Identifying Barriers to Building a Diverse Physician Workforce: A National Survey of the ACR Membership
Article InfoArticle Outline
Abstract
Purpose
The aim of this study was to identify potential barriers to building a diverse workforce in radiology and radiation oncology by conducting a national survey of physicians in these fields and studying their reported career experiences.
Methods
An electronic survey of ACR members (February 27, 2018, to April 26, 2018) was conducted in which physicians’ attitudes about their work environment, relationships, and culture were queried. The aim was to determine if responses differed by gender or race/ethnicity. In total, 900 invitations were issued; women were oversampled with the goal of equal representation. Descriptive summaries (proportions of yes or no responses) were calculated per item, per subgroup of interest. Logistic regression analysis was used to identify significant associations between gender- and item-specific responses; it was not used in the race/ethnicity analysis because of the small sizes of many subgroups.
Results
The response rate was 51.2% (461 of 900). In total, 51.0% of respondents identified as women (235 of 461); the 9.5% (44 of 461) who identified as black or African American, Hispanic, or American Indian or Alaska Native were considered underrepresented minorities. Respondents’ mean age was 40.2 ± 10.4 years. Subgroups varied most in their reporting of unfair or disrespectful treatment. Women were significantly more likely than men to report such treatment attributable to gender (50.6% versus 5.4%; odds ratio, 18.00; 95% confidence interval, 9.29-34.86; P < .001), and 27.9% of underrepresented minorities compared with 2.6% of white non-Hispanic respondents reported such treatment attributable to race/ethnicity.
Conclusions
Women and underrepresented minorities disproportionately experience unfair or disrespectful treatment in the workplace. Addressing this problem is likely to be critically important for improving workforce diversity.
Key Words:
Diversity, gender, underrepresented minority, workforce, barriers, survey, radiologist, radiation oncologistIntroduction
To achieve greater diversity in the US physician workforce, agents of change must consider the composition of the population of medical school graduates, as well as that of individual medical specialties [1x1Lett, L.A., Orji, W.U., and Sebro, R. Declining racial and ethnic representation in clinical academic medicine: a longitudinal study of 16 US medical specialties. PLoS One. 2018;
13: e0207274
Crossref | PubMed | Scopus (11) | Google ScholarSee all References, 2x2Chapman, C.H., Hwang, W.T., Both, S., Thomas, C.R. Jr., and Deville, C. Current status of diversity by race, Hispanic ethnicity, and sex in diagnostic radiology. Radiology. 2014;
270: 232–240
Crossref | PubMed | Scopus (55) | Google ScholarSee all References, 3x3Chapman, C.H., Hwang, W.T., and Deville, C. Diversity based on race, ethnicity, and sex, of the US radiation oncology physician workforce. Int J Radiat Oncol Biol Phys. 2013;
85: 912–918
Abstract | Full Text | Full Text PDF | PubMed | Scopus (32) | Google ScholarSee all References, 4x4Association of American Medical Colleges. Diversity in medical education: facts and figures 2016.. (Available at:)http://www.aamcdiversityfactsandfigures2016.org/.
Google ScholarSee all References]. Approximately 1 in 2 medical school graduates are women and 1 in 10 are underrepresented minorities (URMs) [4x4Association of American Medical Colleges. Diversity in medical education: facts and figures 2016.. (Available at:)http://www.aamcdiversityfactsandfigures2016.org/.
Google ScholarSee all References]. In radiology and radiation oncology, women represent 1 in 4 physicians, and URMs represent 1 in 14 to 15 physicians [2x2Chapman, C.H., Hwang, W.T., Both, S., Thomas, C.R. Jr., and Deville, C. Current status of diversity by race, Hispanic ethnicity, and sex in diagnostic radiology. Radiology. 2014;
270: 232–240
Crossref | PubMed | Scopus (55) | Google ScholarSee all References, 3x3Chapman, C.H., Hwang, W.T., and Deville, C. Diversity based on race, ethnicity, and sex, of the US radiation oncology physician workforce. Int J Radiat Oncol Biol Phys. 2013;
85: 912–918
Abstract | Full Text | Full Text PDF | PubMed | Scopus (32) | Google ScholarSee all References]. These findings indicate that our professions have been unable to successfully recruit women and URMs to the same extent as many other medical specialties.
Factors driving the lack of gender, racial, and ethnic diversity in radiology and radiation oncology are poorly understood. To date, most related research has focused on gender disparities in markers of professional success (eg, leadership, practice ownership) [5x5Baker, S.R., Barry, M., Chaudhry, H., and Hubbi, B. Women as radiologists: are there barriers to entry and advancement?. J Am Coll Radiol. 2006;
3: 131–134
Abstract | Full Text | Full Text PDF | PubMed | Scopus (44) | Google ScholarSee all References, 6x6Lewis, R.S., Bhargavan, M., and Sunshine, J.H. Women radiologists in the United States: results from the American College of Radiology’s 2003 survey. Radiology. 2007;
242: 802–810
Crossref | PubMed | Scopus (49) | Google ScholarSee all References, 7x7Jalilianhasanpour, R., Charkhchi, P., Mirbolouk, M., and Yousem, D.M. Underrepresentation of women on radiology editorial boards. J Am Coll Radiol. 2019;
16: 115–120
Abstract | Full Text | Full Text PDF | PubMed | Scopus (6) | Google ScholarSee all References, 8x8Piper, C.L., Scheel, J.R., Lee, C.I., and Forman, H.P. Representation of women on radiology journal editorial boards: a 40-year analysis. Acad Radiol. 2018;
25: 1640–1645
Abstract | Full Text | Full Text PDF | PubMed | Scopus (13) | Google ScholarSee all References, 9x9Jena, A.B., Olenski, A.R., and Blumenthal, D.M. Sex differences in physician salary in US public medical schools. JAMA Intern Med. 2016;
176: 1294–1304
Crossref | PubMed | Scopus (169) | Google ScholarSee all References, 10x10Kapoor, N., Blumenthal, D.M., Smith, S.E., Ip, I.K., and Khorasani, R. Sex differences in radiologist salary in U.S. public medical schools. AJR Am J Roentgenol. 2017;
209: 953–958
Crossref | PubMed | Scopus (7) | Google ScholarSee all References]. To our knowledge, a body of literature that similarly addresses URMs is absent. Although efforts to describe such disparities are an important starting point for diversifying the workforce of our specialties, alone, they are likely to be insufficient. For example, recent data suggest that in academic settings, gender-based salary discrepancies may be lower in radiology compared with other medical specialties when adjusting for specific factors (eg, publication count) [9x9Jena, A.B., Olenski, A.R., and Blumenthal, D.M. Sex differences in physician salary in US public medical schools. JAMA Intern Med. 2016;
176: 1294–1304
Crossref | PubMed | Scopus (169) | Google ScholarSee all References]. Efforts to close this gap have been important, but women remain underrepresented in academic radiology [2x2Chapman, C.H., Hwang, W.T., Both, S., Thomas, C.R. Jr., and Deville, C. Current status of diversity by race, Hispanic ethnicity, and sex in diagnostic radiology. Radiology. 2014;
270: 232–240
Crossref | PubMed | Scopus (55) | Google ScholarSee all References]. This circumstance suggests that workforce disparities may be driven in part by core problems that remain unexposed and that attenuate the benefits of such corrective efforts.
