Diversity Trends by Sex and Underrepresented in Medicine Status Among US Radiation and Medical Oncology Faculty Over 5 Decades

Key Points

Question  How has the composition (by sex and by race and ethnicity) of US radiation oncology (RO) and medical oncology (MO) academic departments’ workforces evolved over time?

Findings  This cross-sectional study assessed demographic trends among RO and MO academic departments during the past 5 decades. Sex and racial and ethnic diversity of RO and MO faculty has increased over time but has not kept pace with that of the US population, particularly with respect to individuals designated as underrepresented in medicine (URM).

Meaning  Efforts to recruit and retain URM individuals among academic oncology departments are critical to maintain a diverse workforce and training environment; the increase in women faculty in both specialties may inform measures to achieve similar progress among URM faculty.

Importance  It remains unclear how the historical exclusion of women and racial and ethnic minority groups from medical training, and therefore the oncologic subspecialties, has contributed to rates of faculty diversity among oncology departments over time. Oncologic faculty diversity is an important initiative to help improve care and address health disparities for an increasingly diverse US population with cancer.

Objectives  To report trends in academic faculty representation by sex and by race and ethnicity for radiation oncology (RO) and medical oncology (MO) departments and to describe comparisons with the general US population, medical students, RO and MO trainees, clinical department chairs, and faculty in other departments.

Design, Setting, and Participants  This cross-sectional analysis used data from the Association of American Medical Colleges to analyze trends by sex and by race and ethnicity among full-time US faculty in RO and MO departments from 1970 through 2019. Data were analyzed between October 2020 and April 2021.

Main Outcomes and Measures  Proportions of women and individuals from underrepresented in medicine (URM) racial and ethnic groups (Black, Hispanic, and Indigenous individuals) were calculated among RO and MO academic departments; trends were analyzed over 5 decades. These proportions were compared with cohorts already described. In addition, proportions of women and URM individuals were calculated by faculty rank among RO and MO departments.

Results  In 1970, there were 119 total faculty in RO (10 women [8.4%] and 2 URM [1.7%]) and 87 total faculty in MO (11 women [12.6%] and 7 URM [8.0%]). In 2019, there were 2115 total faculty in RO (615 women [29.1%] and 108 URM [5.1%]) and 819 total faculty in MO (312 women [38.1%] and 47 URM [5.7%]). Total faculty numbers increased over time in both RO and MO. Faculty representation of URM women proportionally increased by 0.1% per decade in both RO (95% CI, 0.005%-0.110%; P <. 001 for trend) and MO (95% CI, −0.03% to 0.16%; P = .06 for trend) compared with non-URM women faculty, which increased by 0.4% (95% CI, 0.25%-0.80%) per decade in RO and 0.7% (95% CI, 0.47%-0.87%) per decade in MO (P < .001 for trend for both). Faculty representation of URM men did not significantly change for RO (0.03% per decade [95% CI, −0.008% to 0.065%]; P = .09 for trend) or MO (0.003% per decade [95% CI, −0.13% to 0.14%]; P = .94 for trend). Representation of both women and URM individuals among both specialties was lower than their representation in the US population in both 2009 and 2019. Across all cohorts studied, RO faculty had the lowest URM representation in 2019 at 5.1%. At every rank in 2019, the number of total URM faculty represented among both MO and RO remained low (MO: instructor, 2 of 44 [5%]; assistant professor, 18 of 274 [7%]; associate professor, 13 of 177 [7%]; full professor, 13 of 276 [5%]; and RO: instructor, 9 of 147 [6%]; assistant professor, 57 of 927 [6%]; associate professor, 20 of 510 [4%]; full professor, 18 of 452 [4%]).

Conclusions and Relevance  This cross-sectional study suggests that RO and MO academic faculty have increased the representation of women over time, while URM representation has lagged. The URM trends over time need further investigation to inform strategies to improve URM representation in RO and MO.

