Gender differences in self-assessed performance and stress level during training of basic interventional radiology maneuvers

Gender differences are known to influence physician-patient interaction and patient satisfaction [1‐5], medical practice, and therapy outcomes [1, 2, 4‐6]. In the case of objectively comparable medical performance, for example, patients rated treatment success depending on the doctor’s gender, which resulted in differences in receiving appreciation and positive public representation [3, 7‐10]. Likewise, gender differences in medical specialty preferences have been described [11, 12].

Focusing on the broad field of radiology, less women working in the field of interventional radiology (IR) can be found [9, 10, 13]. Accordingly, Li et al also reported a significant gender disparity in IR compared to general radiology in Canada [14], which they attributed to possible misconceptions about IR among medical students and female physicians, presumably due to insufficient mentoring. However, they also mentioned pregnancy-related issues and gender differences in teamwork to contribute to these findings [14].

As various gender-related differences have been reported for medical education, it is imperative to discern their origin and the processes involved in their emergence to improve understanding and to take these disparities into consideration in the training of interventionalists. The underlying study therefore focuses on self-assessment in learning basic interventional endovascular tasks. Psychological variables, such as perceived stress levels and self-assessed ratings of performance and difficulty of different tasks were included in the paradigm to grant a differentiated understanding of the subjective processes and experiences that lead to these gender differences.

Data from 64 participants (26 female and 38 male) were acquired. We created a set of four tasks (1.1, 1.2, 2.1, and 2.2) with increasing and decreasing difficulty levels. The third task (task 2.1) was the most difficult one, whereby the first and the fourth task were the easiest, as indicated by the times required to solve the tasks. Thus, the most significant differences were revealed while absolving task 2.1. One female participant decided to quit during task 2.1 due to high stress levels. Data of this participant were only considered for experiments 1.1 and 1.2, as well as for the NASA-TLX (n = 64), but not for analysis of the overall performance (n = 63).

The number of studies analyzing “gender differences” increased exponentially within the last three decades, as confirmed by a PubMed search. The same applies to the search terms gender differences and medicine. Focusing on medicine, significant gender differences in medical treatment and outcomes have been described for innumerous diseases [1, 2, 4‐6, 13]. Likewise, significant differences between male and female physicians affecting medical training as well as daily clinical practice have been identified [3, 8, 10, 14]. In the underlying cohort study not only technical skills but even more important factors such as perceived stress levels and the ability to assess one self’s performance correctly were analyzed.

Basic objective parameters, such as time and a number of catheter movements, were the most primitive and superficial sort of parameters, which we acquired by analysing the procedures. From these data, we additionally calculated further parameters such as velocity and agitation. Our experimental setup contained four different tasks with different difficulty levels. The most challenging task was task 2.1, which significantly (by about 40–55%) contributed to the overall results of the objective parameters (because the participants took more time to manage this task). Whereas there were no significant differences in objective parameters in the easier tasks, we observed that the female participants took more time to solve the given tasks and asked earlier for help than the male participants. On the other hand, women were able to estimate their own objective performance much better, whereas men failed to correctly assess their own performance. The difference in working speed might result from different working strategies and it could also be triggered by different factors. For example, all participants were informed that the time was taken during the experiments. However, it was never said to any of the participants that time is of relevance to judge their performance. Nevertheless, this information alone could be sufficient to trigger male competitiveness (which from an evolutionary aspect may be expected to have been exposed to strong selective pressure throughout human history) [24]. Therefore, it could have been that male participants simply hurried up because they felt the experiment to be a challenge, while women did not see this experiment as a challenge and simply took their time to solve the tasks. This would be in line with the hypothesis, that reduced working speeds yield a higher degree of precision thereby avoiding more complications and resulting in a better outcome. This interpretation would also be supported by the findings of Barr et al, who reported women take more time in treating patients [1, 2, 4], which resulted in an increased level of patient satisfaction [1, 5]. Gender-based treatment differences were also observed in terms of 30-day mortality and readmission in an internists’ study: patients treated by female internists presented a significantly lower readmission rate and mortality after 30 days than patients treated by male internists [6]. Furthermore, Alcaide-Leon et al observed fewer mistakes made by female doctors in diagnostic radiology than by their male colleagues [25]. Irrespective of disease severity, patient-specific characteristics, and gender, treatment disparities have also been related to a surgeon’s age [26]. This, however, is not an unexpected finding as older surgeons in most cases will be more experienced.