Discovering how physicians perceive key attributes of their work environments and professional relationships, and how such perceptions may differ among women and URMs, is critically important for identifying barriers to increasing diversity in our professions. Past studies have investigated factors that could represent deterrents for women, including limited patient contact, an emphasis on technology, gender-based discrimination or harassment, and a lack of role models [11x11Fielding, J.R., Major, N.M., Mullan, B.F., Neutze, J.A., Shaffer, K., Wilcox, C.B. et al. Choosing a specialty in medicine: female medical students and radiology. AJR Am J Roentgenol. 2007;
188: 897–900
Crossref | PubMed | Scopus (49) | Google ScholarSee all References, 12x12Roubidoux, M.A., Packer, M.M., Applegate, K.E., and Aben, G. Female medical students’ interest in radiology careers. J Am Coll Radiol. 2009;
6: 246–253
Abstract | Full Text | Full Text PDF | PubMed | Scopus (52) | Google ScholarSee all References, 13x13Frank, E. and Vydareny, K. Characteristics of women radiologists in the United States. AJR Am J Roentgenol. 1999;
173: 531–536
Crossref | PubMed | Scopus (11) | Google ScholarSee all References, 14x14Zener, R., Lee, S.Y., Visscher, K.L., Ricketts, M., Speer, S., and Wiseman, D. Women in radiology: exploring the gender disparity. J Am Coll Radiol. 2016;
13 (344-50.e1)
Abstract | Full Text | Full Text PDF | PubMed | Scopus (29) | Google ScholarSee all References]. Again, to our knowledge, corresponding data for URMs are absent. A contemporary, national survey has not been conducted to further query and expose these and other potential factors.
To address this paucity of evidence, we developed, conducted, and analyzed a national survey of the physician membership of the ACR. We asked members questions about their attitudes and experiences in their professional environments to understand the climate of their workplaces and determine if their responses differed by gender and URM status. In so doing, our goal was to shed light on potential barriers to building a diverse physician workforce in radiology and radiation oncology. In this report we present the first and primary analyses from this study.
Methods
Human Subjects
This study received an exemption from the Massachusetts General Hospital (MGH) institutional review board.
Overview of the Study Design
We prospectively designed and executed a national survey of the membership of the ACR. We reviewed deidentified, aggregate data from more than 36,000 members (July 2017) to develop our study design. In the eligible respondent pool, we included all active (nonretired) physician members, including trainees and attending (posttraining) radiologists and radiation oncologists. We excluded physicists, students, and retired physician members and individuals without e-mail addresses. We decided, on the basis of the extant distribution of men and women in the overall membership data, to oversample by gender to ensure sufficient completed cases for an analysis of major outcomes by gender; we sought to achieve 50% female and 50% male respondents. We were unable to oversample URMs because ACR member profiles were sparsely populated with race/ethnicity data; such data were available for less than 600 members. We invited 900 members to participate. Further details regarding our sampling plan are included in the Appendix.
Study Procedures
Development of the Survey Instrument
We developed the ACR Diversity Survey from 2017 to 2018, drawing on and adapting from items previously used by Dr. Karen Donelan to measure a culture of diversity and respect in surveys of patients and the hospital workforce [15x15Donelan K, Bareto E, Michael C, Bonner S. 2016 Diversity workforce culture survey, final report. Boston: Massachusetts General Hospital. March, 2017.
Google ScholarSee all References, 16x16Donelan K, Michael C. MGH diversity metrics project: Supporting effective strategies to develop a diverse workforce, final report, Boston: Massachusetts General Hospital. March, 2017.
Google ScholarSee all References] and using additional items contributed by members of the expert team. A small subset of items were adapted from the Association of American Universities Climate Survey on Sexual Assault and Sexual Misconduct [17x17Association of American Universities. Climate Survey on Sexual Assault and Sexual Misconduct. (Available at:)https://www.aau.edu/key-issues/aau-climate-survey-sexual-assault-and-sexual-misconduct-2015.
Google ScholarSee all References]. The survey was drafted and underwent revisions by the chair of the ACR Commission for Women and Diversity, and by members of the Research Workgroup of this commission, before its finalization. The survey was designed to be self-administered online; prior versions (of the MGH survey) have also been self-administered on paper. The ACR version was pretested by survey experts and physicians before administration in the full sample. In the Appendix, we show all survey items, mapping them to the analyses we report in the results. The survey included five domains of items: (1) work environment and relationships, (2) profession attributes (eg, attitudes about radiation exposure, technology), (3) culture of diversity and respect, (4) career recommendation (ie, would you recommend your choice of profession to medical students?), and (5) personal and professional characteristics. The survey included 29 items in domains 1 to 4. In analyses of responses by gender and race/ethnicity, respectively, subsets of items were analyzed.
Importantly, in the administration of surveys of diversity, respondents are free to identify or omit personal characteristics. This approach is used because respondents may have concerns that their personal information will allow them to be identified, especially in organizations with small numbers of women or URM employees. Although this approach can introduce limitations related to missing data when analyzing the survey results, it ensures a measure of safety and respect for the respondents.