Academic oncology faculty members are requisite in training future generations of oncologists to care for an increasingly diverse population of patients with cancer. It is estimated that 49% of all new cancer cases in 2021 will occur in women.¹ Among Black, Indigenous, and Hispanic populations, new cancer cases are estimated at 457.6, 379.8, and 346.9 cases per 100 000 people, respectively,¹ with higher incidence and mortality compared with their non-Hispanic White counterparts.^2,3 It has been documented that a health care workforce that better reflects the demographic characteristics of those it serves may lead to improved patient satisfaction, compliance, and outcomes.^4,5 Historically, women and certain racial and ethnic minority groups were excluded from and/or had limited pathways into medicine and, consequently, oncology training.⁶ Calls to increase health care workforce diversity date back several decades^6-9 and resulted in the creation of some opportunities and pathways to increase the representation of women and racial and ethnic minority groups in medicine, albeit with a documented period of stagnation⁶ for racial and ethnic minority groups between 1974 and 1990 because of a backlash against such policies (notably, the 1978 University of California v Bakke decision, which directly challenged affirmative action in medical schools).¹⁰ Calls for improved diversity in the health care workforce have increased in urgency and number in recent years. It is unclear whether the growing imperative to address disparities in representation by sex and by race and ethnicity in the medical field has resulted in corresponding progress in the composition of academic radiation oncology (RO) and medical oncology (MO) departments. This is an increasing priority for oncology—the American Society of Clinical Oncology released a strategic plan for workforce diversity in 2017,¹¹ and, more recently, a 2020 report by the American Association for Cancer Research highlighted the need to train a diverse oncology workforce to overcome cancer health disparities.¹² Although the lack of diversity among RO trainees and in the general RO and MO workforce has been previously highlighted,^13-17 trends in representation both by race and ethnicity and by sex, specifically among academic RO and MO departments over 5 decades, have not been reported. In addition, while the overall diversity of medical school faculty has been increasing by race and ethnicity and by sex,¹⁸ the change is of noticeably lesser magnitude than that seen among medical school applicants, matriculants, and graduates.¹⁹ The increase in medical school faculty diversity also has not kept pace with the increasing diversity of the US population.¹⁸ Herein, we report trends in faculty diversity, overall and by academic rank, among US RO and MO departments during the past 5 decades.

Data were acquired from the Association of American Medical Colleges (AAMC) full-time Faculty Roster between 1970 and 2019²⁰ for academic RO and MO departments, to assess sex and racial and ethnic trends over 5 decades. As the AAMC does not separately list RO and MO departments but instead groups each under larger department classifications, RO and MO departments were identified per eTable 1 in the Supplement. Data were also acquired for surgery, internal medicine, and radiology departments, as well as all clinical department chairs. Individual decades were defined as 1970-1979, 1980-1989, 1990-1999, 2000-2009, and 2010-2019. The AAMC FACTS^21,22 and Electronic Residency Applicant Service tables²³ were used to obtain data on medical school applicants, matriculants, and graduates; RO residency applicants; hematology and oncology and oncology fellowship (hereafter referred to as MO fellowship) applicants; and current trainees. This study was determined by the Mass General Brigham institutional review board to be exempt from human subjects research guidelines because secondary analysis of existing deidentified data.

Similar to Xierali et al,²⁴ we defined racial and ethnic groups as mutually exclusive (eMethods in the Supplement). Underrepresented in medicine (URM) status refers to Black, Hispanic, and Indigenous individuals, per the AAMC definition of racial and ethnic groups that are underrepresented in the medical profession relative to their numbers in the general population.^25,26

Proportions of faculty by sex, URM status, and faculty rank were calculated to assess long-term trends between 1970 and 2019. Simple linear regression models were used to estimate trends and associated P values, with year as the independent variable. Linear trends of absolute numbers as well as proportions by sex and URM status were evaluated by year and by decade for faculty departments and for faculty rank. As yearly changes were small, change per decade (based on the year representing the end of the decade) is presented. Absolute percentage change between 2019 and 1979 was calculated by sex and URM status among different clinical specialties. The Marascuilo procedure²⁷ was used to perform pairwise comparisons by sex and by race and ethnicity among the US population; medical school applicants, matriculants, and graduates; RO and MO trainees; RO and MO faculty; and department chairs for 2019. The absolute difference in proportions represented in the US population and the aforementioned cohorts was calculated by sex and URM status. A 2-sample test of proportions was used to compare the difference to a value of 0 (null hypothesis). P ≤ .05 was considered statistically significant. All statistical analyses were performed using R, version 4.0.1 (R Group for Statistical Computing).