Accordingly, gender differences in self-assessment and perception of one’s own abilities and achievements have been reported. Women, for example, do not only seem to rate their performance generally lower than men [7] but women and men also seem to perceive personal success and failure differently [27, 28]. Whereas women tend to rate their personal success as “lucky”, and failures are considered as “lack of skills”, men tend to rate their personal success as “based on skill” and failures as being “unlucky” [27, 28]. Interestingly, when being unobserved, women rated their performance similar to those of their male counterparts [7]. Ludwig et al thus hypothesized that women, when being under observation, seem to judge their own performance more accurately than men, presumably to avoid being judged negatively for their failings [7]. This explanation would also match our observation with women requiring more time to solve the interventional tasks in our experimental setup while correctly interpreting their own performance. These findings are also supported by Gill et al, who exposed male and female participants to tasks harbouring a high likelihood of failure [3]: They observed a decreased performance of male participants concomitantly with a decreased chance of success, whereas female performance decreased independently of that. In our study, higher stress levels of male participants positively correlated with an increase in time and number of catheter movements required to solve the tasks. Thus, when men were more stressed out, they tended to become more agitated. When women indicated lower stress levels and lower difficulty of tasks and self-assessed their performance higher, they showed higher catheter acceleration values than men, which could mean, that these female participants performed catheter movements more efficiently than men.

The present study has some limitations. Whereas over 60 volunteers were recruited, the number of participants still could have been higher. On the other hand, the number of participants yielded several significant results. We tried to eliminate any bias from the study using a standardized teaching video and recruitment, training, and the experiments were performed by the same male MD student. This, in fact, might have introduced some bias, as male or female participants might have reacted differently depending on the gender (or even personality) of the study supervisor. This problem, however, is almost intractable. Interestingly, we realized that men were more willing to participate in the study than women, which is another potential source for bias, which hardly can be corrected. Reasons for this might be that men are more prone to accept challenges. This theory would also be supported by Ludwig et al and Gill et al, who reported a higher willingness of male participants to accept challenges and to compare with others than women. Ludwig et al hypothesized that women could tend to rate their performance lower when being observed, which would match our results. Whether this difference affects the results of our study finally remains unclear, and matched-pair analyses of a larger cohort might yield even more exact results. The fact that we investigated the performance of volunteers naïve to catheter techniques puts the results of this study mainly into an early educational/training setting: comparing well-trained (and it would be difficult to clearly define “well-trained”) male and female interventionalists might have yielded different results. Therefore, performing this experiment with participants naïve to catheter handling provides the considerable advantage of avoiding confounding factors such as years of training. Another important point is that differences in objective catheter handling performance might have influenced the results. Whereas we ensured that none of the participants had any experience with catheter handling, it was more difficult to exclude different backgrounds regarding other situations in which hand-eye coordination is important. We therefore checked the background of the participants using a questionnaire including activities possibly affecting manual dexterity including sports and playing an instrument but found no significant differences here.

Differences in self-assessed performance, stress level, and self-assessed difficulty observed in our study were comparable to the results of other studies [3, 7, 27, 28] and could explain differences in performance between men and women. Increased awareness of failing in women might already have influenced our objective parameters, such as perceived stress levels, difficulty, and own performance. One consequently may interpret that the female participants presented with a remarkably critical self-perception, which may have resulted in a raised willingness to ask for help. Most interestingly, when our male participants felt their skills were insufficient, their self-assessed stress levels increased proportionally.

In conclusion, we observed several interesting gender-related differences not only regarding working speed and catheter movements in some parts of the experiment but also in self-perception of individual performance and stress levels in this educational setting. We learned that inexperienced male volunteers solved the most difficult interventional task somewhat faster than women, who in contrast took themselves more time and performed catheter movements with less agitation. Female participants asked more frequently and earlier for help than their male counterparts, while they showed a more “aggressive tactic” with significantly more catheter movements per second. Female participants inexperienced in endovascular procedures rated their performance in general lower than inexperienced men and indicated higher stress levels during simulated angiography, but were much better able to correctly assess their objective performance. Whereas men tended to overestimate their own performance and failed in self-assessment of objective parameters. Although these gender-based differences do not necessarily reflect the situation in well-trained male and female interventionalists, instructors should be aware of these differences as they may affect the interpretation of young colleagues’ behavior and performance, thereby improving training and teaching of not only interventionalists.