Analysis of gender in radiology in Australia and its importance to the profession and workforce planning

Medical workforce diversity is important with gender constituting a significant role. Male and female medical practitioners participate in the workforce differently: understanding the cultural and social expectations, economic productivity, professional opportunities, and the effects on workforce supply, will aid workforce planning. Having a workforce that reflects the diversity of the community is important in providing patient‐centred care. As more than half of medical graduates are female, it would be expected that this is reflected in radiology specialty. We analyse the Australian gender‐specific data from the Royal Australian and New Zealand Clinical Radiology (RANZCR) clinical radiology workforce census from 1992 to 2020, focusing on changes in gender representation, number of hours worked per week, differences in subspecialisation and geographical distribution. This analysis found that the proportion of the female radiologists increased from 13% to 29%: still an underrepresentation of women radiologists when compared with the gender distribution of medical students and junior doctors. This will persist in the short to medium term, given the tapering of female doctors entering radiology training. In terms of workforce planning, women are more likely to work less than their male counterparts in the early to mid‐career. Women are underrepresentated in interventional and neurointerventional radiology. There is more self‐reported subspecialty interest in breast and women's imaging. A review of the literature demonstrated a similar situation in comparable countries. We also considered the reasons, potential solutions for this, and knowledge gaps where research is needed.


Introduction
Over the previous decade, gender parity as part of diversity, has become a major subject of interest, as evidence that diversity improves workforce engagement and innovation by diversity in skill-sets and viewpoints. 1 In the corporate world, it is recognised that women in leadership positions results in greater competitiveness and consequently more economic success. 1 Workforce diversity, including gender parity in medical specialties, has been identified as a way of addressing the social inequalities and hierarchies associated with gender, and is used as a measure of workplace performance. [2][3][4] A suitable metaphor would be the lyrics of the song titled 'We're all fruit salad!' from the Wiggles, a popular Australian children's band, citing that, 'It takes all different kinds of fruit to make a perfect bowl. ' Many countries have reported that within the health sector, women comprise over 75% of the workforce including nurses and allied health, making them indispensable in the delivery of health care. 5 In contrast, globally, there is, proportionately, fewer females entering medical specialties compared with general practice. A recent cross-sectional study by Cater et al. looked at 95 professional radiology organisations in 75 countries: only 34% of radiologists were female despite women making up half of American and European medical graduates. 1,5 The 2016 Royal Australian and New Zealand College of Radiologists (RANZCR) activities report stated only 26% and 33% of radiologists in Australia and New Zealand, respectively, were female. 6 There are social perceptions that women are poorer workforce investments as they contribute less to the workforce given concomitant societal expectations or demands. 7 These biases are being addressed with institutional policies for equal opportunities, such as parental leave and flexible training programmes. Some studies report that female clinicians show greater empathy, are more patient-centred, and are better communicators. 8 Patient demand for female doctors has increased, especially when the problem is gender-related. These attributes should be embraced and leveraged.
Efforts to achieve gender equity, in turn, can affect differences in allocation of resources 7 as male and female medical practitioners are known to participate differently in the workforce. These differences include the type of work, hours of work, approach to clinical practice, and are influenced by marital status and home or family circumstances. For instance, where there is difference in hours worked per week, there would need to be a change in trainee numbers to ensure that there will be no workforce shortage down the line. What about the gender variation in subspecialty decision? 9 For this reason, appropriate and responsive workforce planning is essential to meet the needs of the Australian population. This is increasingly important, given the ever-increasing workforce gap in radiology, with excess demand growth for radiologists at a rate of 3.8% per annum. 10 RANZCR has been conducting census surveys of radiologists in Australia and New Zealand every 4 years since 1992. These are a rich source of longitudinal data about the clinical radiologist workforce.

Aim
This paper reviews the gender dynamics in the clinical radiology workforce in Australia in relation to: 1 Trends in gender representation in the clinical radiologist workforce over time. 2 Differences by gender in hours worked, and subspecialisation. 3 The experiences of the workforce by gender.
In the context of staffing levels and adequacy, 4 Understand the effects on workforce participation which need to be considered in workforce planning and related policies.

