Abstract
Background
This study aims to quantify the extent and diversity of the cancer care workforce, beyond medical oncologists, to inform future demand because the number of cancer survivors is expected to grow in the United States.
Methods
Surveillance, Epidemiology, and End Results-Medicare data were used to evaluate health-care use of cancer survivors diagnosed between 2000 and 2014, enrolled in fee-for-service Medicare Parts A and B, and 65 years or older in 2008-2015. We calculated percentage of cancer survivors who saw each clinician specialty and their average annual number of visits in each phase of care. We projected the national number of individuals receiving care and number of annual visits by clinician specialty and phase of care through 2040.
Results
Cancer survivors had higher care use in the first year after diagnosis and last year of life phases. During the initial year after cancer diagnosis, most survivors were seen for cancer-related care by a medical oncologist (59.1%), primary care provider (55.9%), and/or other cancer-treating physicians (42.2%). The percentage of survivors with cancer-related visits to each specialty declined after the first year after diagnosis, plateauing after year 6-7. However, at 10 or more years after diagnosis, approximately 20% of cancer survivors had visits to medical oncologists and an average of 4 visits a year.
Conclusions
Cancer survivors had higher care use in the first year after diagnosis and last year of life. High levels of care use across specialties in all phases of care have important implications for models of survivorship care coordination and workforce planning.
The number of cancer survivors in the United States is expected to grow 54% from 16.9 million in 2019 (1) to 26.1 million in 2040 (2). Nearly two-thirds of cancer survivors are 65 years or older, and this percentage is expected to increase to 73% by 2040 (2). The substantial increase in survivors has raised concerns about the adequacy of the current health-care workforce to provide high-quality cancer care (3-8).
Two studies commissioned by the American Society of Clinical Oncology projected a growing gap between the care capacity of oncologists and the number of cancer survivors needing care (4,9). There have been other studies exploring the workforce by specialty (10) and of medical (11), surgical (7), and radiation (8,12) oncologists. In response, the National Cancer Policy Forum of the National Academies of Sciences, Engineering, and Medicine (NASEM) convened a workshop in 2009 (13) to discuss strategies to address the projected oncology care shortages, including the increased use of advance practice providers (APPs), for example, nurse practitioners and physician assistants. In 2019, another NASEM workshop explored the oncology care force, including professional and informal caregivers and strategies to improve its efficiency, effectiveness, and resilience (14,15). Although prior studies have examined potential oncology workforce shortages by subspecialty, they have not comprehensively examined the diverse set of clinicians and APPs, including delivery of primary care, who concurrently care for cancer survivors from diagnosis through end of life (EOL) (16).
To address these gaps, we examined cancer survivors’ use of clinician care, specifically from medical, surgical, and radiation oncologists, other cancer-treating physicians (OCTPs), primary care physicians (PCPs), and APPs, from cancer diagnosis through EOL. We also project the number of cancer survivors and the amount of clinician care these survivors will need through 2040. These estimates can be used to inform cancer care workforce demand.
Methods
Data Sources
Surveillance, Epidemiology, and End Results (SEER)-Medicare Claims
We used the SEER cancer registry data (17) linked with Medicare data from the Centers for Medicare and Medicaid (CMS) to estimate clinician care use among cancer survivors aged 65 years and older (18). We used administrative claims for persons enrolled in fee-for-service (FFS) inpatient (Part A) and outpatient (Part B) and not in Medicare Advantage (MA) for clinician care use. We used the National Claims History (NCH) records to identify claims submitted to Medicare by clinicians (eg, physicians and APPs). The NCH claims include each clinician’s unique National Provider Identifier (NPI), which is required for reimbursement of services, and self-reported specialty.
American Medical Association (AMA) Physician Masterfile
The AMA Physician Masterfile includes all physicians who are completing or have completed an accredited residency training program in the United States or have a valid US state license (19). For each physician, the database includes their NPI and their specialty. Specialty is compiled from multiple sources, including credentialing institutions and organizations, and physician self-reporting. We used AMA data current as of July 18, 2016 (19).
National Plan and Provider Enumeration System (NPPES) Registry
The NPPES is the NPI registry maintained by CMS. NPIs are assigned by CMS when a provider enrolls in the NPPES. Self-reported specialty is recorded when NPIs are assigned and is coded using standardized Healthcare Provider Taxonomy Codes (20). NPPES data as of July 11, 2017, were included.
Cancer Survivor Cohorts
We included persons in the SEER-Medicare data who had an incident cancer diagnosed between 2000 and 2014 (diagnosis period). Persons were required to be enrolled in Medicare FFS (Parts A and B) for at least 1 month postdiagnosis and have at least 1 Medicare NCH claim between 2008 and 2015 (observation period). The observation period began in 2008 because NPIs, allowing determination of clinician specialty, were introduced in July 2007. Persons were excluded if their cancer was reported via death certificate or if they survived less than 1 month after cancer diagnosis. Persons diagnosed with cancer before 2008 and their available claims were included to estimate use by “long-term” survivors. Included persons (“cancer survivors”) were required to be 65 years and older during the observation period and were grouped into cohorts based on phases of care, as described below. These cohorts were not mutually exclusive. We used the SEER race and ethnicity recodes Hispanic, non-Hispanic American Indian and Alaska Native, non-Hispanic Asian Pacific Islander, non-Hispanic Black, non-Hispanic White, and unknown race or ethnicity.
