Understanding Equity, Diversity, and Inclusion Within Canadian Radiation Oncology Training Programs: Experiences of Residents and Fellows

1. Introduction

Equity, diversity, and inclusion principles (EDI) have become embedded within strategic priorities in postgraduate medicine, academic medical institutions, and professional organizations (1-6). Promoting a diverse workforce in healthcare has the potential to enhance cultural competency, reduce health disparities, and improve patient outcomes (7). Furthermore, it is important that the healthcare workforce represents the population they serve. Within the United States (US) radiation oncology (RO) workforce, the proportion of women and underrepresented minority groups is less than that of the general population (8). Studies have shown that cancer patient outcomes are worse for under-represented populations and therefore, efforts to promote diversity within medical schools and training programs have been suggested to improve patient care (9, 10).

The American Society for Radiation Oncology (ASTRO) has established a Health Equity, Diversity and Inclusion (HEDI) Council and the American Society of Clinical Oncology (ASCO) has established priorities for diversification as a mechanism for addressing health disparities and improving cancer care and research (1, 2). Additionally, within postgraduate medicine, the Association of Residents in Radiation Oncology (ARRO) established an Equity and Inclusion Subcommittee (EISC) in 2020 with an aim to create a safe space for all trainees, promote diversity, equity, inclusion and social justice and assess diversity in the workforce (6). Furthermore, the Canadian Association of Radiation Oncology (CARO) established an Equity, Diversity, Inclusion and Indigenous Care standing committee in 2023 to promote awareness and support initiatives that improves diversification within the specialty to ensure the provision of high-quality, inclusive patient care (11).

Previously, most RO EDI research has been conducted in the US and has focused primarily on gender, specifically the underrepresentation of women (12, 13). However, recently more comprehensive studies evaluating and exploring EDI within RO and medical physics have been conducted. The European Society for Radiotherapy and Oncology (ESTRO) conducted a collaborative study to collect benchmark data on EDI and workforce engagement (14). They found overall lower inclusion scores compared to US data and concluded that initiatives to improve EDI and engagement are needed. However, of the 812 respondents, only 12.6% were in-training and unfortunately, their data was not reported separately. Furthermore, within the field of medical physics, the Canadian Organization of Medical Physics (COMP) in collaboration with the American Association of Medical Physics (AAMP) recently published their faculty climate survey results (15, 16). Although most respondents reported positive workplace experiences, they identified equity-lacking groups, including women, racialized minorities, Indigenous peoples, and people with disabilities. Recommendations to improve EDI and accessibility included conducting routine EDI climate surveys, creating EDI training opportunities for society members, and conducting an EDI climate survey for trainees (15, 16).

EDI data regarding RO trainees is emerging. In 2022, the ARRO EISC surveyed RO residents regarding diversity, equity, inclusion, and belonging. With a 42% response rate, they found significant racial, ethics and gender differences in the areas of support, mentorship, inclusion and bias (17). They recommended that RO training programs collaborate to focus on initiatives directed towards residents from under-represented groups to foster a supportive and unbiased learning environment, which in turn, should facilitate recruitment and retention of a diverse workforce in the future.

There is limited data regarding the demographics and experiences of RO trainees worldwide, and Canadian-specific data is lacking. In the context of interest in this topic amongst ROs generally in Canada, simultaneously a similar survey was administered to RO faculty, and we surveyed Canadian trainees. The objectives of this study were to understand the sociodemographic and lived experiences of RO trainees in Canada to bring awareness to any gaps that exist and propose EDI strategies to improve representation within the specialty to enhance clinical care.

2. Materials and Methods

Study Population and Survey Measures

Following a comprehensive literature review of other published EDI surveys and international input from content and methodological experts, two questionnaires were created in tandem to collect cross-sectional data for Canadian Radiation Oncology trainees and faculty (14-16).

Prior to distribution, the surveys underwent cognitive pre-testing, pilot testing, and iterative revisions by the study team until consensus was reached. The final trainee survey included 59 items: Socio-demographics (n=24), Training Perceptions (n=5), Mentorship (n=8), Experiences of Discrimination/Harassment (n=19), and open-ended questions (n=3). Bilingual surveys were created to encourage survey participation in either of Canada’s official languages (Supplemental Materials). Description and results of the faculty survey is the topic of a separate manuscript (18).

