Introduction

Surgery has long been considered an old boys’ club [1, 2]. In the UK in 2013, women made up less than 10% of consultant surgeons, while a majority of new medical graduates responding to a survey on career intentions perceived a surgical career to be unwelcoming to women [3]. Proportions of women and graduate students entering medical school are increasing, but they are less likely to pursue a career in surgery, prompting appeals in the literature for renewed emphasis on promoting surgery to medical students [2,3,4].

Much of the literature assessing motivations to pursue surgical careers is restricted to a high-income country context, wherein a lack of female role models and a hostile culture towards women are frequently cited as key barriers to pursuing a surgical career as a woman [2, 4,5,6,7]. Yet while significant work has been completed in Western countries on the benefits of mentorship and positive role models [2, 4, 8, 9], there is little evidence that this trend is generalizable worldwide; indeed, work that has examined global generalizability of such barriers has found significant regional variation [8]. Furthermore, some studies from high-income countries dispute that a lack of role models or lifestyle factors influence surgical career choices in women relative to men, instead underscoring a hostile surgical culture and a perceived “surgical personality” as deterrent factors specific to women [10]. Workload and “work-life balance” have also frequently been cited as deterrents for women undertaking a surgical career [11, 12]. There is a paucity of research examining factors influencing a decision to pursue a surgical career in low- and middle-income countries.

With one-third of surgical disease burden occurring in low- and middle-income countries, there is a strong impetus to study barriers to entry for surgical careers given increased interest in increasing surgical workforce in these countries by the year 2030 [13, 14]. Therefore, we composed a survey on attitudes and barriers to pursuing a surgical career and undertook a cross-sectional analysis with the resultant data, aiming to examine discrepancies between countries of differing income strata and between genders (online Supplemental Material).

Methods

In collaboration with the University of Oxford, we developed a cross-sectional, mixed methods survey. The survey was created by a focus group of medical students from Asia, Africa, Europe and the Americas. Over a series of online meetings, we determined focus areas and performed face validation of the survey. A literature review was performed to attest for criterion validity, and all previously surveyed areas of interest studied regarding medical students interested in surgery were included in the final survey. The survey contained 37 questions arranged in five sections: basic demographics, including medical school location and length of study, as well as questions relating to the following thematic areas: education, resources, perceptions and profession-related. Equivalent-form reliability was considered and tested. The language of the survey was deemed suitable for non-native speakers by trialling on students from four different continents. Equivalent-form reliability was tested using different question wordings relating to access to postgraduate training and surgical teaching. The Cronbach’s alpha test was used to determine internal validity of study questions. The questions were measured on a Likert scale (1 strongly disagree, 2 disagree, 3 neither agree of disagree, 4 agree, 5 strongly agree), which was converted to binary on multivariate analysis. Analysis was completed using Intercooled Stata software, version 13.0 (StateCorp College Station, TX, USA).

The survey was hosted on a web-based survey platform via the University of Oxford and was distributed via multiple mailing and social media of the International Federation Medical Student Association (IFMSA) and its Global Surgery Working Group. The study was open to all medical students and was entirely anonymous and confidential. Questionnaires from graduates after their internship year (defined as one year post medical school) were excluded, in keeping the inclusion criteria the same as IFMSA membership. The outcome of interest was consideration of pursuing a surgical career (yes/no). A univariate analysis was performed of individual questions against this outcome. A logistic regression model was then constructed with all factors that were statistically significant on univariate analysis, plus other questions of interest that were added adjustment, including workforce density in country. A stepwise progression was performed, and factors that showed no association were withdrawn from the model. The final model contained mentorship, acute care element of surgical careers, cost of postgraduate training, workload, academic competitiveness, academic competitiveness, year of study and World Bank Income status, which was divided into HICs and UMICs ad LICs and LMICs. Interactions were tested in the model with gender. Subsequently, a subgroup analysis was performed on male and female participants separately and then by income strata.

Results

Our study population consisted of 639 medical students from 75 countries (Fig. 1). The demographic profile of respondents is given in Table 1. Overall, 58.5% (n = 374) of respondents were female. 4.5% (n = 29) of respondents were from low-income countries (LICs), 27% (n = 172) from lower-middle-income countries (LMICs), 20% from upper-middle-income countries (UMICs) and 48% (n = 308) from high-income countries (HICs). Table 1 shows student’s intentions to pursue a surgical career by the stage of training, separating male and female respondents. The univariate analysis of questions is displayed in Table 2, and multivariable logistic regression results are shown in Table 3.