Clinician Specialty Classification
Clinicians were included if the clinician’s NPI was listed in at least 1 survivor’s claim. A 3-step process classified clinician specialty (Figure 1). First, all NPIs on the included SEER-Medicare NCH claims from 2008 to 2015 were compiled and matched against the AMA data, considered to be gold-standard physician specialty data (21). Second, NPIs not found in the AMA data and NPIs in the AMA data without a listed specialty were compared with the NPPES to identify additional physicians (eg, NPI associated with a primary Taxonomy Code beginning with “20”). Third, for all physician NPIs without an AMA specialty and for all nonphysician NPIs, specialty was determined based on self-reported information on the NCH claims.
Figure 1.
Three-step process used to classify clinician specialty. First, all unique National Provider Identifiers (NPI) included in Surveillance, Epidemiology, and End Results (SEER)-Medicare National Claims History (NCH) are matched against American Medical Association (AMA) Physician data. Second, NPIs not found in the AMA data, and NPIs in the AMA data without a listed specialty were compared with the National Plan and Provider Enumeration System (NPPES) to identify additional physician specialty. Third, for all physician NPIs without an AMA specialty and for all nonphysician NPIs, specialty was determined based on self-reported information on the NCH claims. aDermatologists, endocrinologists, gynecologists, and urologists.
Physician specialty was categorized using a hierarchical scheme. First, if any type of oncology specialty, including hematology, was found in the 3-step process described above, the physician was assigned to the oncology category. Among physicians classified as oncologists, subspecialty was also assigned using a hierarchical classification scheme of radiation oncologist, surgical oncologist, and medical oncologist. Recognizing that dermatologists, endocrinologists, gynecologists, and urologists (classified as other cancer treating physicians) and “nononcology surgeons” (eg, general, thoracic, oral and maxillofacial, and neurologic) often provide cancer care, these physicians were included in our assessment. Finally, PCPs were identified among the remaining unclassified physicians. NPIs not identified as being associated with a physician specialty, as described above, were included in the analyses if they were associated with a nurse practitioner or a physician. See Supplementary Table 1 (available online) for the list of clinician specialty codes considered.
Cancer-Related Visit for Nononcology Physicians
We defined cancer-related care for the nononcology group of clinicians (nononcology surgeons, PCPs, and APPs) if the line-item diagnosis code, associated with the listed procedure code on the claim, was consistent with malignant cancer (listed in Supplementary Table 2, available online) (22,23). The line-item diagnosis codes represent the reason for which the billed care or services were provided.
Clinician Care Use by Phase of Care
We evaluated clinician care use among cancer survivors during different phases of care defined by years since cancer diagnosis and, for those who died, we defined the EOL as the 12 months before death. Based on SEER cause of death information, we further categorized the EOL phase into cancer death and noncancer death. For persons who died during the observation period, months were first assigned to the EOL phase (up to 12 months before death), and any remaining months were then assigned to the applicable years since cancer diagnosis phase. EOL months were excluded for persons with unknown cause of death. Months of observation were censored at the diagnosis of a subsequent tumor, age 99 years, date of death, end of Medicare Parts A and B enrollment, or end of observation (December 31, 2015), whichever occurred first. For example, a person diagnosed on January 1, 2005, who died of cancer on December 31, 2010, would contribute claims in 2008, 2009 and 2010 to the use estimates for the fourth and fifth year post-diagnosis phases and EOL cancer phase, respectively.
Visits to surgical oncologists and cancer visits to nononcology surgeons were combined to best capture cancer surgical care.
We calculated a weighted percentage of cancer survivors who saw each clinician specialty during each phase of care. The denominator included cancer survivors who were enrolled in Medicare FFS (Parts A and B) during the phase of care, and the numerator included persons who had at least 1 claim from the specified clinician type during the phase of care. The weights were the inverse of the length phase of care enrollment. Then, among cancer survivors who visited the specified clinician type (numerator), the average monthly number of visits per specified clinician and phases of care was calculated and multiplied by 12 to allow reporting of average annual number of visits. Also, for survivors who visited medical oncologists, we estimated the proportion who died of cancer in the next phase of care as a proxy for patients in need of more intensive care.
National Health-Care Use
We calculated US (national) cancer prevalence projections for 2020, 2030, and 2040 by phase of care, sex, and age group (0-64 years, 65-74 years, 75-84 years, 85+ years) and cancer type (24). To project national clinician care use, we multiplied the stratified cancer prevalence projections by the respective proportions of cancer survivors obtaining care from each specialty (Supplementary Table 3, available online) and by the average annual number of visits per specialty (Supplementary Table 4, available online). We did summations across the strata products to obtain national projections. We assumed that the use among cancer survivors younger than 65 years was the same as the youngest cancer survivors for which data were available, that is, the younger cancer survivors aged 65-74 years.
Results
Cancer Survivors Characteristics
Characteristics of the cancer survivors who were included in the first and sixth year after diagnosis and EOL cancer cohorts are displayed in Table 1. Cancer survivors were predominantly male (>51%) and non-Hispanic White (>80%) (Table 1). Cancer survivors included in the first year after diagnosis cohort were younger at observation than those in the other cohorts. The most common cancer diagnoses for the first and sixth year after diagnosis cohorts were prostate, breast, melanoma, colorectal, lung, and bladder cancer. Among the 442 140 EOL cancer cohort, lung cancer was the most common cancer type and the majority, 52.1%, died in the 0-1 year from diagnosis (Supplementary Table 5, available online).
Table 1.