The survey was electronically distributed via RedCAP July and December 2023. The study population consisted of RO residents and fellows currently enrolled in a Canadian training program. Residency training in Canada is 5 years (PGY 1-5) and applicants apply through the Canadian Resident Matching Service (CaRMS). Fellows are trainees specializing in a particular field after completing a RO residency. The initial invite and three email reminders were sent to all RO trainees (n=173) through Program Directors and to RO trainee members of CARO via their email registry resource. This study was approved by our local research ethic board (***).

Data Analysis

Descriptive statistics summarized responses. Comparisons between demographic groups (gender, race, level of training, location of training) were performed using Chi-square and Fisher’s exact tests. Univariable and multivariate logistic regression analyses were performed to identify factors associated with the outcome of considering transferring to a different training program. Thematic analysis was performed on open-ended items (19). Recurring responses were grouped into categories and ultimately themes were developed that encapsulate the findings. Themes were reviewed by collaborators for consensus and approval.

3. Results

This section may be divided by subheadings. It should provide a concise and precise description of the experimental results, their interpretation, as well as the experimental conclusions that can be drawn.

3.1. Survey Response

Using data from the 2023-2024 Canadian Post-MD Education Registry (CAPER), we identified 115 residents and 58 fellows (173 trainees) enrolled in Canadian RO training programs. Among these, 98 responded to the survey (57% response rate) (20).

3.2. Respondent Demographics

Complete self-reported respondent demographics are highlighted in Table 1. Regarding level of training, 29% (28/98) were fellows and 65% (64/98) were residents with all postgraduate years represented. Of resident respondents, most were Canadian Medical Graduates (CMGs) (61/64, 95%), whereas most fellow respondents were International Medical Graduates (IMGs) (24/28, 86%). Regarding gender identity, 52% (51/98) identified as men, 42% (41/98) as women and 3% (3/98) as non-binary. Gender distribution according to level of training was similar (p=0.42). Twenty-eight respondents (28/98, 29%) reported their primary was different than English/French. Most respondents (61/98, 62%) reported that both of their parents/guardians had a college/university degree.

Most respondents (61/98, 62%) identified as a racialized minority, with the most common self-reported ethnicities being Chinese (16/98, 16%), South Asian (15/98, 15%), and Arab (10/98, 10%). Gender distribution in Caucasian and racialized groups were similar (p=0.57). Fellow respondents were more likely to self-identify as a racialized minority compared to resident respondents (82% vs. 53%, p=0.016). Regarding sexual orientation, 10% (10/98) identified as a sexual minority (2SLGBTIA+) (21). When trainees were asked if they had a disability, 4% (4/98) responded “yes,” with a mental health/emotional disability being most common.

Having a career with strong research or education components was desired/strongly desired by 31% (28/98) and 87% (79/98) of respondents, respectively. In addition to having a medical degree, 32% (31/98) reported having a Masters/PhD/equivalent with 14% (13/98) currently pursuing an additional postgraduate degree. Half of the respondents (50%) reported being an author on 1-5 peer reviewed publications and 6% (5/98) reported being an author on greater than 25 publications.

Respondent demographics were compared to CAPER and Canadian census data, when available. Our study population was representative of CAPER data with respect to level of training (e.g. resident vs. fellow; p=0.80) and representative of CAPER and Canadian census data with respect to gender identity (p=0.24 compared to CAPER and p=0.39 compared to census). Racialized groups were over-represented in our study compared to Canadian census data (62% vs. 27%, p<0.001). This remained true when residents were analysed separately (53% vs. 27%, p<0.001). Black respondents were representative of the Canadian population (3% vs. 4.3%, p=0.55); however, Indigenous respondents were under-represented (1% vs. 4.9%, p=0.007).

3.3. Training Perceptions and Mentorship

Table 2 summarizes responses regarding training perceptions and mentorship. The majority reported program satisfaction (83%, 78/94) and a “culture of respect” (69%; 64/93). However, 40% (37/92) reported they have considered switching to a different training program at least once and 74% (69/93) reported feeling regret about deciding to become a physician at least once. When asked if “People like myself are under-represented in my training program,” 28% (25/90) responded “yes,” and 15% (13/88) responded “yes” to “There are equity and diversity obstacles for entry in my radiation oncology training program.”

Most respondents (58/91, 64%) reported that a formal mentorship program existed within their training program, with 73% (66/91) reporting having at least one faculty mentor and 81% (74/91) turning to peers for mentorship. However, 21% (19/91) reported it was difficult/very difficult to identify a mentor. The most common topics discussed with a mentor included research (53%), clinical work (52%), and career planning (43%). As residents progressed in their training, they reported increased mentorship discussions in career planning (p=0.005) and fellowships (p=0.003). Most respondents (67/91, 75%) were satisfied/very satisfied with their current mentorship and 40% (36/91) agreed/strongly agreed with the statement, “It is important I have a mentor with similar demographic characteristics to me.”