In 1970, there were 119 total faculty in RO (10 women [8.4%] and 2 URM [1.7%]) and 87 total faculty in MO (11 women [12.6%] and 7 URM [8.0%]). In 2019, there were 2115 total faculty in RO (615 women [29.1%] and 108 URM [5.1%]) and 819 total faculty in MO (312 women [38.1%] and 47 URM [5.7%]). Overall, the total number of faculty members increased over time in both the RO and MO departments. A breakdown of faculty growth trends by sex and URM status demonstrated an increase of non-URM women faculty by 0.4% (95% CI, 0.25%-0.80%) per decade in RO (P <.001 for trend) and by 0.7% (95% CI, 0.47%-0.87%) in MO (P < .001 for trend) (Figure 1; eTable 2 in the Supplement). The number of URM women faculty significantly increased by 0.1% per decade in RO (95% CI, 0.005%-0.110%; P < .001 for trend) and nonsignificantly by 0.1% per decade in MO (95% CI, −0.03% to 0.16%; P = .06 for trend). Faculty representation of URM men did not statistically significantly proportionally increase per decade in either MO or RO. Men comprised the majority among both URM and non-URM faculty for both RO and MO; URM faculty remained below 10% of total faculty throughout the entire study period for both RO and MO. Between 1970 and 1990, there was a steady decrease in the proportion of MO URM faculty, after which there was slow increase. For every 1 additional URM man per decade, there were 16 (25.0/1.6) and 13 (10.0/0.8) additional non-URM men in RO and MO, respectively, while for every 1 additional URM woman per decade, there were 13 (13.0/1.0) and 13 (8.0/0.6) additional non-URM women in RO and MO, respectively.

There were significant differences in diversity by faculty rank over time (eFigure 1 in the Supplement). The proportion of women increased more than the proportion of URM individuals among RO and MO faculty among the different ranks, yet a higher proportion of women faculty members had lower academic rank than men. There was a higher proportion of women faculty members with more advanced rank at the end of modern decades (1999, 2009, and 2019) among MO faculty than RO faculty (MO faculty full professors: 1999, 15% [20 of 138]; 2009, 20% [42 of 206]; and 2019, 28% [77 of 276]; RO faculty full professors: 1999, 8% [24 of 285]; 2009, 14% [47 of 348]; and 2019, 21% [94 of 452] Table). Among both RO and MO, there were significant proportional increases in women at the full, associate, and assistant professor ranks. When examining trends by URM status, URM representation significantly increased at the full and assistant professor levels for RO and at the associate professor level for MO faculty (Table), but in general, numbers remained very low throughout.

When comparing trends of proportions of women RO and MO faculty throughout each decade relative to other departments, women’s representation increased significantly for all evaluated departments (Figure 2A). Comparing proportions between 1979 and 2019, internal medicine had the greatest absolute percentage change (10% [1033 of 10 307] in 1979 to 41% [17 616 of 43 219] in 2019), followed by MO (10% [18 of 178] in 1979 to 38% [312 of 819] in 2019) and surgery (5% [184 of 4099] in 1979 to 26% [4251 of 16 221] in 2019), while both radiology and RO had similar absolute percentage changes (radiology, 12% [396 of 3249] in 1979 and 30% [2933 of 9895] in 2019; and RO, 10% [42 of 427] in 1979 and 29% [615 of 2115] in 2019).