Characteristics of cancer survivors in the first year and fifth year after diagnosis and end-of-life cancer death phases of care for SEER-Medicare patients diagnosed with cancer between 2000 and 2014 and 65 years or older at the observation period (2008-2015).
| Characteristic | Phase of carea |
||
|---|---|---|---|
| Time since cancer diagnosis |
End-of-life cancer deathb | ||
| 0 to < 1 y | 5 to <6 y | ||
| Total No. | 793 741 | 601 508 | 442 140 |
| Age at diagnosis, % | |||
| <65 y | 4.4 | 27.1 | 6.9 |
| 65-69 y | 26.8 | 24.2 | 18.4 |
| 70-74 y | 23.8 | 20.1 | 19.1 |
| 75-79 y | 19.6 | 15.6 | 19.5 |
| ≥80 y | 25.4 | 13.0 | 36.1 |
| Age at observation, % | |||
| 65-69 y | 31.3 | 27.3 | 20.6 |
| 70-74 y | 23.8 | 24.1 | 19.0 |
| 75-79 y | 19.6 | 20.1 | 19.7 |
| ≥80 y | 25.3 | 28.5 | 40.8 |
| Sex, % | |||
| Male | 52.4 | 52.8 | 51.1 |
| Female | 47.6 | 47.2 | 48.9 |
| Race and ethnicity, % | |||
| Hispanic | 5.30 | 5.20 | 5.60 |
| Non-Hispanic American Indian and Alaska Native | 0.40 | 0.40 | 0.50 |
| Non-Hispanic Asian Pacific Islander | 4.10 | 4.10 | 4.60 |
| Non-Hispanic Black | 7.10 | 7.10 | 8.60 |
| Non-Hispanic White | 81.70 | 82.10 | 80.50 |
| Unknown race or ethnicity | 1.40 | 1.20 | 0.10 |
| Cancer site, % | |||
| Bladder | 6.3 | 5.3 | 3.6 |
| Female breast | 17.1 | 21.4 | 5.9 |
| Colorectal | 9.5 | 10.1 | 9.8 |
| Lung | 9.2 | 3.7 | 28.1 |
| Melanoma | 9.8 | 8.2 | 1.9 |
| Prostate | 17.5 | 27.2 | 5.2 |
The cohorts are not mutually exclusive. SEER = Surveillance Epidemiology, and End Results.
Defined as the 12 months preceding death.
Percentage of Cancer Survivors Receiving Care and Number of Visits
During the first year after cancer diagnosis, most cancer survivors had at least 1 cancer-related visit with a medical oncologist (59.1%), surgeon (48.7%), OCPT (42.1%), and a PCP (55.9%), and smaller percentages of survivors had cancer-related visits with radiation oncologists (35.5%) and APPs (14.7%) (Table 2; Supplementary Figure 1, available online). Following the initial year after diagnosis (and presumed initial treatment), the percentage of survivors with cancer-related visits to each specialty declined over the next 6-7 years and were relatively stable between 7 years and 15 years postcancer diagnosis. Approximately 20% of survivors had visits with medical oncologists 15 years post cancer. We estimated that 11.6%, 3.8%, and 2.8% of cancer survivors who visited medical oncologists in 0-1 years, 5-6 years, and 10 or more years since diagnosis died of cancer in the following year (Supplementary Table 6, available online). During the EOL cancer phase, the percentage of survivors having cancer-related visits was substantially higher, especially with medical oncologists (86.2%), PCPs (84.7%), and APPs (30.6%). The percentage of survivors with any visit to a PCP or APP was fairly stable for each cohort from time of cancer diagnosis until the EOL, when care was higher irrespective of the cause of death. In the first year after diagnosis, women more frequently visited medical oncologists (73.9% vs 45.3%) and surgeons (61.0% vs 37.3%) and less frequently visited OCTP (24.8% vs 57.6%) than men (Supplementary Table 3, available online). The percentage of survivors with a cancer visit or any visit to a PCP or APP was slightly higher for women.
Table 2.
Percent of SEER-Medicare cancer survivors with visits to select clinicians by phase of care
| Phase of carea | No. of cancer survivorsb | Cancer visit (oncology specialty) |
Cancer visit (cancer diagnosis code)c |
Any visit |
|||||
|---|---|---|---|---|---|---|---|---|---|
| Medical oncologists, % | Radiation oncologists, % | Surgeonsd, % | Other cancer treating physicianse, % | PCPs, % | APPs, % | PCPs, % | APPs, % | ||
| 0 to <1 y | 793 741 | 59.1 | 35.5 | 48.7 | 42.2 | 55.9 | 14.7 | 92.4 | 36.6 |
| 1 to <2 y | 783 695 | 46.1 | 18.6 | 19.2 | 29.1 | 32.3 | 6.6 | 88.2 | 29.9 |
| 2 to <3 y | 722 818 | 41.6 | 13.4 | 14.6 | 26.7 | 27.3 | 5.3 | 87.6 | 29.0 |
| 3 to <4 y | 677 104 | 38.5 | 10.6 | 12.3 | 24.9 | 24.6 | 4.5 | 87.4 | 28.5 |
| 4 to <5 y | 636 848 | 36.0 | 8.7 | 10.6 | 23.5 | 22.7 | 4.1 | 87.2 | 28.2 |
| 5 to <6 y | 601 508 | 33.1 | 7.2 | 9.2 | 22.2 | 21.1 | 3.5 | 87.1 | 27.7 |
| 6 to <7 y | 568 996 | 29.1 | 5.6 | 7.9 | 21.0 | 19.6 | 2.9 | 87.1 | 27.2 |
| 7 to <8 y | 535 314 | 26.8 | 4.8 | 7.3 | 20.1 | 18.6 | 2.7 | 87.0 | 27.1 |
| 8 to <9 y | 453 887 | 25.0 | 4.2 | 6.6 | 18.9 | 17.6 | 2.5 | 87.0 | 27.3 |
| 9 to <10 y | 373 020 | 23.7 | 3.7 | 6.0 | 17.6 | 16.8 | 2.4 | 87.0 | 28.4 |