Univariable logistic regression analysis was used to identify factors associated with considering transferring to a different training program and included dissatisfaction with current mentorship (OR=17.14, p=0.002), lack of comfort reporting harassment incidents (OR=10.56, p=0.001), dissatisfaction with their training program (OR=5.77, p=0.034), perceived equity and diversity obstacles for entry into a RO training program (OR=4.15, p=0.028), personal experience with discrimination (OR=3.05, p=0.025), and personal experience with harassment (OR=2.90, p=0.049) (Table 3).

3.4. Discrimination and Harassment

Table 4 summarizes experiences of discrimination and harassment. At least one episode of discrimination during training was reported by 38% of respondents (28/75), primarily due to gender (16%), race (15%) and age (10%). At least one episode of personal harassment was reported by 24% and was mainly perpetrated by a patient/family member (56%, 9/16). Less than half of respondents (45%) felt comfortable reporting discrimination and/or harassment and only 11% reported disclosing harassment/discrimination experiences.

Training to address sexual harassment, anti-racism, LGBTQ health, and other forms of discrimination was available to 47%, 53%, 33%, and 50% of respondents, respectively. Up to 49% of respondents were not aware if these types of training existed.

3.5. Comparisons by demographics groups

Women were more likely to feel under-represented in their training program (17/41, 46% vs. 6/51, 13%, p=0.001) and perceived equity and diversity obstacles (8/41, 22% vs. 2/51, 4%, p=0.02); however, there were no differences between level of training (residents vs. fellows) (p=0.43 and p=1.0), location of training (CMGs vs. IMGs) (p=0.30 and p=0.43), race/ethnicity (p=0.81 and p=1.0) nor primary language (p=0.87 and p=0.38). Women were also more likely to strongly agree/agree to the statement “It is important I have a mentor with similar demographic characteristics to me” (21/41, 55% vs. 13/51, 28%, p=0.03).

Women perceived more discrimination events (64% vs. 19%; p<0.001), whereas male respondents felt more comfortable reporting harassment incidents (60% vs. 26%; p=0.01). Women were also more likely to perceive discrimination based on gender (15/41, 37% vs. 1/51, 2%, p<0.001) and age (8/41, 20% vs. 2/51, 4%, p=0.02) compared to men. Harassment events were more commonly perceived by resident respondents compared to fellow respondents (16/64, 25% vs. 1/28, 4%, p=0.018), Caucasian respondents compared to racialized minorities (32% vs. 10%, p=0.011) and CMG respondents compared to IMGs (18/70, 26% vs. 0/28, 0%, p<0.001). Racialized minorities were more likely to rate an “excellent/very good” culture of respect compared to Caucasian respondents (44/61, 77% vs. 20/37, 55%, p=0.043), although they reported more discrimination due to race/ethnicity and national origin (25% vs. 0%; p<0.001 and 11% vs. 0%, p=0.043, respectively). Residents were more likely to feel discriminated against by patients/family due to gender compared to fellow respondents (14/64, 22% vs. 1/28, 4%, p=0.03).

3.6. Comparisons by demographics groups

Open-ended questions were included to further explore trainees’ experiences and to seek suggestions for improving the learning environment and advancing EDI. Questions included: “What should training programs do to address learner mistreatment?”, “What should training programs do to advance equity, diversity and inclusion in the workplace?”, and “What should training programs do to make resident selection or faculty hiring practices more equitable?” There were 24 (24%), 20 (20%), and 21 (21%) free-text responses, respectively, to these questions.

Three themes were generated (Table 5). Firstly, respondents felt education on mistreatment and EDI were needed. Secondly, improved pathways for trainee reporting of mistreatment were recommended, including providing supportive space, ensuring anonymity and ensuring specific actions afterward. Thirdly, respondents recommended diversity in trainee and faculty selection and recruitment committees.

4. Discussion

This is the first national survey to characterize the current sociodemographics and learning environment experiences within Canadian RO trainees. Self-reported under-represented groups were identified, highlighting opportunities to improve diversity and inclusion. Although most respondents reported satisfaction and respectful cultures, key differences between groups were observed, including experiences with discrimination and harassment.