When comparing trends of proportions for URM faculty throughout each decade, URM representation increased significantly for all evaluated departments (Figure 2B). Comparing proportions between 1979 and 2019, internal medicine had the greatest statistically significant absolute percentage change in URM faculty (4% [419 of 10 307] in 1979 to 10% [4229 of 43 219] in 2019), followed by surgery (4% [176 of 4099] in 1979 to 9% [1507 of 16 221] in 2019) and radiology (4% [129 of 3249] in 1979 to 7% [665 of 9895] in 2019). Change in URM faculty proportion was not significant for MO (5% [9 of 178] in 1979 to 6% [47 of 819] in 2019; P = .86).

To understand the current diversity among RO and MO faculty relative to other clinical departments; clinical department chairs; medical school applicants, matriculants, and graduates; RO and MO trainees; and the 2019 US population,²⁸ we compared proportions of women and URM individuals in each of these cohorts for 2019 (eTable 3 in the Supplement).²³ The proportion of women applying to (52% [27 847 of 53 341]), matriculating at (52% [11 461 of 21 863]), and graduating from (48% [9557 of 19 938]) medical school in 2019 was similar to representation among the US population (51% [166 650 550 of 328 239 523]). Representation of women among RO residency applicants (30% [194 of 637]), RO residents (30% [220 of 724]), and RO faculty (29% [615 of 2115]) in 2019 was lower than MO fellowship applicants (46% [514 of 1128]), MO fellows (48% [831 of 1744]), and MO faculty (38% [312 of 819]). Department chairs across all clinical departments had the lowest representation of women in 2019 (17% [402 of 2398]) of all groups compared.

When comparing proportions of URM individuals, the highest was observed in the US population at 31% (104 879 607 of 328 239 523), followed by medical school applicants (18% [9867 of 53 341]) and matriculants (17% [3942 of 21 863]). The proportion of URM medical school graduates was lower at 11% (2348 of 19 938), similar to RO residency applicants (13% [78 of 637]) and MO fellowship applicants (10% [115 of 1128]). Representation of URM individuals among RO faculty (5% [108 of 2115]) and MO faculty (5% [47 of 819]) was lower than URM representation among all clinical department chairs (9% [224 of 2398]). Radiation oncology faculty had the highest representation of Asians in 2019 (34% [711 of 2115]), which was significantly higher than the US population (6% [18 427 914 of 328 239 523]), clinical department chairs (10% [237 of 2398]), radiology faculty (27% [2660 of 9895]), internal medicine faculty (25% [10 735 of 43 219]), surgery faculty (18% [2873 of 16 221]), and medical school applicants (21% [11 027 of 53 341]), matriculants (22% [4687 of 21 863]), and graduates (22% [4300 of 19 938]). Department chairs had the highest proportion of White individuals (79% [1888 of 2398]) in comparison with all groups examined.

When evaluating these trends between the end of the 2 most recent decades (2009 and 2019), in general, the absolute difference between the proportion of women in an evaluated group and the proportion of women in the US population decreased for all groups (ie, became more representative) between 2009 and 2019, except for medical school graduates and RO residency applicants (Figure 3A). The proportion of the US population from racial and ethnic backgrounds considered to be URM has increased between 2009 and 2019. The absolute difference between the proportion of URM in an evaluated group and the proportion of URM in the US population increased for all groups (ie, became less representative) between 2009 and 2019, except for medical school applicants and matriculants (Figure 3B). Yearly trends from 2009 to 2019 are presented in eFigure 2 in the Supplement.

Creating and maintaining a diverse health care workforce is a priority to help combat societal inequities and health disparities, particularly in light of the evolving demographic characteristics of the general US population. The US Census Bureau estimates that non-Hispanic White individuals, currently the majority group accounting for greater than 50% of the nation’s total population, will no longer hold this title by 2044, as it will decrease below 50%.²⁹ Separately, women make up just above 50% of the entire US population, and the life expectancy of US women is longer than that of US men.³⁰ As the general population ages and diversifies, we can anticipate its implications on our cancer population. Cancer burden has been recognized as unequally affecting certain racial and ethnic groups in the US, with one example being the highest cancer death rate among African Americans compared with all other racial and ethnic groups.³¹