Methods
The RANZCR Clinical Radiology workforce Census was analysed to extract data on gender in relation to workforce participation including hours worked, practice type and location, subspecialisation, the current experience of radiologists in the workplace, and the effects on staffing levels and skill-mix throughout Australia. A longitudinal, descriptive analysis of the demographic profiles was conducted by gender, across the years, to 2020, exploring the patterns of workforce participation. This is a combination of regression and time-series analysis, observing a broad cross-section of subjects who are repeatedly surveyed over time.
The frequency of responses in the most recent census was used to ensure the response structure was appropriate to analyse gender differences.
A review of the literature relating to gender in the radiology workforce was performed to examine international trends, and was used to compare with the situation in Australia and inform the discussion about gender dynamics in radiology workforce.

Results from the RANZCR census
The 2020 census was distributed to 2611 radiologists in Australia based on the RANZCR membership database, 2332 of whom are currently practising. This included all educational affiliates and international medical graduates in the database. Of the 2332 radiologists invited, 1658 (71.1%) were male and 674 (28.9%) female. There was a response rate of 45.4%, with 1186 responding to the questionnaire, of whom 1098 are currently practising as radiologists. Of the 1098 radiologists, 765 (69.7%) were male and 333 (30.3%) female; 46.1% of all male radiologists and 49.4% of all female radiologists who are currently practising, responded to the census. This response rate is similar to the previously collected censuses.
Female radiologists make up 26.5% of the radiology workforce in 2020. The proportion of female radiologists making up the workforce was 15.8% in 2016 and 13% in 1992. The female radiologist cohort is relatively younger than their male colleagues, with mean age of 47 and 51 years, respectively ( Fig. 1). The proportion of female trainees was 38.0%, 35.2%, and 33% in 2012, 2016, and 2020, respectively (Fig. 2). These numbers are people in training posts, and not the percentage entering training that year.
In a typical working week, the mean hours worked by women is 34 h/week, and for men 41 h/week, a difference of 6 h/week (Figs 3, 4). Women are more likely to work part-time in early and mid-career, whilst men are more likely to work part-time in late career (Fig. 5).
Women, on average, spend 6.5 hours a week on procedures whilst men spend 10 hours a week on procedures. Women spend 3.8 hours in teaching and supervising registrars whilst men spend an average of 3.2 hours doing so. The proportion of each gender in after-hours and teleradiology work are similar. The ratio of each gender in private and public settings are also similar.
There is no significant difference in the proportion of male and female radiologists who have self-reported subspecialty interests. More women identified as subspecialists in breast imaging, compared with men. There is a disparity in gender representation in interventional and neuro-interventional subspecialties, with women being less likely to be interventional or neuro-interventional radiologists.
The Modified Monash Model (MMM) was developed in 1999 to target workforce programmes aiming at attracting health professionals to more remote and smaller communities. It classifies metropolitan, regional, rural, and remote areas according to remoteness as defined by the Australian Bureau of Statistics (ABS) and town size. The radiologist distribution across the different MMM categories was calculated to understand workforce distribution and there is roughly similar proportion of the   genders when plotted against the different MMM regions. For instance, in metropolitan (MM1) areas, the proportion of men versus women is 71% and 29%. The only deviation from the ratio of male:female radiologists is in the medium rural towns (MM4), with 90% male predominance, but the numbers are small with only 9 men and 1 woman radiologist, so statistical significance cannot be ascribed to this given the small number. Additional information is provided in the Supporting Information, as excerpts from the RANZCR 2020 census.