Survey Execution
We conducted the survey online in partnership with Sage Computing, a vendor experienced in conducting surveys for the ACR, from February 27 to April 26, 2018. Respondents were contacted by e-mail and directed to a unique link they could use to complete the survey. In the event of a “bad address” (ie, if an e-mail could not be successfully sent to an available address), the respondent was replaced, such that 900 e-mail invitations were successfully sent to 450 randomly selected male ACR members and to 450 randomly selected female ACR members, in keeping with the study’s inclusion and exclusion criteria. Respondents were remunerated electronically with a $50 gift card. Deidentified data were returned to MGH for subsequent analysis.
Data Analysis
We evaluated differences in responses by gender and by race/ethnicity. For each analysis, a subset of relevant survey questions was evaluated: 25 items for gender and 19 items for race/ethnicity. Items included in each respective analysis are indicated in the Appendix. Survey responses were dichotomized into “yes” or “no,” as detailed in the Appendix. When respondents indicated “prefer not to say” or “I don’t know,” such responses were grouped with missing responses.
Analysis of Differences in Responses by Gender
For the 25 survey items of interest, descriptive summaries (percentages of “yes” versus “no” responses per item) were first calculated by gender. Then, for each item, a multivariate logistic regression model was constructed in which the dichotomized response was regressed on gender and trainee status (ie, independent variables: gender [predictor], trainee status [control]; dependent variable: yes/no response). Odds ratios (ORs), their 95% confidence intervals (CIs), and associated P values were calculated for all gender effects (ie, for all associations evaluated between gender and survey item responses). To account for multiple comparisons, we calculated false discovery rate–adjusted P values (PFDR), which served as the standard for discerning significance in this study [18x18Benjamini, Y. and HY. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Stat Soc B. 1995;
57: 289–300
Google ScholarSee all References]. We used a level of significance (α) of .05.
As noted previously, respondents had the option to omit identifying characteristics and item responses. Missing data were handled in two ways: (1) itemwise complete-case analysis (primary analysis) and (2) multiple imputation (secondary analysis) [19x19Rubin, D.B. Multiple imputation for nonresponse in surveys. John Wiley,
New York; 1987
Crossref | Google ScholarSee all References]. Item-wise complete-case analysis restricted the regression analysis to those respondents who indicated their gender, trainee status, and response for each given survey item of interest. Multiple imputation used all respondents’ data by “filling in” missing demographics and/or survey items; both demographics and survey responses were imputed.
Analysis of Differences in Responses by Race and Ethnicity
For the 19 survey items of interest, descriptive summaries were calculated for five subgroups: (1) white, non-Hispanic; (2) Asian or Asian American, non-Hispanic; (3) URM, defined as individuals identifying as black or African American, Hispanic, American Indian or Alaska Native, or Native Hawaiian or other Pacific Islander; (4) more than one race (multiracial), non-Hispanic; and (5) respondents who did not report race/ethnicity information. Regression modeling was not considered a valid approach for hypothesis testing in the race/ethnicity analysis because of the small size and lack of variability in the distribution of responses of some subgroups.
Results
In total, 900 survey invitations were issued. The response rate was 51.2% (461 of 900). Table 1 includes demographic information provided by the respondents. Respondents included 235 women and 204 men; 22 individuals did not indicate their gender. The ethnic and racial composition of the respondent group was 5.6% Latino or Hispanic (26 of 461), 61.8% white (285 of 461), 20.4% Asian or Asian American (94 of 461), 3.5% black or African American (16 of 461), 0.4% American Indian or Alaska Native (2 of 461), 0.0% Native Hawaiian or other Pacific Islander (0 of 461), 5.6% multiracial (26 of 461), and 8.2% not reported (38 of 461). For our study, race/ethnicity was categorized as (1) white, non-Hispanic, 57.5% (265 of 461); (2) Asian or Asian American, non-Hispanic, 20.2% (93 of 461); (3) URM, 9.5% (44 of 461); (4) multiracial, non-Hispanic, 3.7% (17 of 461); and (5) not reported, 9.1% (42 of 461). Radiologists (attending or trainee) constituted 91.3% (421 of 461) of respondents; radiation oncologists constituted 6.3% (29 of 461). In total, 39.3% of respondents (181 of 461) were trainees. The mean age of respondents was 40.2 ± 10.4 years.
| Characteristic | n | % |
|---|---|---|
| Sex | ||
| Male | 204 | 44.3∗ |
| Female | 235 | 51.0∗ |
| Did not report†† | 22 | 4.8 |
| Race | ||
| White | 285 | 61.8 |
| Asian or Asian American | 94 | 20.4 |
| Black or African American | 16 | 3.5 |
| American Indian or Alaska Native | 2 | 0.4 |
| Native Hawaiian or Pacific Islander | 0 | 0.0 |
| More than one race | 26 | 5.6 |
| Did not report†† | 38 | 8.2 |
| Latino or Hispanic | ||
| Yes | 26 | 5.6 |
| No | 406 | 88.1 |
| Did not report†† | 29 | 6.3 |
| Specialty | ||
| Radiology | 421 | 91.3 |
| Radiation oncology | 29 | 6.3 |
| Did not report‡‡ | 11 | 2.4 |
| Academic practice | ||
| Yes | 109 | 23.6 |
| No | 137 | 29.7 |
| Did not report‡‡ | 215 | 46.6 |
Analysis of Differences in Responses by Gender
Table 2 summarizes a subset of key findings from the gender analysis; additional results are included in the Appendix. Responses to several items differed by gender, with statistical significance, when controlling for trainee status. Key findings are also summarized in the following discussion, organized by survey domain.