| 10 to <11 y | 301 356 | 22.4 | 3.2 | 5.4 | 16.3 | 16.2 | 2.3 | 86.9 | 29.3 |
| 11 to <12 y | 237 952 | 21.0 | 2.8 | 4.7 | 14.7 | 15.6 | 2.2 | 86.7 | 30.5 |
| 12 to <13 y | 179 877 | 20.2 | 2.6 | 4.2 | 13.4 | 15.0 | 2.1 | 86.6 | 31.8 |
| 13 to <14 y | 127 836 | 19.5 | 2.4 | 4.1 | 12.5 | 14.4 | 2.1 | 86.7 | 33.5 |
| 14 to <15 y | 78 553 | 19.2 | 2.2 | 3.8 | 11.3 | 13.6 | 2.0 | 86.4 | 35.2 |
| 15 to <16 y | 35 477 | 20.5 | 2.3 | 3.2 | 8.9 | 10.7 | 1.5 | 86.3 | 38.1 |
| EOL cancer death | 442 140 | 86.2 | 44.1 | 45.4 | 23.3 | 84.7 | 30.6 | 96.2 | 54.6 |
| EOL noncancer death | 302 828 | 40.2 | 10.4 | 15.1 | 18.7 | 35.1 | 7.7 | 96.5 | 52.8 |
Years since cancer diagnosis or EOL: defined as the 12 months preceding death and classified by cause of death. APPs = advance practice providers (nurse practitioners and physician assistants); EOL = end of life; PCPs = primary care physicians; SEER = Surveillance, Epidemiology, and End Results.
Cancer survivors are those diagnosed with any cancer between 2000 and 2014 and age 65 years or older at observation in 2008-2015.
Visit included only if the associated claim listed a cancer diagnosis code as the reason for the service provided.
Includes all surgical oncologist visits and nononcology surgeons visits only if the associated claim listed a cancer diagnosis code as the reason for the service provided.
Dermatologists, endocrinologists, gynecologists, and urologists.
During the initial year after cancer diagnosis, the average number of cancer-related visits per survivor who visited a clinician varied by clinician specialty (radiation oncologists = 13.1, medical oncologists = 11.7, surgeons = 3.4, OCTP = 4.5, PCPs = 3.9, and APPs = 2.5) (Table 3; Supplementary Figure 2, available online). By 5 years post diagnosis, cancer-related visits were lower at approximately 2 annual visits per clinician, except for medical and radiation oncologists, with an annual average of 4 and 3 visits, respectively. Visits to all clinician types were dramatically higher during the EOL phase, particularly to medical oncologists, radiation oncologists, and PCPs, and highest among those who died of cancer. The mean number of cancer-related visits to oncologists and OCTP was slightly higher for men compared with women (Supplementary Table 4, available online). In general, the mean number of cancer-related visits decreased with age but not any visits to PCPs or APPs.
Table 3.
Number of average annual visits per cancer survivor who visited select clinicians by phase of carea
| Phase of careb | No. of cancer survivors | Cancer visit (oncology specialty) |
Cancer visit (cancer diagnosis code)c |
Any visit |
|||||
|---|---|---|---|---|---|---|---|---|---|
| Medical oncologists | Radiation oncologists | Surgeonsd | Other cancer treating physicianse | PCPs | APPs | PCPs | APPs | ||
| 0 to <1 y | 793 741 | 11.7 | 13.1 | 3.4 | 4.5 | 3.9 | 2.5 | 10.5 | 3.2 |
| 1 to <2 y | 783 695 | 6.9 | 3.6 | 2.3 | 3.1 | 2.7 | 2.2 | 8.1 | 2.9 |
| 2 to <3 y | 722 818 | 5.9 | 3.0 | 2.1 | 2.7 | 2.5 | 2.0 | 7.9 | 2.9 |
| 3 to <4 y | 677 104 | 5.3 | 2.9 | 2.0 | 2.5 | 2.4 | 1.9 | 7.8 | 2.9 |
| 4 to <5 y | 636 848 | 4.9 | 2.8 | 1.9 | 2.4 | 2.3 | 1.9 | 7.8 | 2.8 |
| 5 to <6 y | 601 508 | 4.6 | 2.8 | 1.9 | 2.4 | 2.3 | 1.8 | 7.7 | 2.9 |
| 6 to <7 y | 568 996 | 4.5 | 2.9 | 1.9 | 2.4 | 2.2 | 1.9 | 7.7 | 2.9 |
| 7 to <8 y | 535 314 | 4.4 | 2.9 | 1.9 | 2.3 | 2.2 | 1.9 | 7.8 | 2.9 |
| 8 to <9 y | 453 887 | 4.3 | 3.1 | 1.8 | 2.3 | 2.1 | 1.8 | 7.7 | 2.9 |
| 9 to <10 y | 373 020 | 4.2 | 3.1 | 1.8 | 2.3 | 2.1 | 1.8 | 7.7 | 2.9 |
| 10 to <11 y | 301 356 | 4.1 | 3.5 | 1.8 | 2.3 | 2.1 | 1.8 | 7.6 | 2.9 |
| 11 to <12 y | 237 952 | 4.0 | 3.6 | 1.8 | 2.2 | 2.1 | 1.8 | 7.5 | 3.0 |
| 12 to <13 y | 179 877 | 3.9 | 3.9 | 1.8 | 2.2 | 2.0 | 1.9 | 7.4 | 3.1 |
| 13 to <14 y | 127 836 | 3.9 | 4.1 | 1.8 | 2.2 | 2.0 | 2.0 | 7.3 | 3.1 |
| 14 to <15 y | 78 553 | 3.8 | 5.0 | 1.9 | 2.2 | 2.0 | 2.0 | 7.3 | 3.3 |
| 15 to <16 y | 35 477 | 4.1 | 5.7 | 2.1 | 2.1 | 2.1 | 2.0 | 7.6 | 3.7 |
| EOL cancer death | 442 140 | 21.1 | 10.2 | 3.7 | 4.4 | 9.2 | 4.0 | 22.1 | 5.3 |
| EOL noncancer death | 302 828 | 10.9 | 7.7 | 3.1 | 3.2 | 5.4 | 3.0 | 23.7 | 5.8 |
Cancer survivors are those diagnosed with any cancer between 2000 and 2014 and age 65 years or older at observation in 2008-2015. APPs = advance practice providers (nurse practitioners and physician assistants); EOL = end of life; PCPs = primary care physicians; SEER = Surveillance, Epidemiology, and End Results.
