% of counties in the United States did not have any physicians specialized in oncology with primary practice site located in that county, and 12% of counties did not have any oncologists in the local county as well as the adjacent counties. Using county cancer rates as proxies of demand for oncology care, we found an alarming pattern that oncologists tended to be less available in counties with higher demand, suggesting that patients with cancer residing in higher-demand counties will either need to travel farther to receive cancer treatment or rely on physicians or other health professions not specialized in oncology to provide the cancer care they need. Indeed, a recent study found that approximately 24% of Medicare beneficiaries traveled 1 hour or more to receive their cancer care.18 We also found wide variations in the availability of oncologists by Census division. Counties in West North Central and West South Central regions of the United States are at highest risk of experiencing shortage of oncologists.
The number of US-based hematologists/oncologists and medical oncologists reported in our study was 16,499 in September 2019, which was higher than the number (12,432 in April 2017) reported in ASCO’s 2018 State of Oncology Practice in American report.8 Although having more recent data partially explained the higher number reported in our study, another reason accounting for the discrepancy was the difference in data sources. Specifically, our study used the NPI data to capture oncologists serving Medicare as well as non-Medicare patients with cancer, whereas the ASCO report was based on data from Medicare Physician Compare; therefore, the ASCO report was limited to physicians who billed Medicare for their services. Although we have excluded deactivated NPIs from our study cohort, the deactivation is not automatic and has to be initiated by the provider. Therefore, oncologists identified from the NPI data could also include those who are no longer in practice, retired or have full-time management or research positions. It is worth noticing that prior to ASCO’s 2018 State of Oncology Practice report, the ASCO report relied on surveys of ASCO members and the AMA Physician Masterfile as the primary data sources for estimation of the supply of oncologists. While the AMA Masterfile is commonly used to estimate the supply of physicians, studies have raised concerns regarding the accuracy of geographic information or the number of specialists reported in the AMA data.19,20 This may explain ASCO’s decision to replace the AMA Masterfile with the Medicare Physician Compare in its latest oncology workforce report.8
Expanding the role of nonphysician health professions, such as pharmacists or APPs, has been suggested as a viable strategy to alleviate shortage in the supply of oncologists, especially for survivorship and palliative care.5,21 We identified 1,090 oncology pharmacists in 2019 from provider taxonomy code, whereas the Board of Pharmacy Specialties and the American Society of Health-System Pharmacists reported 1,863 Board of Pharmacy specialties-certified oncology pharmacists in 2014.22 Therefore, we used oncology pharmacists to demonstrate how to visually inspect whether counties with shortage of oncologists could rely on nonphysician health professionals to fulfill the unmet need for oncology care. We showed that oncology pharmacists were more likely to be present in counties with higher densities of oncologists. If the pattern observed in oncologist pharmacists is true for other types of nonphysician oncology health professionals, expanding the role of these professions may not fully address geographic shortages in oncologists.
Moving forward, to comprehensively examine the geographic distribution of the oncology workforce that include both oncologists and oncology nonphysicians, novel data linkage strategies are necessary. For example, Bruinooge et al9 combined membership database and a proprietary healthcare provider database to estimate the number of APPs who provided oncology care. The authors used the provider taxonomy to identify APPs while relying on other data sources to determine whether these APPs delivered oncology care. Recognizing the need to expand the oncology workforce to include physicians and nonphysicians not specialized in oncology, the Institute of Medicine (now National Academies of Sciences, Engineering, and Medicine [NASEM]) Committee on Improving the Quality of Cancer Care recommended developing and enhancing the core competencies of health professions not specialized in oncology to deliver cancer care through accreditation, certification, and training activities.23 In addition, a recent workshop (titled “Developing and Sustaining an Effective and Resilient Oncology Careforce”) hosted by the National Cancer Policy Forum at the NASEM further broadened the concept of cancer care team from oncology workforce to oncology careforce by including family caregivers.24 Future research should supplement the NPI files with other data sources (as in Bruinooge et al9) to examine the geographic codependency of oncologists and oncology nonphysicians. This will offer insight regarding whether unmet need for oncology care in certain geographic areas could be fulfilled by providers not specialized or subspecialized in oncology.