Fig. 1
figure 1

Country distribution of student responses

Full size image
Table 1 Population characteristics
Full size table
Table 2 Univariate analysis
Full size table
Table 3 Multivariate analysis
Full size table

On multivariate analysis, mentorship [OR 3.42 (CI 2.29–5.12) p = 0.00], the acute element of the surgical specialties [OR 2.22 (CI 1.49–3.29) p = 0.00], academic competitiveness [OR 1.61 (CI 1.07–2.42) p = 0.02] and being from a high or upper-middle-income country (HIC and UMIC) [OR 1.56 (CI 1.021–2.369) p = 0.04] all increased likelihood to be considering a surgical career. Perceived access to postgraduate training [OR 0.63 (CI 0.417–0.943) p = 0.03], increased year of study [OR 0.68 (CI 0.57–0.81) p = 0.00] and perceived heavy workload [OR 0.47 (CI 0.31–0.73) p = 0.00] all decreased likelihood to consider a surgical career. Perceived educational quality however showed no affect on students’ decision to consider a surgical career when tested using several variations of education-related questions. It was thus excluded from the model. The gender of the mentor was significantly associated with pursuing a surgical career in the univariate analysis but became insignificant when controlling for mentorship as a whole on multivariate analysis, and this question was also thus excluded from the final model.

On subset analysis, perceived poor access to postgraduate training made women 60% less likely to consider a surgical career [OR 0.381 (CI 0.217–0.671) p = 0.00], whilst not showing an effect on men [OR 1.13 (CI 0.61–2.12) p = 0.70. Concerns about high cost of training halved the likelihood of students from low and low-middle-income countries (LICs and LMICs) considering a surgical career [OR 0.45 (CI 0.25–0.82) p = 0.00] whilst not demonstrating a significant relationship in HIC or UMIC countries. Women from LICs and LMICs were 40% less likely to consider surgical careers than men, when controlling for other factors [OR 0.59 CI (0.342–1.01 p = 0.053]. Figure 2 demonstrates the raw data answers from the questions “How do you rate access to postgraduate surgical training programmes?” dividing answers by income strata, whilst Fig. 3 shows the responses to this question by both gender and income strata as a percentage of the total responders within each category. Finally, Fig. 4 demonstrates the regression coefficients for the association of access to postgraduate training opportunities versus decision to pursue surgery as a career, with coefficients greater than one representing a positive association, whilst controlling for the other questions described in the multivariate analysis.

Fig. 2
figure 2

Answers to "How do you rate access to postgraduate surgical training programmes?" by income strata of country of study

Full size image
Fig. 3
figure 3

Answers to "How do you rate access to postgraduate surgical training programmes?" both gender and income strata as a percentage of the total responders within each category

Full size image
Fig. 4
figure 4

The regression coefficients for the association of access to postgraduate training opportunities versus decision to pursue surgery as a career, with coefficients greater than one representing a positive association (with agreement that access to postgraduate training affects their decision-making associated with higher likelihood to pursue surgery). Each bar represents one multivariable analysis, adjusting for percentage medical school completed, acute care element of surgical careers, cost of postgraduate training, workload and academic competitiveness

Full size image

Comparing groups by World Bank income strata in country of study, concerns about high cost of training halved the likelihood of students from low and low-middle-income countries (LICs and LMICs) considering a surgical career [OR 0.45 (CI 0.25–0.82) p = 0.00] whilst not demonstrating a significant relationship in students from HIC or UMIC countries. Women from LICs and LMICs were 40% less likely to consider surgical careers than men, when controlling for other factors [OR 0.59 CI (0.342–1.01 p = 0.053].

A number of questions were used to test internal survey internal validity, using the Cronbach’s alpha test. “Surgery is taught well at my university” and “How would you rate the quality of your surgical teaching overall?” demonstrated a reliability coefficient of 0.7741 (good). “Do you think that doctors in your country have good access to postgraduate surgical” and “How do you rate access to postgraduate surgical training programmes?” were tested, resulting in a reliability coefficient 0.5464 (acceptable).