Years since cancer diagnosis or EOL: defined as the 12 months preceding death and classified by cause of death.
Visit included only if the associated claim listed a cancer diagnosis code as the reason for the service provided.
Includes all surgical oncologist visits and nononcology surgeons visits only if the associated claim listed a cancer diagnosis code as the reason for the service provided.
Dermatologists, endocrinologists, gynecologists, and urologists.
Clinical Care Use by Cancer Type
Clinician care use varied by cancer site and reflected different treatment modalities (Supplementary Table 7, available online). For example, the percent of cancer survivors obtaining care from medical oncologists was higher among persons diagnosed with digestive system, female breast and gynecologic, hematologic, brain, and lung cancers.
Table 4 provides projections of clinician visits by specialty based on current treatment patterns. For example, the number of survivors cared for by medical oncologists is estimated to be 5.60 million in 2020 and projected to increase to 6.93 million in 2030 and 7.93 million in 2040, representing a 24% and 42% increase, respectively, compared with 2020. When stratified by phase of care, the greatest projected increases were estimated to be among longer-term survivors of 10 or more years, with an increase of approximately 35% in medical oncologist visits and number of visits from 2020 through 2030. Use was projected to increase at a lower rate from 2030 through 2040 because of a slowing in aging of the US population.
Table 4.
Projected number of cancer survivors in the United States and, for selected clinicians, the associated number of patients for whom cancer care will be needed and the number of annual visitsa, overall and by phase of care for 2020, 2030, and 2040b
| Phase of care | No. of cancer survivors | Increase from 2020, % | Medical oncologists |
Radiation oncologists |
Surgeonsc |
Other cancer treating physiciansd,e |
PCPse |
APPse |
||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| No. of patients | Annual visits | No. of patients | Annual visits | No. of patients | Annual visits | No. of patients | Annual visits | No. of patients | Annual visits | No. of patients | Annual visits | |||
| 2020 | ||||||||||||||
| Total | 17.39 | 5.60 | 38.75 | 1.61 | 11.35 | 2.01 | 5.24 | 3.21 | 9.03 | 4.00 | 12.91 | 0.81 | 1.96 | |
| 0 to <1 y | 1.22 | 0.72 | 8.66 | 0.45 | 6.01 | 0.59 | 2.03 | 0.51 | 2.32 | 0.70 | 2.72 | 0.19 | 0.46 | |
| 1 to <5 y | 4.14 | 1.71 | 9.91 | 0.57 | 1.76 | 0.61 | 1.26 | 1.07 | 2.86 | 1.16 | 2.88 | 0.23 | 0.46 | |
| 5 to <10 y | 3.75 | 1.09 | 4.72 | 0.20 | 0.60 | 0.29 | 0.53 | 0.73 | 1.66 | 0.73 | 1.58 | 0.11 | 0.21 | |
| 10+ y | 7.52 | 1.59 | 6.38 | 0.18 | 0.85 | 0.28 | 0.56 | 0.74 | 1.58 | 0.94 | 1.93 | 0.13 | 0.26 | |
| EOL cancer death | 0.36 | 0.32 | 7.00 | 0.17 | 1.72 | 0.17 | 0.64 | 0.08 | 0.36 | 0.31 | 2.92 | 0.12 | 0.47 | |
| EOL noncancer death | 0.41 | 0.18 | 2.08 | 0.05 | 0.40 | 0.07 | 0.22 | 0.08 | 0.25 | 0.15 | 0.88 | 0.04 | 0.11 | |
| 2030 | ||||||||||||||
| Total | 22.17 | 27.5 | 6.93 | 47.66 | 1.96 | 13.77 | 2.45 | 6.39 | 4.07 | 11.44 | 4.98 | 16.08 | 0.99 | 2.40 |
| 0 to <1 y | 1.46 | 19.9 | 0.85 | 10.23 | 0.53 | 7.13 | 0.70 | 2.41 | 0.62 | 2.83 | 0.83 | 3.26 | 0.22 | 0.54 |
| 1 to <5 y | 4.91 | 18.7 | 2.01 | 11.70 | 0.67 | 2.07 | 0.72 | 1.49 | 1.29 | 3.47 | 1.38 | 3.44 | 0.27 | 0.54 |
| 5 to <10 y | 4.55 | 21.5 | 1.30 | 5.69 | 0.25 | 0.73 | 0.35 | 0.64 | 0.91 | 2.10 | 0.89 | 1.93 | 0.13 | 0.25 |
| 10+ y | 10.26 | 36.4 | 2.14 | 8.65 | 0.24 | 1.18 | 0.38 | 0.76 | 1.04 | 2.23 | 1.28 | 2.62 | 0.18 | 0.34 |
| EOL cancer death | 0.47 | 28.9 | 0.41 | 8.84 | 0.21 | 2.18 | 0.22 | 0.81 | 0.11 | 0.47 | 0.40 | 3.75 | 0.15 | 0.59 |
| EOL noncancer death | 0.53 | 28.2 | 0.22 | 2.56 | 0.06 | 0.48 | 0.09 | 0.27 | 0.10 | 0.33 | 0.19 | 1.09 | 0.04 | 0.14 |
| 2040 | ||||||||||||||
| Total | 26.07 | 49.9 | 7.93 | 53.83 | 2.19 | 15.27 | 2.78 | 7.22 | 4.74 | 13.30 | 5.72 | 18.47 | 1.12 | 2.69 |
| 0 to <1 y | 1.64 | 34.5 | 0.95 | 11.31 | 0.59 | 7.80 | 0.79 | 2.69 | 0.70 | 3.19 | 0.94 | 3.68 | 0.24 | 0.60 |
| 1 to <5 y | 5.46 | 31.9 | 2.22 | 12.95 | 0.73 | 2.25 | 0.79 | 1.65 | 1.43 | 3.90 | 1.53 | 3.84 | 0.29 | 0.59 |
| 5 to <10 y | 5.13 | 37.1 | 1.45 | 6.37 | 0.27 | 0.81 | 0.39 | 0.73 | 1.04 | 2.43 | 1.01 | 2.19 | 0.15 | 0.28 |
| 10+ y | 12.64 | 68.1 | 2.57 | 10.50 | 0.30 | 1.43 | 0.46 | 0.93 | 1.31 | 2.81 | 1.55 | 3.21 | 0.22 | 0.41 |