| Survey Item | Women | Men | Missing | OR (95% CI) | P Value | ||||
|---|---|---|---|---|---|---|---|---|---|
| Total n | % Agree (n) | Total n | % Agree (n) | Total n | % Agree (n) | Unadjusted | False Discovery Rate Adjusted | ||
| Work environment and relationships | |||||||||
| I am treated with respect by patients and their family members | 235 | 52.3 (123) | 204 | 65.2 (133) | 22 | 59.1 (13) | 0.59 (0.40-0.86) | .007 | .021 |
| In my career, having at least one mentor that is the same gender as I am is important to me | 235 | 71.9 (169) | 204 | 34.3 (70) | 22 | 22.7 (5) | 4.90 (3.26-7.36) | <.001 | <.001 |
| The opportunity to interact directly with patients is important to me | 235 | 68.5 (161) | 204 | 72.1 (147) | 22 | 59.1 (13) | 0.84 (0.56-1.27) | .418 | .564 |
| Profession attributes | |||||||||
| Radiation exposure is a drawback of my profession | 235 | 41.7 (98) | 204 | 52.9 (108) | 19 | 52.6 (10) | 0.64 (0.44-0.93) | .019 | .054 |
| The opportunity to apply principles of physics and mathematics at work is important to me | 235 | 31.9 (75) | 204 | 51.5 (105) | 19 | 42.1 (8) | 0.44 (0.30-0.65) | <.001 | <.001 |
| The opportunity to work with state-of-the-art imaging, therapeutic, and/or informatics technology is important to me | 234 | 91.0 (213) | 203 | 91.6 (186) | 19 | 100.0 (19) | 0.93 (0.47-1.81) | .824 | .859 |
| I believe that I am paid fairly compared with others of similar experience in my department or practice | 233 | 73.4 (171) | 203 | 73.4 (149) | 19 | 68.4 (13) | 1.00 (0.65-1.53) | .998 | .998 |
| Opportunities for leadership and career advancement are less available to female physicians than to male physicians in my department or practice | 234 | 38.9 (91) | 204 | 8.3 (17) | 19 | 5.3 (1) | 7.00 (3.99-12.30) | <.001 | <.001 |
| Culture of diversity and respect | |||||||||
| My department or practice fosters a culture of respect for all people | 235 | 83.4 (196) | 204 | 90.7 (185) | 17 | 100.0 (17) | 0.52 (0.29-0.93) | .027 | .067 |
| My department or practice has a fair and transparent maternity leave policy | 187 | 28.3 (53) | 164 | 10.4 (17) | 14 | 7.1 (1) | 3.42 (1.88-6.21) | <.001 | <.001 |
| My department or practice has a comfortable and private place for women to breastfeed or pump | 192 | 39.6 (76) | 146 | 13.7 (20) | 12 | 16.7 (2) | 4.13 (2.37-7.19) | <.001 | <.001 |
| In the past year, in your work as a radiologist or radiation oncologist, have you personally felt you were treated unfairly or with disrespect… | |||||||||
| because of your gender? | 235 | 50.6 (119) | 204 | 5.4 (11) | 15 | 13.3 (2) | 18.00 (9.29-34.86) | <.001 | <.001 |
| because of your age? | 234 | 23.9 (56) | 203 | 11.3 (23) | 15 | 6.7 (1) | 2.46 (1.45-4.18) | .001 | .003 |
| Career recommendation | |||||||||
| Thinking about all of the experiences you have had while working as a radiologist or radiation oncologist, would you recommend your choice of profession to qualified medical students | 235 | 96.6 (227) | 204 | 95.1 (194) | 19 | 89.5 (17) | 1.56 (0.58-4.20) | .38 | .558 |
Note: Items with responses that differ significantly, between women and men, are highlighted in boldface type. Results for additional items with responses that did not differ significantly between men and women are included in the Appendix. CI = confidence interval; OR = odds ratio.
Work Environment and Relationships
Women were less likely than men to report that they were “treated with respect by patients and their family members” (52.3% versus 65.2%; OR, 0.59; 95% CI, 0.40-0.86; PFDR = .021). Women were more likely than men to respond that a same-gender mentor was important (71.9% versus 34.3%; OR, 4.90; 95% CI, 3.26-7.36; PFDR < .001).
We did not detect a significant difference between women and men in their perceived importance of the “opportunity to interact directly with patients” (68.5% versus 72.1%; OR, 0.84; 95% CI, 0.56-1.27; PFDR = .564).
Profession Attributes
Women were less likely than men to indicate that “the opportunity to apply principles of physics and mathematics at work” was important (31.9% versus 51.5%; OR, 0.44; 95% CI, 0.30-0.65; PFDR < .001). Women were more likely to indicate that “opportunities for leadership and career advancement [were] less available” to female physicians in their workplace (38.9% versus 8.3%; OR, 7.00; 95% CI, 3.99-12.30; PFDR < .001).
There was evidence to suggest that women may be less likely to perceive radiation exposure as a professional drawback; however, the observed difference did not attain the false discovery rate significance threshold (41.7% versus 52.9%; OR, 0.64; 95% CI, 0.44-0.93; Punadjusted = .019, PFDR = .054).
We did not detect a significant difference between women and men in their perceptions regarding the importance of “the opportunity to work with state-of-the-art imaging, therapeutic, and/or informatics technology” (91.0% versus 91.6%; OR, 0.93; 95% CI, 0.47-1.81; PFDR = .859) or fair compensation (73.4% versus 73.4%; OR, 1.00; 95% CI, 0.65-1.53; PFDR = .998).
Culture of Diversity and Respect
The largest discrepancy between women and men was observed when respondents were asked about unfair or disrespectful treatment in the workplace. Women were substantially more likely than men to respond that in the past year, they have been “treated unfairly or with disrespect” because of their gender (50.6% versus 5.4%; OR, 18.00; 95% CI, 9.29-34.86; PFDR < .001). Women were also more likely to respond that they have received such treatment because of their age (23.9% versus 11.3%; OR, 2.46; 95% CI, 1.45-4.18; PFDR = .003).
There was evidence to suggest that women may be less likely to agree with the statement that their workplace “fosters a culture of respect for all people”; however, the observed difference did not attain the false discovery rate significance threshold (83.4% versus 90.7%; OR, 0.52; 95% CI, 0.29-0.93; Punadjusted = .027, PFDR = .067).