Oncology physician workforce diversity is identified as a critical aspect to reduce cancer health disparities in multiple ways. Patients report greater satisfaction with physicians of concordant race and ethnicity,³² and published data demonstrate that this concordance promotes adherence to treatment recommendations and improved outcomes.^33-35 A diverse physician workforce expands health care access³⁶; URM physicians are more likely to practice in medically underserved areas,^37-39 an essential initiative for the RO workforce.⁴⁰ Workforce diversity ensures inclusion of more viewpoints and enhances innovation, creativity, problem solving, and productivity.^7,41,42 Furthermore, a representative cancer workforce can make much-needed strides in ensuring that biomedical research is inclusive and extends to all sectors of the US population. This inclusion is particularly urgent with regard to cancer clinical trials, which are underrepresentative of the US cancer population.^43,44

Despite previous recognition and calls to increase physician workforce diversity more broadly, it is not clear how these appeals have translated to gains in diversity over time among oncology faculty. Given that oncology faculty are typically responsible for training the next generation of oncologists, with data suggesting that URM faculty can aid with recruitment and retainment of URM individuals,^36,38,39,45 we sought to understand these long-term trends spanning 5 decades, with particular attention to recent decades. We demonstrate that, despite increased diversity of RO and MO faculty over time, this diversity has not kept pace with that of the US population, particularly with respect to URM status. Diversity of RO and MO faculty lags behind medical school diversity, which has grown during the past decade, presumably in large part because of high-profile efforts by the AAMC to promote diversity, equity, and inclusion among medical schools and the physician workforce.^6-8,19 In addition, our results highlight significant diversity differences along the career ladder in both specialties, with women having lower academic rank than men throughout the study period, and URMs underrepresented at every rank. The lack of women in higher academic ranks has been demonstrated in other clinical departments²⁴ as well as in academic medicine as a whole⁴⁶; this finding is particularly true for URM women.⁴⁷ Medical oncology outperformed RO, with more women at higher ranks in the modern era, while URM physicians continued to remain underrepresented.

In comparisons across different groups for 2019, we observed that representation of women in medical school has improved dramatically, with applicants and matriculants exceeding the representation of women in the US population. However, we see lower representation of women among MO faculty, RO residency applicants, and RO faculty, and the lowest representation of women among clinical academic chairs, consistent with prior studies.^17,46 Despite approximately 31% URM representation among the US population, the health care workforce trails at all stages in the pipeline. The efforts to improve representation of URM groups among medical school applicants and matriculants has translated to gains over time,^6,48 steady at 18% in 2019, increased from 15% in 2009. A gap between URM representation among medical school applicants and matriculants (approximately 18%) and graduates (approximately 11%) has emerged. A small study of a survey distributed to 167 URM medical students, residents, and practicing physicians identified a lack of faculty member role models from racial and ethnic minority groups in academics as a potential barrier, but further work is necessary to understand all contributory factors.⁴⁹ Additional gaps in URM representation emerged between 2009 and 2019. Very low URM representation among RO faculty (from 6% in 2009 to 5% in 2019) does not bode well for improving representation at higher academic ranks or department chair level. This study highlights the need for more initiatives to ensure successful retention of URM individuals throughout the academic promotional pipeline for RO specifically, to better mirror the success of other specialties (ie, surgery and internal medicine). This is a multifactorial issue, with focus not only on increasing diversity of the upstream pipeline but maintaining diversity throughout the entire pipeline, requiring difficult but necessary conversations about racial and ethnic systemic bias, lack of exposure and opportunities, and financial toxicities and pressures, to name a few.^13,50 Until these factors are further delineated and better addressed, focused and targeted mentorship is key.