Several limitations warrant discussions. First, we reported the landscape of oncology workforce at a snapshot in time (ie, September 2019) and were not able to document the trend over time because historical data were not available for download from the NPPES website. Second, we relied on the healthcare provider taxonomy to define and identify oncology workforce, which led to a conservative estimate of the size of oncology workforce because we were not able to fully capture nonphysician health professions who treated patients with cancer. For example, a close examination of the latest version (version 20.1) of healthcare provider taxonomy code indicates that there is no code representing oncology subspecialty for nurse practitioners or physician assistants. Third, the NPI assigns a single location for each provider. While oncology providers may have more than one practice location, the ASCO 2018 State of Oncology Practice in American report (using Physician Compare data) that 72% of hematology/oncology practices have only one location/site.8 We also addressed this limitation by examining counties with no oncologist in the local or adjacent counties. The concern that the location reported in the NPI data may not reflect providers’ change of practice location is lessened by a policy under the NPI Final Rule, which requires all covered entities to update their data within 30 days of any changes.25 Therefore, providers who are active in practice will be incentivized to follow this NPI rule. Finally, while the list of healthcare provider taxonomy was created as a joint effort between the CMS and the Accredited Standards Committee X12N Health Care and maintained by the National Uniform Claim Committee, taxonomy codes in the NPI data are self-selected by providers. Therefore, we cannot rule out the possibility of misclassification bias.
In conclusion, our analysis characterized the current landscape of the oncology workforce as geographically mismatched demand and supply, in that counties with higher demand for oncology care were often those with lower supply of oncologists. We also observed a large discrepancy in the supply of oncologists across geographic regions. One strategy to mitigate the shortage of oncologists is to leverage oncologists available from adjacent counties. Even with this expanded network of oncologists, approximately 12% of counties remained without access to oncologists. Counties with shortage of oncologists may need to rely on other health professionals to fulfill the unmet need for oncology care, especially for patients who cannot travel longer distances due to health or financial reasons. This highlights the importance to develop and standardize core competencies for health professions not specialized in oncology to deliver quality cancer care.
ACKNOWLEDGMENT
Dr Shih acknowledges funding from the National Cancer Institute (R01CA207216, R01CA225647 and CCSG P30 CA016672). Ya-Chen Tina Shih received consulting fees, travel, and accommodations for serving on a grants review panel for Pfizer Inc and an advisory board for AstraZeneca in 2019.
This research was partially supported by the Hubert L. and Olive Stringer Professorship for Cancer Treatment and Research.
DISCLAIMER
The views expressed here are those of the authors and do not represent any official position of the National Cancer Institute or National Institutes of Health.
AUTHOR CONTRIBUTIONS
Conception and design: All authors
Financial support: Ya-Chen Shih
Administrative support: Ya-Chen Shih
Provision of study material or patients: Ya-Chen Shih
Collection and assembly of data: All authors
Data analysis and interpretation: All authors
Manuscript writing: All authors
Final approval of manuscript: All authors
Accountable for all aspects of the work: All authors
AUTHORS’ DISCLOSURES OF POTENTIAL CONFLICTS OF INTEREST
State of Physician and Pharmacist Oncology Workforce in the United States in 2019
The following represents disclosure information provided by authors of this manuscript. All relationships are considered compensated unless otherwise noted. Relationships are self-held unless noted. I =Immediate Family Member, Inst = My Institution. Relationships may not relate to the subject matter of this manuscript. For more information about ASCO’s conflict of interest policy, please refer to www.asco.org/rwc or ascopubs.org/op/authors/author-center.
Open Payments is a public database containing information reported by companies about payments made to US-licensed physicians (Open Payments).
Ya-Chen Tina Shih
Consulting or Advisory Role: Pfizer, AstraZeneca
Research Funding: Novartis (to the institution)
No other potential conflicts of interest were reported.