Discussion

This study demonstrates that a wide variety of factors influence student’s decisions pursue a surgical career. Mentorship, the acute element of surgical care and academic competitiveness increased students’ likelihood to consider surgical career across gender and income strata, whereas concerns over access to postgraduate training, heavy workload and progression through medical school all decreased likelihood to consider surgery. Access to postgraduate training in particular demonstrated a gender interaction and reduced women’s likelihood to consider surgical careers by around 60% when controlling for other factors (p = 0.0281). Whilst the importance of mentorship is in concordance with existing literature, perceived poor access to postgraduate training preventing women from considering surgical careers is a new finding. This finding was generalizable across income groups and may demonstrate more subtle biases throughout the surgical application process.

Whilst many global surgery programmes focus on teaching surgical skills, our study has demonstrated that perceived quality of surgical teaching does not affect student’s decisions to pursue surgery, nor do concerns regarding the overall surgical quality in their country of study. It may be that focusing global surgical programmes on improving access to postgraduate opportunities or lowering the cost of postgraduate training may be more effective in bringing medical students into surgical careers in these settings. A concerted effort into improving mentorship structures would likely also be beneficial, given that in our study, agreement with the importance of mentorship made students four times more likely to consider a surgical career. The effect of year group and progression through medical school is in accordance with the literature. Surgery is widely regarded as a popular specialty of choice amongst first year medical students [15], but importantly most students change their specialty choices, regardless of initial interest [16].

There has been a growing interest in looking at women in surgery. With almost all literature examining European and American trainees, our study is the first attempt to see whether such findings are consistent worldwide and across income stratas. Access to and cost of postgraduate training made women in specific regions and income groups significantly less likely to pursue a surgical career, and on subset analysis of the LIC and LMIC students, being female made students 40% less likely to consider surgery, independent of other factors (see Tables 4, 5) Trials of women-only scholarships for postgraduate surgical training may be useful in eliminating financial barriers from young female doctors, and further work should be pursued to elicit the causes of this perceived financial impediment.

Table 4 Multivariate by gender—(i) multivariate analysis in female students and (ii) multivariate analysis in male students
Full size table
Table 5 Multivariate by income strata—(i) multivariate analysis in students studying in HICs and UMICs and (ii) multivariate analysis in students studying in LMICs and LICs
Full size table

The finding that perceived educational quality bears no significance on students’ decision to pursue a surgical career has important implications for the success of mission trips from HICs focusing on teaching surgical skills. It suggests that such efforts may be better focused on increasing access to sustainable postgraduate training programs, or else offering meaningful mentorship opportunities. However, whilst we have demonstrated that gender has an effect on decisions to pursue surgery, there may also be interactions with other factors, such as the income status of the students’ country of study. Women therefore should not be treated as a global, homogenous group.

It is important that medical educators consider how medical school-based experiences, interactions with mentors and perceived postgraduate opportunities, impact students’ specialty choices and decision-making. Mentorship clearly plays an important role in the decisions of both men and women, but access to and funding for postgraduate training may also play an important role in encouraging women to take up this career path.

This study was limited by its selection bias. IFMSA is made up of motivated and internationally active students and may not be generalizable to the true medical student population. However, the authors believe that this bias is limited by the fact that this was true in both male and female populations, and also that no questions pertained to travel or global health, as would preference students attending international events. The study is also limited by disproportionate representation of certain countries. This has led to larger confidence intervals especially on subset analysis and may lead to type II error. However, by targeting existing medical students, this study eliminates any recall bias by of questioning surgical trainees. This study focused entirely on pursuing surgical careers, and does not question participants on their consideration of anaesthesia or obstetrics, both of which are important for reducing the burden of surgical disease. Further work looking into these areas would be beneficial.

Conclusion

This study is the first of its kind globally to quantify differences in barriers to surgical careers in medical students worldwide. Whilst we have been unable to draw firm conclusions on all of the driving forces behind decision-making, we have determined many patterns and areas for further study and potential investment. Mentorship the acute element of surgical care and academic competitiveness were positively impactful on medical students’ decision to pursue a surgical career across genders and income strata, while perceived poor access to postgraduate training and heavy workload made women in particular less likely to consider a surgical career. While there are some similarities amongst women worldwide, it is important to ensure that cultural assumptions and evidence amassed in high-income country are not generalized to other regions without evidence of consistency in behaviour or perception. There is considerable interaction between income strata and gender differences, as well as different factors that may be targeted to encourage female trainees in different regions. Emphasis on effective mentorship, improving postgraduate opportunities and reducing workload may all contribute to improving opportunities for medical students, and particularly women to pursue surgical careers worldwide.