Women were more likely than men to indicate “a fair and transparent maternity leave policy” in their workplace (28.3% versus 10.4%; OR, 3.42; 95% CI, 1.88-6.21; PFDR < .001). Women were also more likely to indicate that their workplace “has a comfortable and private place for women to breastfeed or pump” (39.6% versus 13.7%; OR, 4.13; 95% CI, 2.37-7.19; PFDR < .001). However, interpretation of differential responses was limited by the lower response rates for these items compared with others (Table 2).
Career Recommendation
We did not detect a significant difference when comparing women and men in whether they would recommend their profession to medical students (96.6% versus 95.1%; OR, 1.56; 95% CI, 0.58-4.20; PFDR = .558).
Secondary Analysis (Imputation)
There were no differences in determinations of statistical significance between the itemwise complete-case analysis and the imputation analysis, except regarding one item related to radiation exposure. In the imputation analysis, gender-based differences regarding the perception of radiation exposure as a professional drawback achieved statistical significance, suggesting that women would be less likely than men to perceive radiation exposure as a professional drawback (OR, 0.63; 95% CI, 0.43-0.91; PFDR = .041). This was considered a minor difference in results. Either way, the findings indicate that radiation exposure was not perceived to be a greater drawback by women (41.7%) compared with men (52.9%).
Analysis of Differences in Responses by Race and Ethnicity
Table 3 summarizes results of the descriptive analysis focused on race and ethnicity. Key findings are addressed in the following discussion. Importantly, unlike in the gender analysis, comparisons between subgroups in the race/ethnicity analysis should be considered purely descriptive in nature (ie, when a given proportion is stated to be higher or lower than another, the comparison was not evaluated for statistical significance). See the “Data Analysis” subsection of “Methods” for further details.
| Survey Item | White Non-Hispanic | Asian Non-Hispanic | URM | Multiracial Non-Hispanic | Missing | |||||
|---|---|---|---|---|---|---|---|---|---|---|
| Total n | % Agree (n) | Total n | % Agree (n) | Total n | % Agree (n) | Total n | % Agree (n) | Total n | % Agree (n) | |
| Work environment and relationships | ||||||||||
| I am treated with respect… | ||||||||||
| at work by my supervisor | 265 | 90.6 (240) | 93 | 89.2 (83) | 44 | 97.7 (43) | 17 | 94.1 (16) | 42 | 78.6 (33) |
| by my colleagues and co-workers | 264 | 97.3 (257) | 93 | 96.8 (90) | 44 | 93.2 (41) | 17 | 94.1 (16) | 42 | 90.5 (38) |
| by patients and their family members | 265 | 57.7 (153) | 93 | 51.6 (48) | 44 | 65.9 (29) | 17 | 94.1 (16) | 42 | 54.8 (23) |
| I have one or more colleagues or co-workers whom I trust to give me advice and support | 264 | 97.3 (257) | 93 | 96.8 (90) | 44 | 88.6 (39) | 17 | 94.1 (16) | 42 | 83.3 (35) |
| I have a mentor in my department or practice who supports my career development | 265 | 58.9 (156) | 93 | 61.3 (57) | 44 | 54.5 (24) | 17 | 70.6 (12) | 42 | 61.9 (26) |
| I have a mentor in my profession who supports my career development | 265 | 63.0 (167) | 93 | 68.8 (64) | 44 | 61.4 (27) | 17 | 70.6 (12) | 42 | 57.1 (24) |
| In my career, having at least one mentor that is the same race or ethnicity as I am is important to me | 265 | 18.5 (49) | 93 | 32.3 (30) | 44 | 40.9 (18) | 16 | 18.8 (3) | 42 | 19.0 (8) |
| The opportunity to interact directly with patients is important to me | 265 | 69.8 (185) | 93 | 72.0 (67) | 44 | 72.7 (32) | 17 | 70.6 (12) | 42 | 59.5 (25) |
| Profession attributes | ||||||||||
| I believe that I am paid fairly compared with others of similar experience in my department or practice | 262 | 75.2 (197) | 93 | 66.7 (62) | 44 | 75.0 (33) | 17 | 76.5 (13) | 39 | 71.8 (28) |
| My department or practice makes deliberate efforts to recruit women and under-represented minority physicians | 264 | 40.5 (107) | 93 | 40.9 (38) | 44 | 31.8 (14) | 17 | 35.3 (6) | 39 | 59.0 (23) |
| Opportunities for leadership and career advancement are less available to under-represented minority (URM) physicians than to other physicians in my department or practice | 263 | 13.3 (35) | 93 | 25.8 (24) | 44 | 31.8 (14) | 17 | 5.9 (1) | 39 | 12.8 (5) |
| Culture of diversity and respect | ||||||||||
| My colleagues and co-workers are respectful of staff and employees who come from diverse backgrounds | 265 | 92.1 (244) | 93 | 88.2 (82) | 44 | 81.8 (36) | 17 | 100 (17) | 37 | 89.2 (33) |
| My department or practice fosters a culture of respect for all people | 265 | 88.7 (235) | 93 | 83.9 (78) | 44 | 81.8 (36) | 17 | 88.2 (15) | 37 | 91.9 (34) |
| In my department or practice, I feel comfortable sharing information about my personal identity, culture, religion, or background with my colleagues and co-workers | 264 | 79.9 (211) | 93 | 81.7 (76) | 44 | 63.6 (28) | 17 | 82.4 (14) | 37 | 83.8 (31) |
| My colleagues or co-workers… | ||||||||||
| sometimes tell offensive jokes or stories, or make insulting comments, about race or ethnicity∗ | 265 | 20.0 (53) | 93 | 19.4 (18) | 44 | 11.4 (5) | 17 | 5.9 (1) | 37 | 8.1 (3) |
| sometimes send or post material that is offensive to me (such as cartoons, memes, or images) on social media∗ | 264 | 3.4 (9) | 93 | 7.5 (7) | 44 | 2.3 (1) | 17 | 0.0 (0) | 37 | 2.7 (1) |
| have made inappropriate remarks to me about my body, physical appearance, or clothing∗ | 265 | 10.2 (27) | 93 | 9.7 (9) | 44 | 2.3 (1) | 17 | 5.9 (1) | 37 | 8.1 (3) |
| In the past year, in your work as a radiologist or radiation oncologist, have you personally felt you were treated unfairly or with disrespect because of your race and/or ethnicity? | 265 | 2.6 (7) | 93 | 20.4 (19) | 43 | 27.9 (12) | 17 | 0.0 (0) | 35 | 5.7 (2) |
| Career recommendation | ||||||||||
| Thinking about all of the experiences you have had while working as a radiologist or radiation oncologist, would you recommend your choice of profession to qualified medical students | 265 | 95.8 (254) | 93 | 95.7 (89) | 44 | 100 (44) | 17 | 94.1 (16) | 39 | 89.7 (35) |
Note: Items with responses with the greatest differences between URM and white non-Hispanic individuals’ responses are highlighted in boldface type (see “Results” for further details). URM = underrepresented minority (inclusive of all individuals identifying as Latino or Hispanic, black or African American, or American Indian or Alaska Native).