In our survey, women comprised 42% of respondents, which is representative of the Canadian RO trainee population and census data. Our results are comparable to US findings for RO residents (39.2%); however, less than the 59% of current medical students identifying as women in Canadian medical schools (22). Additionally, we found no significant differences regarding female representation according to race/ethnicity nor PGY-level. Women more often perceived encountering equity and diversity obstacles in their training, perceived more discrimination events, and felt less comfortable reporting harassment incidents compared to male respondents. Similarly, ARRO reported that women were less likely to agree that they were able to voice a contrary opinion without fear of retaliation compared to male residents (17). This highlights that gender-based disparities continue to exist within RO. We recommend targeted mentorship programs, and strategies to mitigate bias and increase representation of female medical students interested in RO.

Unexpectedly, Caucasians respondents were under-represented compared to Canadian census data (38% vs. 73%, p<0.001) with most respondents (62%) identified as racialized minority. The most common self-reported race/ethnicities were Chinese, Southeast Asian and Arab. Respondents from these groups were over-represented compared to Canadian census data even when fellows, who were more commonly IMGs, were analyzed separately from resident respondents. In the ARRO survey, 54.6% identified as Caucasian (vs. 32% in our study), with Asian being the most common racialized group (29.5%) (17). Indigenous respondents were underrepresented (1% vs. 5%, p=0.007). Efforts to improve the representation of Indigenous populations in medicine have been recognized as needed and include breaking down systemic barriers to make education more accessible, establishing strategies to support students longitudinally and adopting a distinctions-based approach to recruitment and selection processes (23).

Similar to ARRO’s findings, most respondents reported a respectful workplace culture (17). However, at least one episode of discrimination during training was reported by 38% of respondents, which is concerning and warrants further assessment. The top 3 perceived reasons for personal discrimination were gender, race/ethnicity, and age. Work done within Medical Physics faculty showed that experiences of discrimination were primarily related to gender, age and pregnancy/caretaking responsibilities (15, 16). In response, AAPM/COMP created recommendations to improve the workplace environment including creating an online, non-punitive incident reporting system for discrimination/harassment (15, 16). Similar policies should also be implemented within RO training programs, as nearly half of respondents felt uncomfortable reporting discrimination/harassment experiences.

Content analysis provided insights into how training programs can address learner mistreatment, advance EDI, and improve equity in selection processes. Our theme are similar to recommendations conceived by ARRO, which in addition also included developing inclusive cultures for psychological safety, establishing cross-institutional and -specialty collaborations, and creating mentorship opportunities for underrepresented groups (24). In Canada our professional organization CARO has implemented an Under-represented in Radiation Oncology Mentorship Program (UROMP) (25) to increase access to mentorship, foster inclusive environments and help improve diversity within our specialty. Additionally, ESTRO piloted a multidisciplinary, international mentoring programme to facilitate professional development (26). The pilot was positive endorsed by participants and is now an annual programme commencing at the ESTRO congress. Furthermore, we also recommend that efforts be directed to undergraduate medical education, with initiatives focused on medical school diversification and removal of barriers to entry to residency. For example, 67% of medical students interested in Radiology reported that a diverse and inclusive learning environment impacted their program ranking selections (27). In 2021, CaRMS established EDI as a strategic priority, with the goals of collecting diversity data to inform the current environment and future improvement of residency applications (25). One way of accomplishing this is through a voluntary CaRMS self-identification questionnaire that allows applicants to confidentially self-report whether they identify as a member of an equity deserving group (28). These initiatives will lay the foundation to identify inequities, which in turn can help training programs address these gaps.

Our study has several limitations. First, there is a potential selection/response bias in that those with strong opinions (positive or negative) may have been more likely to respond. For example, in our study Caucasian respondents were under-represented compared to Canadian census data. Race data is not collected via CAPER; however, our respondents were representative of the Canadian radiation oncology trainee population regarding gender identity and level of training Second, some demographic groups were small (e.g. 2SLGBTQIA+ and Indigenous respondents), and given the low numbers, further analyses were not possible, limiting the conclusions that can be made.

5. Conclusions

We characterized the current EDI characteristics of sociodemographic groups within Canadian RO training programs and assessed trainee lived experiences. Although the majority reported satisfaction and a respectful culture within their programs, key differences between groups were observed, including experiences with discrimination and harassment. Targeted strategies and stronger institutional policies to improve representation and reduce rates of discrimination and harassment in Canada’s RO training programs are needed. Future work includes collaboration with CARO and RO training program leadership to formally implement EDI initiatives and establish priorities to promote, recruit and retain diverse members, and improve workplace environments and training programs for all.