| EOL cancer deathf | 0.54 | 49.2 | 0.47 | 9.76 | 0.24 | 2.43 | 0.25 | 0.92 | 0.13 | 0.56 | 0.46 | 4.27 | 0.17 | 0.66 |
| EOL noncancer deathf | 0.65 | 59.4 | 0.27 | 2.95 | 0.07 | 0.54 | 0.10 | 0.32 | 0.13 | 0.42 | 0.24 | 1.28 | 0.05 | 0.16 |
Visit included only if the associated claim listed a cancer diagnosis code as the reason for the service provided. APPs = advance practice providers (nurse practitioners and physician assistants); EOL = end of life; PCPs = primary care physicians; SEER = Surveillance, Epidemiology, and End Results.
Prevalence estimates are from SEER data and projected to the US population. All projected values for the numbers of patients and numbers of annual visits are in the millions.
Includes all surgical oncologists regardless of diagnosis code and nononcology surgeons if the claim listed a cancer diagnosis code.
Dermatologists, endocrinologists, gynecologists, and urologists.
Cancer-related visits.
Years since cancer diagnosis or EOL: defined as the 12 months preceding death and classified by cause of death.
Discussion
This study provides comprehensive estimates of the range and type of clinicians concurrently caring for cancer survivors throughout the cancer continuum and projections for future workforce needs. Cancer survivors had higher care use in the first year after diagnosis and EOL cancer death phase. We project that compared with 2020 levels, the total number of survivors receiving care by medical oncologists is expected to increase 24% by 2030 and 42% by 2040. The highest increase in use is projected to occur among persons who survive 10 or more years following diagnosis between 2020 and 2030 and has important implications for models of survivorship care coordination and workforce planning.
During the initial year after cancer diagnosis, most survivors were seen by either a medical oncologist or OCTPs. Cancer-related visits diminish over time for all specialties, except for medical oncologists, but remain high for primary care any visit. The percentage of survivors visiting medical oncologists was estimated to be 20% at 10 or more years after diagnosis. The results highlight the need for more research on the reasons long-term survivors continue to see medical oncologists, to better coordinate care between the oncology and primary care workforce, and to develop more efficient care strategies (25–27). Our estimates also demonstrated many visits across clinicians in any given phase of care, which raises questions about care coordination, communication, and the need for all these visits. PCPs and APPs consistently saw cancer survivors through all phases of care for cancer and noncancer care, which provides an opportunity for more efficient coordination strategies for gradual transition to primary care alone for those with no evidence of disease. These findings have implications for an already challenged primary care workload and the need to develop more in-depth expertise in long-term and late effects of cancer survivors, including the risk for late recurrences or new cancers (28,29).
Our study updates and extends prior studies (10,11,30,31) in important ways. We included a broad range of providers and specialties to estimate their concurrent use, considered cancer-related and any visit for APPs and PCPs, and evaluated a larger number of cancer sites. We also used a more detailed phases of care approach that include each year from diagnosis and EOL phases, providing insight into care use across providers in the cancer care continuum that can be used to evaluate and coordinate care across clinicians. Compared with earlier research that included SEER-Medicare data from 1998 to 2003 (11), our first year after diagnosis estimates (initial care) were slightly higher for the percent of survivors who saw a medical oncologist (59% vs 47%) and the national estimates of survivors seeing medical oncologists (0.72 M vs 0.66 M). These differences are likely due to more recent data and updated methods. For example, this study includes survivors diagnosed with multiple cancers who were excluded in the previous study. Additionally, the algorithm to identify physician specialty is improved. Coombs et al. (30,31) found that approximately 10% of cancer survivors received care from APPs. We estimated that 15% and 56% of survivors received cancer care from APPs and PCPs, respectively, in the first year after diagnosis, decreasing to approximately to 5% and 20% at 5 years after diagnosis.