REFERENCES
- 1.Shapiro E, Cooper CS, Greenfield S, et al. : The American Academy of Pediatrics Workforce Survey for the Section on Urology 2015. J Pediatr Urol 13:68-72, 2017 [DOI] [PubMed] [Google Scholar]
- 2.Kim JS, Cooper RA, Kennedy DW: Otolaryngology-head and neck surgery physician work force issues: An analysis for future specialty planning. Otolaryngol Head Neck Surg 146:196-202, 2012 [DOI] [PubMed] [Google Scholar]
- 3.AAMC : The 2019 Update: The Complexities of Physician Supply and Demand: Projections From 2017 to 2032. Washington, DC, AAMC, 2019 [Google Scholar]
- 4.Hortobagyi GN, American Society of Clinical Oncology : A shortage of oncologists? The American Society of Clinical Oncology workforce study. J Clin Oncol 25:1468-1469, 2007 [DOI] [PubMed] [Google Scholar]
- 5.Erikson C, Salsberg E, Forte G, et al. : Future supply and demand for oncologists: Challenges to assuring access to oncology services. J Oncol Pract 3:79-86, 2007 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.ASCO : Key Trends in Tracking Supply of and Demand for Oncologists. Alexandria, VA, ASCO, 2018. https://www.asco.org/sites/new-www.asco.org/files/content-files/practice-andguidelines/documents/2018-ASCO-Workforce-Information-System.pdf [Google Scholar]
- 7.Yang W, Williams JH, Hogan PF, et al. : Projected supply of and demand for oncologists and radiation oncologists through 2025: An aging, better-insured population will result in shortage. J Oncol Pract 10:39-45, 2014 [DOI] [PubMed] [Google Scholar]
- 8.Kirkwood MK, Hanley A, Bruinooge SS, et al. : The State of Oncology Practice in America, 2018: Results of the ASCO Practice Census Survey. J Oncol Pract 14:e412-e420, 2018 [DOI] [PubMed] [Google Scholar]
- 9.Bruinooge SS, Pickard TA, Vogel W, et al. : Understanding the role of advanced practice providers in oncology in the United States. J Oncol Pract 14:e518-e532, 2018 [DOI] [PubMed] [Google Scholar]
- 10.CMS : Data Dissemination: Announcing Changes to the National Plan and Provider Enumeration System (NPPES) Downloadable File, 2019. https://www.cms.gov/Regulations-andGuidance/Administrative-Simplification/NationalProvIdentStand/DataDissemination. [Google Scholar]
- 11.Miller BF, Petterson S, Burke BT, et al. : Proximity of providers: Colocating behavioral health and primary care and the prospects for an integrated workforce. Am Psychol 69:443-451, 2014 [DOI] [PubMed] [Google Scholar]
- 12.CDC : United States Cancer Statistics Centers for Disease Control and Prevention, 2019. https://www.cdc.gov/cancer/uscs/dataviz/download_data.htm [Google Scholar]
- 13.Cancer Query, MN Department of Health, 2020. https://data.web.health.state.mn.us/cancer_query [Google Scholar]
- 14.Kansas Cancer Registry, University of Kansas Medical Center, 2018. https://www.kumc.edu/kcr/zsearch.aspx [Google Scholar]
- 15.County Population Totals: 2010-2019, United State Census Bureau, 2020. https://www.census.gov/data/datasets/time-series/demo/popest/2010s-counties-total.html [Google Scholar]
- 16.County Adjacency File, United States Census Bureau. https://www.census.gov/geographies/reference-files/2010/geo/county-adjacency.html [Google Scholar]; 16a. https://opencagedata.com/
- 17.Hord J, Shah M, Badawy SM, et al. : The American Society of Pediatric Hematology/Oncology workforce assessment: Part 1-Current state of the workforce. Pediatr Blood Cancer 65:e26780, 2018. 10.1002/pbc.26780 [DOI] [PubMed] [Google Scholar]
- 18.Rocque GB, Williams CP, Miller HD, et al. : Impact of travel time on health care costs and resource use by phase of care for older patients with cancer. J Clin Oncol 37:1935-1945, 2019 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Freed GL, Nahra TA, Wheeler JR, et al. : Counting physicians: Inconsistencies in a commonly used source for workforce analysis. Acad Med 81:847-852, 2006 [DOI] [PubMed] [Google Scholar]
- 20.McLafferty S, Freeman VL, Barrett RE, et al. : Spatial error in geocoding physician location data from the AMA Physician Masterfile: Implications for spatial accessibility analysis. Spat Spatiotemporal Epidemiol 3:31-38, 2012 [DOI] [PubMed] [Google Scholar]
- 21.Lehn JM, Gerkin RD, Kisiel SC, et al. : Pharmacists providing palliative care services: Demonstrating a positive return on investment. J Palliat Med 22:644-648, 2019 [DOI] [PubMed] [Google Scholar]
- 22.Ignoffo R, Knapp K, Barnett M, et al. : Board-certified oncology pharmacists: Their potential contribution to reducing a shortfall in oncology patient visits. J Oncol Pract 12:e359-e368, 2016 [DOI] [PubMed] [Google Scholar]
- 23.IOM : Delivering High-Quality Cancer Care: Charting a New Course for a System in Crisis. Washington, DC, Institute of Medicine, 2013 [PubMed] [Google Scholar]
- 24.Developing and Sustaining an Effective and Resilient Oncology Careforce: Proceedings of a Workshop. Washington, DC, National Academies of Sciences, Engineering, and Medicine, 2019 [PubMed] [Google Scholar]
- 25.HIPAA Administrative Simplification : National Plan and Provider Enumeration System Data Dissemination, 2007. https://www.federalregister.gov/documents/2007/05/30/07-2651/hipaa-administrative-simplification-national-plan-and-provider-enumeration-system-data-dissemination [PubMed] [Google Scholar]