Discussion
The situation in Australia The proportion of women radiologists has increased in the period from 1992 to 2020. The 28.9% proportion of women radiologist is only slightly below the Australia Institute of Health and Welfare (AIHW) report that women comprise 29.5% of the most common specialties of practice. This proportion is similar when plotted against different MMM regions. The mean age of female radiologists being younger than their male colleagues can be expected given the recent trend of increasing percentage of females entering the training programme and workforce. Saying that, there is steady decrease in the proportion of female trainees since 2012.
Women hold a lower full-time equivalent (FTE) when compared with men although women are to work more hours than men in late career. This has been fairly stable when compared with the prior censuses. Women spend disproportionately less time doing procedures and more time in teaching and supervising registrars. Women are also less likely to be interventional or neurointerventional radiologists. 12 How does Australia compare with other countries?
Since 1990, the percentage of Australian female medical students has increased to approximately 50%. Despite the good gender balance in most medical schools, and hence, an available recruitment pipeline, the current RANZCR census shows underrepresentation of females in the radiology training programmes. 4,13 Similar trends have also been observed in the United States. 14 The data indicated a roughly equal proportion of females (50%) versus males in the US population in medical schools, with a sharp decline in female representation in radiology training programmes (27.8%), and as qualified radiologists (26.1%) (<0.001). 15,16 In Canada, 57% of medical students are female, but only 2% of females as compared with 7.4% of males choose  radiology as a specialty 17,18 with 32% of radiologists being women. In a recent survey of 26 countries and 29 professional radiology organisations, approximately 33.5% of radiologists were women with variation in the distribution across the countries, lower in the USA at 27% and around 50% in Spain and Thailand. 5

Current trends in gender gap
In Australia there are proportionately more female trainees than female radiologists, indicating partial redress of the gender imbalance. 6 The 2014 RANZCR activity indicated 37.5% trainees in Australia were female. 6 Similarly, this trend is also observed in other countries. In the most recent 2016 ACR workforce survey, only 21.4% of radiologists are women, yet women make up a higher proportion of the younger generation of radiologists, with 26% of those under 45 years of age being women. 19 In the 2014 United Kingdom Royal College of Radiologists census, one-third of the radiology consultants are women and 46% of current trainees are women. 20 Whilst this is promising, it should be noted that in the last 4 years, the ratio of Australian trainees has plateaued at 33%. If efforts are to be taken to encourage more women to take up radiology in their specialty training, one must understand the factors that influence the choice of specialty.

Reasons for the gender gap
Drinkwater's 2008 qualitative study of medical students conducted using semistructured interviews of third-and fourth-year medical students in the UK identified work and family balance as a core value influencing medical students' choice of careers. 21 Women rated flexible work hours, intellectual challenge, more patient contact and role models as having more positive impact on their choice of specialty. Women were also more ready to compromise professional goals and achievements to fulfil their family roles such as providing childcare. Roubidoux et al.'s survey of 250 first-year medical students before and after a 7-week introductory radiology course, found that medical students' interest in radiology careers did not differ by gender. 22 This interest in the radiology specialty by female students was similarly demonstrated in Zener's survey of all Canadian medical students. 23 Totally, 917 medical students responded with 514 (56%) females who reported radiology as intellectually stimulating and offering more job opportunities. 23 From these, we can conclude that there is a healthy interest in radiology at the medical student level, and that perks of a career in radiology is closely aligned with what the medical students desire or view favourably. Other perks include reasonable call hours and higher salaries compared with other specialities. For some reason, women still do not choose clinical radiology as a career. 11,24 This is interesting in the context of specialty choice in Australia and New Zealand, as radiology is widely perceived to be a specialty where it is easy to work part-time yet its popularity amongst female medical graduates is declining, and those specialties traditionally considered very demanding in terms of hours worked and on-call commitments are increasing in popularity amongst medical graduates (e.g., obstetrics and gynaecology). 25 In Australia, the gender ratio in junior doctors applying for radiology training programme mirrors the ratio of applicants who are matched, indicating that the disparity cannot be explained by selection bias (Fig. 6). This is shown in the data provided by the Postgraduate Medical Council of Victoria. Studies undertaken in various countries have shown multidimensional factors that affect gender disparity in the radiology workforce. Some of the issues are listed below, which can be investigated further to decide if it may be relevant in the local scene: In some areas, there are reports of radiology being a male-dominated culture that is discouraging to women. 3 Studies have found perceived negative factors and misconceptions about radiology, such as isolated dark rooms, minimal patient contact, level of competitiveness getting into the programme, and knowledge of physics required. This can negatively influence women more than men. 14,16,22 Additional years required for training or subspecialisation may also be a factor, considering family planning reasons. Studies of female medical students undergoing radiology rotations have also revealed that the long-term radiation risk is also a deterrent to the specialty, 26,27 more so than family responsibilities. 28 The on-call requirements of the interventional subspecialties may also be a deterrent for trainees who are primary carers for their young families. 12 It has been reported that the greater the share of women holding leadership positions in a workplace, the greater the likelihood of supporting one another by providing role models and mentorship to the younger, more junior women. 21 There is an underrepresentation of women in academic radiology and leadership roles which can play a role in the gender disparity. 18,27 Saying that, Baker et al. 29 looked at correlating the percentage of women from 186 residency programmes against the number of female chairs of academic departments, and concluded that the low number of female radiology trainees does not appear to be related to the gender of programme directors, positions of influence, and authority.