Our study has limitations. The health-care use measures are based on the SEER-Medicare FFS population, and the observation period was limited to survivors who were 65 years or older. We excluded older survivors enrolled in MA (representing 30% of older Medicare beneficiaries) because these plans were not previously required to provide detailed information about use to CMS. MA encounter data are now being released by CMS. Future efforts to both estimate the use for those younger than 65 years and evaluate use in MA settings will be important. Our national use estimates include cancer survivors of all ages by assuming that survivors younger than 65 years had the same use patterns as those aged 65-74 years, that is, the youngest survivors for whom data are available. Because younger cancer patients have been shown to receive more intensive cancer treatment than older patients, our projections are likely underestimated. Additionally, projections assumed that future use would be the same as that observed in 2008-2015. We were not able to differentiate cancer survivors who had no evidence of disease vs those who had recurrences. However, using cancer death in the following year as a proxy for more intensive care, we estimated that 11.6%, 3.8%, and 2.8% of cancer survivors who visited medical oncologists in 0-1 years, 5-6 years, and 10+ years since diagnosis, respectively, died of cancer in the following year (see Supplementary Table 6, available online), suggesting that few had experienced recurrences. We only quantified health-care workforce demand but not supply. However, the rates of increase in the workforce use projections can provide insight on supply and the needs to meet the demand.
In summary, this study provides detailed estimates of concurrent care use in all phases of the cancer continuum by a wide range of providers. These findings are especially useful in relation to the NASEM workforce reports (14,15) that encouraged new models of care, including organizational changes to delivering cancer care; leveraging technologies to better support clinicians, cancer survivors, and their families; and pursuing policy changes to support high-quality cancer care. The detailed use data and workforce projections through 2040 can be useful for researchers, professional groups, health-care administrators, and health systems for planning purposes and to evaluate different care models.
Funding
The authors received no financial support for the research, authorship, and/or publication of this article.
Notes
Role of the funder: Not applicable.
Disclosures: KRY serves on the Flatiron Health Equity Advisory Board. The other authors have no relationships or conflicts of interest to disclose.
KRY, who is a JNCI Associate Editor and a co-author on the article, was not involved in the editorial review or decision to publish the manuscript.
Author contributions: ABM, LE, HP, KRY, and DKM: conceptualization, methodology, investigation, visualization, writing—original draft and writing—review & editing. CAZ: formal analyses, software and validation.
Disclaimers: The authors are responsible for the research and had full independence in designing the study, interpreting the data, writing, and publishing the report. The content is solely the responsibility of the authors and does not represent the official views of the National Institutes of Health or the American Cancer Society.
Data Availability
The SEER-Medicare data underlying this article are available at https://healthcaredelivery.cancer.gov/seermedicare/obtain/requests.html.
Supplementary Material
Contributor Information
Angela B Mariotto, Division of Cancer Control and Population Sciences, National Cancer Institute, Bethesda, MD, USA.
Lindsey Enewold, Division of Cancer Control and Population Sciences, National Cancer Institute, Bethesda, MD, USA.
Helen Parsons, Division of Health Policy and Management, University of Minnesota, Minneapolis, MN, USA.
Christopher A Zeruto, Information Management Services, Inc, Calverton, MD, USA.
K Robin Yabroff, Surveillance and Health Equity Science, American Cancer Society, Atlanta, GA, USA.
Deborah K Mayer, Division of Cancer Control and Population Sciences, National Cancer Institute, Bethesda, MD, USA; School of Nursing, University of North Carolina Chapel Hill, Chapel Hill, NC, USA; UNC Lineberger Comprehensive Cancer Center, Chapel Hill, NC, USA.