Google ScholarSee all References]; most survey items drew from MGH workforce surveys led by Dr. Karen Donelan [15x15Donelan K, Bareto E, Michael C, Bonner S. 2016 Diversity workforce culture survey, final report. Boston: Massachusetts General Hospital. March, 2017.
Google ScholarSee all References, 16x16Donelan K, Michael C. MGH diversity metrics project: Supporting effective strategies to develop a diverse workforce, final report, Boston: Massachusetts General Hospital. March, 2017.
Google ScholarSee all References] or were contributed by the expert team (see “Methods” for further details).
Work Environment and Relationships
A higher proportion of URMs compared with white non-Hispanic respondents indicated that “having at least one mentor that [was] the same race or ethnicity” was important (40.9% versus 18.5%). The proportion of affirmative responses for Asian or Asian American non-Hispanic respondents was 32.3% and for multiracial non-Hispanic respondents was 18.8%.
The proportion of affirmative responses regarding the importance of the “opportunity to interact directly with patients” was similar when comparing URM (72.7%) with white non-Hispanic (69.8%), Asian or Asian American non-Hispanic (72.0%), and multiracial non-Hispanic (70.6%) respondents.
Profession Attributes
A higher proportion of URMs compared with white non-Hispanic respondents indicated that “opportunities for leadership and career advancement [were] less available to URM physicians than other physicians” in their workplace (31.8% versus 13.3%). The proportion of affirmative responses for Asian or Asian American non-Hispanic respondents was 25.8% and for multiracial non-Hispanic respondents was 5.9%.
The proportion of affirmative responses regarding fair compensation was similar when comparing URMs (75.0%) with white non-Hispanic (75.2%) and multiracial non-Hispanic respondents (76.5%); it was slightly lower among Asian or Asian American non-Hispanic respondents (66.7%).
Culture of Diversity and Respect
A lower proportion of URMs compared with white non-Hispanic respondents reported being comfortable sharing information about their “personal identity, culture, religion or background” with colleagues (63.6% versus 79.9%). The proportion of affirmative responses for Asian or Asian American non-Hispanic respondents was 81.7% and for multiracial non-Hispanic respondents was 82.4%.
A substantially higher proportion of URMs compared with white non-Hispanic respondents indicated that in the past year, they have been “treated unfairly or with disrespect” because of their race and/or ethnicity (27.9% versus 2.6%). The proportion of affirmative responses for Asian or Asian American non-Hispanic respondents was 20.4% and for multiracial non-Hispanic respondents was 0%.
The proportion of affirmative responses regarding whether one’s workplace “fosters a culture of respect for all people” was similar when comparing URMs (81.8%) with white non-Hispanic (88.7%), Asian or Asian American non-Hispanic (83.9%), and multiracial non-Hispanic (88.2%) respondents.
We also found that the proportion of respondents who indicated that their “colleagues or co-workers sometimes tell offensive jokes or stories, or make insulting comments, about race or ethnicity” was high in multiple subgroups: 20.0% among white non-Hispanic respondents, 19.4% among Asian or Asian American non-Hispanic respondents, and 11.4% among URMs.
Career Recommendation
The proportion of affirmative responses regarding the recommendation of one’s profession to medical students was similar when comparing URMs (100%) with white non-Hispanic (95.8%), Asian or Asian American non-Hispanic (95.7%), and multiracial non-Hispanic (94.1%) respondents.
Discussion
We conducted a national survey of ACR physician members to identify potential barriers to diversity in the physician workforce. We focused on describing gender- and race/ethnicity-based differences in physicians’ responses to survey questions about their workplace environment, relationships, and culture. In each analysis, we identified unfair or disrespectful treatment as a dominant concern. Women were significantly more likely than men to report unfair or disrespectful treatment attributed to their gender and age, by patients and families, and in terms of opportunities for career advancement. Compared with white non-Hispanic respondents, a higher proportion of URMs reported such treatment attributed to their race/ethnicity and in terms of opportunities for career advancement. A lower proportion of URMs reported being comfortable sharing information about their identity and background with colleagues. Such differences point to a problem in the culture of diversity in our professions. Addressing this problem is likely to be critically important for improving workforce diversity.
Despite these observations, more than four in five respondents in all gender and race/ethnicity subgroups agreed with the statement that their department or practice “fosters a culture of respect for all people.” Taken together, the findings suggest that unfair or disrespectful treatment of women and URMs is prevalent despite workplace efforts to create environments that discourage such treatment. The results of our study signal the need for every individual to contribute actively to building a culture of diversity and respect. Across multiple race/ethnicity subgroups—including white non-Hispanic, Asian or Asian American non-Hispanic, and URM respondents—a high proportion of physicians reported that their colleagues sometimes made offensive comments about race or ethnicity. Those who are the target of such behavior should be empowered to speak out against it; equally, those who witness it, but are not the target, must take the initiative to speak out.
Significant differences were not observed between women and men regarding the importance of patient interactions, the importance of the opportunity to work with state-of-the-art technology, issues related to radiation exposure, and fairness in compensation. These findings suggest that preconceptions and stereotypes regarding women’s relatively higher desire for patient interaction, dislike of technology, and fear of radiation exposure lack validity. Women were less likely to agree that “the opportunity to apply principles of physics and mathematics at work” was important, suggesting that compared with men, other factors and priorities may have played a greater role at the time of their career selection.