Potential solutions to bridging the gender gap
First and foremost, the policies for non-discrimination, cultural competence, and a supportive environment must be reinforced. Being aware of biases, conscious or unconscious, is also important, so they can be negated, making the workplace more equitable. 1,16 Given the recruitment pipeline, there is a need to expose medical students and junior doctors to the specialty of radiology and what it entails, early on in their careers. This is to pique their interests in radiology, and to dispel some of these common misconceptions such as the lack of patient contact or perceived isolation.
In view of the acceptance rate of men and women getting into radiology training being similar to the proportion of applicants, there is little argument for quota-based selection based on gender, which can be a double-edged sword that can lead to perception of not being selected based on merit.
Better education and guidance on risks of radiation exposure and ways of mitigating the risks can be explored, particularly during pregnancy and training. 26 For example, junior doctors should know that there are safeguarding policies in place to monitor and reduce radiation exposure, and that these legal requirements are more stringent around pregnancy. Training can also be modified so that trainees who have missed out on the various experiences involving radiation (such as interventional procedures) due to pregnancy-related reasons, are able to catch up with their peers. This can, in turn, have a positive impact in addressing the gender disparity in interventional radiology down the line. On the same note, there are also part-time, or flexible work options, available for family reasons that can be more easily arranged for radiology compared with other specialties.
Education and mentoring can help women at different career stages to be better advocates for themselves. 16 This may promote more women to leadership roles which, in turn, may also help cultivate future women radiologists. 4

Addressing workforce planning
Further along the pipeline, with increasing female participation in the radiology workforce, understanding the pattern of contribution is important for workforce planning. In 2014, 22% of radiologists were female, but 46% of all parttime radiologists are female. From the current RANZCR census, women are more likely to be working part-time compared with men. This is also reflected in the 2014 and 2016 ACR surveys, with more female radiologists opting to work part-time compared to their male counterpart. 19,20 Data from the RANZCR and ACR over the last decade also indicated that women differ in terms of their fellowship training and sub-specialisation. In particular, women are underrepresented in neurointerventional radiology and interventional radiology but are overrepresented in breast imaging, obstetrics, and gynaecology imaging. 11,19,20,30 Part of this may be the risk of radiation exposure or on-call responsibilities which is more relevant in women of child-bearing age. Talking to female radiology trainees with young families, it has been raised that limited training exposure to intervention during training given pregnancy status, results in a gap in skills that may limit their career options and opportunities later on in their career.
There is conflicting international data on whether women are more likely to participate in academic medicine or private practice. ACR data suggests the former 4 whilst others suggest that women may find private practice more lucrative, with more flexible working hours that enable a balance of family life and successful career. 18,31 Part of this may be to do with sampling and geographical distribution. Whilst there is a trend with more female authors in publications in recent years, 3 women tend to be at entry-level academia with few being promoted beyond the level of assistant professor or senior faculty. 13,18,27 This is a pattern that is widespread across academia and not specific to radiology.
Internationally, there are also fewer females who are leaders in radiology with only 15% in positions such a directors and chairs of departments. [18][19][20] It is too early to tell whether this is due to the female radiologists entering the workforce later, hence the lag in women representation in more senior positions. Whilst there is scant local data looking at women in academic radiology or leadership roles, it is not expected to deviate too far from the international trends. Future research can focus on this facet of radiology. Nevertheless, fostering leadership roles and facilitating a pathway towards either, would go far in benefiting individuals with academic or leadership interests. 31 There is no report of differences in retirement age by gender in either the RANZCR or ACR activity data. Based on the current RANZCR census, we have observed that female radiologists tend to work more hours per week than men in the later stages of their careers. This may depend on health and life expectancy which varies between the sexes. This literature gap is worth exploring. It may also reflect decreased earning earlier in their careers because of breaks in employment and more part-time work.
An Italian study demonstrated a higher percentage of women radiology trainees with a smaller gender gap. 32 However, these women trainees perceive the work environment as less fair; less satisfying and report anxiety, depression, as well as minor psychiatric disorders. Wellbeing and satisfaction of doctors are important, as a dissatisfied doctor is more likely to experience burnout, reduce hours worked, or leave practice. Other studies have shown that women medical practitioners have 1.6 times the odds of reporting burnout when compared with men, with the odds of burnout increasing by 12-15% for each additional 5 h of work in a 40-h week. 16 So, part of addressing of gender disparity, is to look into the work environment and dynamics.
Finally, whilst working towards a diverse and inclusive workforce that is reflective of the community it serves, this needs to be done with full recognition that aspirations and goals are those of the individual and not inherently tied to gender or other characteristics used to define individuals. An example of this is the increasing number of people taking parental leave irrespective of gender. We are not wanting to add further to gender stereotyping as this is unhelpful. Work-life balance and work-family conflict are issues for men and women radiologists, with perhaps different cultural expectations in men resulting in men not taking more time for family or self than they would have liked. 33 Expanding on this, we hope that this manuscript can start a dialogue for all radiologists, to achieve gender parity in its most general and broadest terms, men and women alike.