References
- 1. Miller KD, Nogueira L, Mariotto AB, et al. Cancer treatment and survivorship statistics, 2019. CA Cancer J Clin. 2019;69(5):363–385. doi: 10.3322/caac.21565. [DOI] [PubMed] [Google Scholar]
- 2. Bluethmann SM, Mariotto AB, Rowland JH. Anticipating the “Silver Tsunami”: prevalence trajectories and comorbidity burden among older cancer survivors in the United States. Cancer Epidemiol Biomarkers Prev. 2016;25(7):1029–1036. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3. Chandak AN, Loberiza FR, Deras M, et al. Estimating the state-level supply of cancer care providers: preparing to meet workforce needs in the wake of health care reform. J Oncol Pract. 2015;11(1):32–37. [DOI] [PubMed] [Google Scholar]
- 4. Erikson C, Schulman S, Kosty M, et al. Oncology workforce: results of the ASCO 2007 program directors survey. J Oncol Pract. 2009;5(2):62–65. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5. Pan HY, Haffty BG, Falit B, et al. Supply and demand for radiation oncology in the United States: updated projections for 2015 to 2025. Int J Radiat Oncol Biol Phys. 2016;96(2):E394–E395. [DOI] [PubMed] [Google Scholar]
- 6.IMS Institute for Healthcare Informatics. Global Oncology Trend Report: Review of 2015 and Outlook to 2020. Parsippany, NJ: IMS Institute for Healthcare Informatics; 2016. https://www.iqvia.com/-/media/iqvia/pdfs/institute-reports/global-oncology-trend-report-2016.pdfd. Accessed July 8, 2020. [Google Scholar]
- 7. Stitzenberg KB, Chang Y, Louie R, et al. Improving our understanding of the surgical oncology workforce. Ann Surg. 2014;259(3):556–562. [DOI] [PubMed] [Google Scholar]
- 8. 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. 2014;10(1):39–45. [DOI] [PubMed] [Google Scholar]
- 9. Erikson C, Salsberg E, Forte G, et al. Future supply and demand for oncologists challenges to assuring access to oncology services. J Oncol Pract. 2007;3(2):79–86. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10. Pollack LA, Adamache W, Ryerson AB, et al. Care of long-term cancer survivors physicians seen by Medicare enrollees surviving longer than 5 years. Cancer. 2009;115(22):5284–5295. [DOI] [PubMed] [Google Scholar]
- 11. Warren JL, Mariotto AB, Meekins A, et al. Current and future utilization of services from medical oncologists. J Clin Oncol. 2008;26(19):3242–3247. [DOI] [PubMed] [Google Scholar]
- 12. Fung CY, Chen E, Vapiwala N, et al. The American Society for Radiation Oncology 2017 Radiation Oncologist Workforce Study. Int J Radiat Oncol Biol Phys. 2019;103(3):547–556. [DOI] [PubMed] [Google Scholar]
- 13.Institute of Medicine National Cancer Policy Forum. Ensuring Quality Cancer Care through the Oncology Workforce: Sustaining Care in the 21st Century: Workshop Summary. Washington, DC: National Academies Press (US; ); 2009.Accessed July 8, 2020 [PubMed] [Google Scholar]
- 14. Takvorian SU, Balogh E, Nass S, et al. Developing and sustaining an effective and resilient oncology careforce: opportunities for action. J Natl Cancer Inst. 2020;112(7):663–670. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15. Kline RM, Arora NK, Bradley CJ, et al. Long-term survivorship care after cancer treatment - summary of a 2017 National Cancer Policy Forum Workshop. J Natl Cancer Inst. 2018;110(12):1300–1310. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.National Academies of Sciences, Engineering, and Medicine. Developing and Sustaining an Effective and Resilient Oncology Careforce: Proceedings of a Workshop. Washington, DC: The National Academies Press; 2019. [PubMed] [Google Scholar]
- 17.SEER Registry Groupings for Analysis. http://seer.cancer.gov/registries/terms.html. Acessed March 10, 2020.
- 18. Enewold L, Parsons H, Zhao L, et al. Updated overview of the SEER-Medicare data: enhanced content and applications. J Natl Cancer Inst Monogr. 2020;2020(55):3–13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.American Medical Association. AMA physician masterfile. https://www.ama-assn.org/life-career/ama-physician-masterfile. Accessed November 10, 2019.
- 20.National Uniform Claim Committee. Health care provider taxonomy. https://www.nucc.org/index.php/code-sets-mainmenu-41/provider-taxonomy-mainmenu-40. Accessed November 10, 2019.
- 21. Warren JL, Barrett MJ, White DP, et al. Sensitivity of Medicare data to identify oncologists. J Natl Cancer Inst Monogr. 2020;2020(55):60–65. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22. Cooper GS, Yuan Z, Stange KC, et al. The sensitivity of Medicare claims data for case ascertainment of six common cancers. Med Care. 1999;37(5):436–444. [DOI] [PubMed] [Google Scholar]
- 23. Lam CJK, Enewold L, McNeel TS, et al. Estimating chemotherapy use among patients with a prior primary cancer diagnosis using SEER-Medicare data. J Natl Cancer Inst Monogr. 2020;2020(55):14–21. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24. Mariotto AB, Enewold L, Zhao JX, et al. Medical care costs associated with cancer survivorship in the United States. Cancer Epidemiol Biomarkers Prev. 2020;29(7):1304–1312. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25. Walker B, Frytak J, Hayes J, et al. Evaluation of practice patterns among oncologists participating in the oncology care model. JAMA Netw Open. 2020;3(5):e205165. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26. Dale W, Chow S, Sajid S. Socioeconomic considerations and shared-care models of cancer care for older adults. Clin Geriatr Med. 2016;32(1):35–44. [DOI] [PubMed] [Google Scholar]
- 27. Lisy K, Kent J, Piper A, et al. Facilitators and barriers to shared primary and specialist cancer care: a systematic review. Support Care Cancer. 2021;29(1):85–96. [DOI] [PubMed] [Google Scholar]
- 28. Nekhlyudov L, O'Malley DM, Hudson SV. Integrating primary care providers in the care of cancer survivors: gaps in evidence and future opportunities. Lancet Oncol. 2017;18(1):e30–e38. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29. Crabtree BF, Miller WL, Howard J, et al. Cancer survivorship care roles for primary care physicians. Ann Fam Med. 2020;18(3):202–209. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 30. Coombs LA, Max W, Kolevska T, et al. Nurse practitioners and physician assistants: an underestimated workforce for older adults with cancer. J Am Geriatr Soc. 2019;67(7):1489–1494. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 31. Coombs LA, Hunt L, Cataldo J. A scoping review of the nurse practitioner workforce in oncology. Cancer Med. 2016;5(8):1908–1916. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The SEER-Medicare data underlying this article are available at https://healthcaredelivery.cancer.gov/seermedicare/obtain/requests.html.