This study has some limitations. The AAMC relies on data provided by medical schools. Thus, it is not known how race and ethnicity or sex data were obtained from faculty. We were unable to further stratify mixed races and ethnicities, which arguably will play more of an important role as the US population further diversifies. We acknowledge that the definitions of racial categories used by the US Census and AAMC do not capture cultural heterogeneity within the groups. We further acknowledge that sex is nonbinary; however, we are not able to capture that depth of data here. Data from the AAMC MO and RO Faculty Roster are based on medical school department classifications—this limited the reporting of MO faculty, as some medical schools have MO faculty under divisions of internal medicine, which are not included in the AAMC Faculty Roster. In addition, the data presented are limited to full-time faculty, as defined by individual departments and medical schools. This affected MO faculty after 2017, as some medical school MO departments changed their definition of full-time faculty, thus decreasing the reported total full-time MO faculty significantly. However, our analysis is strengthened given the comprehensiveness and granularity of data on race and rank over time. Furthermore, we do not think that faculty in a division under internal medicine vs a separate academic department should have inherent differences in hiring and demographics, and thus our analysis is well representative of the academic MO workforce to infer trends. Our MO data reflect the general MO workforce representation as our numbers are similar to the demographics presented in the American Society of Clinical Oncology members self-reported oncology landscape⁵¹ and the 2018 American Medical Association Physician Masterfile⁵² (eFigure 3 in the Supplement). We did not further examine subgroups of URMs (eg, Black vs Hispanic vs Indigenous individuals) over time, given the exceedingly small numbers and proportions and resultant “noise” in the analyses. However, we performed a deeper subgroup analysis for 2019 (eTable 3 in the Supplement) to compare to the US and other relevant populations. Finally, we cannot infer causality from the cross-sectional results presented, such as a relationship between medical school diversity and faculty diversity. Despite these limitations, we believe that the faculty data provided over multiple decades (including by rank) provide a representative snapshot of RO and MO academic faculty, and that this analysis can serve as a benchmark for future studies to assess progress for the coming decades.

Recent events have renewed attention surrounding the importance of diversity and inclusion in the physician workforce.^53-56 Although the existing gaps, particularly in RO, have been previously described,^{13,14,16,17,57-60} to our knowledge, this is the first in-depth examination of the academic oncology workforce trends for women and URM individuals specifically for US-based RO and MO departments. As academic departments bear primary responsibility for recruiting and educating our future oncology workforce, it is critical that academic faculty are inclusive of individuals who reflect the US population. Overall, our results indicate that, despite marginal improvements, more progress is needed to recruit and retain a diverse workforce in both RO and MO, as the current academic workforce still does not reflect the diversity of the general US cancer population nor of the medical school population, particularly regarding URM status. Our cancer population will continue to diversify, and our health care system must prepare to meet the needs of these diverse patients. Combating cancer health disparities requires a multipronged approach, one of which is by diversifying our own oncology physician workforce, along with fostering cultural humility and emphasizing the need for culturally competent delivery of care. A key step is to further understand the barriers and shortcomings of current approaches to recruitment and retainment of individuals who are underrepresented in the academic oncology workforce to inform future efforts, along with deliberate investment into career development and advancement of these traditionally marginalized groups. We urge all stakeholders to support these efforts toward an oncology physician workforce prepared to provide equitable care and reduce disparities for the diversifying US cancer population.

Accepted for Publication: August 28, 2021.

Published Online: December 9, 2021. doi:10.1001/jamaoncol.2021.6011

Corresponding Author: Sophia C. Kamran, MD, Department of Radiation Oncology, Massachusetts General Hospital, Cox 3, 55 Fruit St, Boston, MA 02114 (skamran@mgh.harvard.edu).

Author Contributions: Dr Kamran had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Kamran, Vapiwala.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Kamran, Niemierko, Vapiwala.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Kamran, Niemierko.

Obtained funding: Vapiwala.

Administrative, technical, or material support: Deville, Vapiwala.

Supervision: Kamran, Deville, Vapiwala.

Conflict of Interest Disclosures: Dr Kamran’s spouse is employed by Sanofi Genzyme. No other disclosures are reported.

Additional Contributions: We acknowledge Rae Anne Sloane, BA, Association of American Medical Colleges, for assisting with the generation of the data sets. She was not compensated for her contribution.