Limitations
There are a few limitations to the current RANZCR analysis. The gender analysis was not focus of prior census, so it was difficult to extrapolate detailed data on prior subspecialty or more gender-specific questions. To protect respondent privacy, the original data was not kept, hence cannot be reanalysed for the gender-specific questions. There is also a potential for bias, given that more women radiologists responded to the questionnaire than men.
The data on the intended retirement age, to determine if there is difference between the different genders, has been scant and unreliable, hence prospective questions like this not having been asked in the later census. This literature gap can help with future workforce planning. Further areas to explore include women's representation in academic radiology and leadership roles.
Finally, gender in this manuscript and census, was classified into the two traditional binary forms, and is not adequate in capturing data of those who do not identify themselves within this system. 34 Specifically, the representation, the Sexual and Gender Minority (SGM) in Australia and New Zealand, or amongst doctors, is unknown. SGM encompasses individuals whose sexual orientation, gender identity, and expression or reproductive development varies from traditional societal, cultural, or physiological norms. To further recognise diversity and growth, this will be readdressed and information on this will, hopefully, be captured in the next 2024 census, which will have a more robust sexual orientation and gender identity data. Instead of a single question on gender, a two-tiered question will be adopted, asking about the sex that an individual was assigned at birth, and current gender identity (whether it is male, female, trans-male, trans-female or gender nonconforming).

Conclusion
There is a gender disparity in the Australian radiologist workforces, with underrepresentation of women in radiology. The sharp decline in women representation at the entry to the radiology training programme when compared with the cohort of medical graduates from which they are recruited needs further investigation and addressing. There is evidence of narrowing of the gender gap in the recent census, although the percentage of females in training is decreasing. Closing the gap will be a slow, multifactorial process with factors such as early positive exposure to radiology during medical training for recruitment purposes, dispelling common misconceptions, promotion, and leadership development. More generalised measures such as promoting well-being and preventing burnout are also discussed.
To date, female radiologists have gender-specific work patterns which need to be taken into account for workforce planning. This includes the tendency to engage more in part-time work and different subspecialties. It will be interesting to see if the recent trend in younger males working less than full time continues. There is no perceived difference in gender proportion with respect to geographical distribution. Furthermore, retirement planning, and women's involvement in academic practices and leadership roles needs to be better defined with further data collection initiatives that are currently underway.
A broader discussion of gender topics, such as workfamily conflicts and gender identification, should result in changes that will help to make radiology a more inclusive and welcoming specialty.

Research questions
The following knowledge gaps and research areas have been identified: 1 Using a qualitative and quantitative prospective study design to explore the reasons for the decline in female representation at the transition from medical school into radiology training in Australia and New Zealand 2 Using RANZCR census data and further surveys understand the pattern of female radiologist work practices in part-time work, retirement planning, academic practices, private practices and leadership roles 3 Better representation and capturing of data for the Sexual and Gender Minority (SGM) group.