Another important finding of this study pertained to mentorship. Specifically, women were significantly more likely than men to prefer a same-gender mentor. A higher proportion of URMs and non-Hispanic Asians preferred a mentor of the same race/ethnicity relative to other subgroups. In recent years, a conceptual shift from mentorship to “sponsorship” has been encouraged as a mechanism to more effectively advance the careers of female and URM leaders [20x20Ayyala, M.S., Skarupski, K., Bodurtha, J.N., Gonzalez-Fernandez, M., Ishii, L.E., Fivush, B. et al. Mentorship is not enough: exploring sponsorship and its role in career advancement in academic medicine. Acad Med. 2019;
94: 94–100
Crossref | PubMed | Scopus (13) | Google ScholarSee all References, 21x21Deitte, L.A., McGinty, G.B., Canon, C.L., Omary, R.A., Johnson, P.T., and Slanetz, P.J. Shifting from mentorship to sponsorship—a game changer!. J Am Coll Radiol. 2019;
16: 498–500
Abstract | Full Text | Full Text PDF | PubMed | Scopus (2) | Google ScholarSee all References, 22x22Gottlieb, A.S. and Travis, E.L. Rationale and models for career advancement sponsorship in academic medicine: the time is here; the time is now. Acad Med. 2018;
93: 1620–1623
Crossref | PubMed | Scopus (11) | Google ScholarSee all References]. A sponsor not only affords career guidance but also provides and secures key opportunities for professional advancement [20x20Ayyala, M.S., Skarupski, K., Bodurtha, J.N., Gonzalez-Fernandez, M., Ishii, L.E., Fivush, B. et al. Mentorship is not enough: exploring sponsorship and its role in career advancement in academic medicine. Acad Med. 2019;
94: 94–100
Crossref | PubMed | Scopus (13) | Google ScholarSee all References, 21x21Deitte, L.A., McGinty, G.B., Canon, C.L., Omary, R.A., Johnson, P.T., and Slanetz, P.J. Shifting from mentorship to sponsorship—a game changer!. J Am Coll Radiol. 2019;
16: 498–500
Abstract | Full Text | Full Text PDF | PubMed | Scopus (2) | Google ScholarSee all References, 22x22Gottlieb, A.S. and Travis, E.L. Rationale and models for career advancement sponsorship in academic medicine: the time is here; the time is now. Acad Med. 2018;
93: 1620–1623
Crossref | PubMed | Scopus (11) | Google ScholarSee all References]. Such opportunities can be more difficult to obtain passively for women and URMs. Given the recognized value of mentors and sponsors to professional success, the observed preference among many women and URMs in our study for mentors of the same gender or race/ethnicity points to the importance of diversity among mentors within a practice or department. Ensuring such diversity may increase the likelihood of successful mentoring relationships for women and URMs in our fields.
Our study had important limitations. First, we did not capture the attitudes of medical students, a primary target population of related work in the past [11x11Fielding, J.R., Major, N.M., Mullan, B.F., Neutze, J.A., Shaffer, K., Wilcox, C.B. et al. Choosing a specialty in medicine: female medical students and radiology. AJR Am J Roentgenol. 2007;
188: 897–900
Crossref | PubMed | Scopus (49) | Google ScholarSee all References, 12x12Roubidoux, M.A., Packer, M.M., Applegate, K.E., and Aben, G. Female medical students’ interest in radiology careers. J Am Coll Radiol. 2009;
6: 246–253
Abstract | Full Text | Full Text PDF | PubMed | Scopus (52) | Google ScholarSee all References, 14x14Zener, R., Lee, S.Y., Visscher, K.L., Ricketts, M., Speer, S., and Wiseman, D. Women in radiology: exploring the gender disparity. J Am Coll Radiol. 2016;
13 (344-50.e1)
Abstract | Full Text | Full Text PDF | PubMed | Scopus (29) | Google ScholarSee all References]. Instead, our survey was directed at physicians already practicing within radiology and radiation oncology; moreover, most respondents (across subgroups) indicated that they would recommend their profession to medical students. As a result, we do not have direct evidence to indicate that, by addressing issues regarding the culture of diversity (unfair or disrespectful treatment), our specialties will be more likely to recruit women and URMs. Nevertheless, it is logical to assume that by fostering a stronger culture of diversity in our professions, the likelihood of achieving greater diversity in our workforce in future generations will be higher. Moreover, addressing such issues is critically important for the success and advancement of women and URMs who are already in our fields.
Second, our race/ethnicity analysis was underpowered. For this reason, our observations of proportionate differences in responses between subgroups are based on descriptive judgements rather than inferential methods (ie, formal hypothesis testing). Unlike gender data, race/ethnicity data are not available for the vast majority of ACR members because these measures only recently became available for members to self-report. This lack of data prevented oversampling of candidate URM respondents. Identifying ways to oversample racially and ethnically diverse physician subgroups in radiology and radiation oncology will be helpful for future, similar investigations. However, it should be recognized that oversampling is a research method and not a policy goal: the best solution to capturing the critically important perspectives of different subgroups is to build a diverse workforce.
Third, we did not compare responses across a spectrum of physician groups (ie, those outside radiology and radiation oncology). Physicians in different medical specialties that attract more women and URMs may respond similarly when given the same survey.
Fourth, our respondent pool was constituted by both radiologists and radiation oncologists but was dominated by radiologists. Therefore, our results are most applicable to practicing radiologists.
Fifth, many respondents did not complete all items. In general, the response rate for our survey (51.2%) can be considered relatively high for a physician survey, but missing data still can introduce biases in the survey findings if nonrespondents are likely to respond differently than individuals who completed a given item [23x23Burns, K.E., Duffett, M., Kho, M.E. et al. A guide for the design and conduct of self-administered surveys of clinicians. CMAJ. 2008;
179: 245–252
Crossref | PubMed | Scopus (492) | Google ScholarSee all References].
Conclusions
In a national survey of radiologists and radiation oncologists, we found that women and URMs are substantially more likely to experience unfair or disrespectful treatment in the workplace. Our results suggest that workplace culture could represent a primary barrier to the recruitment, retention, and advancement of these individuals. The formulation of strategies to identify and address sources of this problem, at granular, local levels, should represent a top priority in our professions. Such efforts are likely to represent a critical step toward reducing barriers to building a diverse physician workforce in the years to come.
Take-Home Points
- ▪
In a national survey of radiologists and radiation oncologists, we found that women and URMs are substantially more likely to experience unfair or disrespectful treatment in the workplace.
- ▪
Problems in the culture of diversity and respect may constitute a primary barrier to the recruitment, retention, and advancement of women and URMs.
- ▪
Developing strategies to build a strong culture of diversity and respect in the workplace should represent a top priority in our professions.
Additional Resources
Download
(.04 MB
)
Appendix
References
- 1Lett, L.A., Orji, W.U., and Sebro, R. Declining racial and ethnic representation in clinical academic medicine: a longitudinal study of 16 US medical specialties. PLoS One. 2018; 13: e0207274
- 2Chapman, C.H., Hwang, W.T., Both, S., Thomas, C.R. Jr., and Deville, C. Current status of diversity by race, Hispanic ethnicity, and sex in diagnostic radiology. Radiology. 2014; 270: 232–240
- 3Chapman, C.H., Hwang, W.T., and Deville, C. Diversity based on race, ethnicity, and sex, of the US radiation oncology physician workforce. Int J Radiat Oncol Biol Phys. 2013; 85: 912–918
- 4Association of American Medical Colleges. Diversity in medical education: facts and figures 2016.. (Available at:)http://www.aamcdiversityfactsandfigures2016.org/March 27, 2019
- 5Baker, S.R., Barry, M., Chaudhry, H., and Hubbi, B. Women as radiologists: are there barriers to entry and advancement?. J Am Coll Radiol. 2006; 3: 131–134
- 6Lewis, R.S., Bhargavan, M., and Sunshine, J.H. Women radiologists in the United States: results from the American College of Radiology’s 2003 survey. Radiology. 2007; 242: 802–810
- 7Jalilianhasanpour, R., Charkhchi, P., Mirbolouk, M., and Yousem, D.M. Underrepresentation of women on radiology editorial boards. J Am Coll Radiol. 2019; 16: 115–120
- 8Piper, C.L., Scheel, J.R., Lee, C.I., and Forman, H.P. Representation of women on radiology journal editorial boards: a 40-year analysis. Acad Radiol. 2018; 25: 1640–1645
- 9Jena, A.B., Olenski, A.R., and Blumenthal, D.M. Sex differences in physician salary in US public medical schools. JAMA Intern Med. 2016; 176: 1294–1304
- 10Kapoor, N., Blumenthal, D.M., Smith, S.E., Ip, I.K., and Khorasani, R. Sex differences in radiologist salary in U.S. public medical schools. AJR Am J Roentgenol. 2017; 209: 953–958
- 11Fielding, J.R., Major, N.M., Mullan, B.F., Neutze, J.A., Shaffer, K., Wilcox, C.B. et al. Choosing a specialty in medicine: female medical students and radiology. AJR Am J Roentgenol. 2007; 188: 897–900
- 12Roubidoux, M.A., Packer, M.M., Applegate, K.E., and Aben, G. Female medical students’ interest in radiology careers. J Am Coll Radiol. 2009; 6: 246–253
- 13Frank, E. and Vydareny, K. Characteristics of women radiologists in the United States. AJR Am J Roentgenol. 1999; 173: 531–536
- 14Zener, R., Lee, S.Y., Visscher, K.L., Ricketts, M., Speer, S., and Wiseman, D. Women in radiology: exploring the gender disparity. (344-50.e1)J Am Coll Radiol. 2016; 13
- 15Donelan K, Bareto E, Michael C, Bonner S. 2016 Diversity workforce culture survey, final report. Boston: Massachusetts General Hospital. March, 2017.
- 16Donelan K, Michael C. MGH diversity metrics project: Supporting effective strategies to develop a diverse workforce, final report, Boston: Massachusetts General Hospital. March, 2017.
- 17Association of American Universities. Climate Survey on Sexual Assault and Sexual Misconduct. (Available at:)
- 18Benjamini, Y. and HY. Controlling the false discovery rate: a practical and powerful approach to multiple testing. J R Stat Soc B. 1995; 57: 289–300
- 19Rubin, D.B. Multiple imputation for nonresponse in surveys. John Wiley, New York; 1987
- 20Ayyala, M.S., Skarupski, K., Bodurtha, J.N., Gonzalez-Fernandez, M., Ishii, L.E., Fivush, B. et al. Mentorship is not enough: exploring sponsorship and its role in career advancement in academic medicine. Acad Med. 2019; 94: 94–100
- 21Deitte, L.A., McGinty, G.B., Canon, C.L., Omary, R.A., Johnson, P.T., and Slanetz, P.J. Shifting from mentorship to sponsorship—a game changer!. J Am Coll Radiol. 2019; 16: 498–500
- 22Gottlieb, A.S. and Travis, E.L. Rationale and models for career advancement sponsorship in academic medicine: the time is here; the time is now. Acad Med. 2018; 93: 1620–1623
- 23Burns, K.E., Duffett, M., Kho, M.E. et al. A guide for the design and conduct of self-administered surveys of clinicians. CMAJ. 2008; 179: 245–252
Funding for this study was provided by the ACR. Dr Pandharipande, Ms Lietz, Ms Neal, and Dr Donelan received funding from the ACR through an institutional contract (with Massachusetts General Hospital) for the conduct of this work. Ms Bansal received funding from the ACR through a contract with Sage Computing for the conduct of this work. The design and conduct of the study, the analysis and interpretation of the data, and the drafting and approval of the manuscript were under the full responsibility of the author team. The collection of data was executed by Sage Computing. After data collection, a deidentified data set was returned to the authors for analysis. Dr Pandharipande has received research funding from the Medical Imaging and Technology Alliance (outside the submitted work). Dr Sadigh has received an Association of University Radiologists GE Radiology Research Academic Fellowship and an RSNA Seed Grant (outside the submitted work). Ms Bansal received personal fees from the ACR during the conduct of this study. All other authors state that they have no conflict of interest related to the material discussed in this article. Drs Macura and Donelan contributed equally to this report.
