Potential biases

Ten studies recruited participants through non-random selection, introducing a risk of sampling bias.^{Reference Takeda, Morio, Snell, Otaki, Takahashi and Kai7,Reference Powell, Cooles, Carrie and Paleri10,Reference Doshi and Carrie12–Reference Qorban, Al-Khatib, Howldar, Allinjawi, Jawa and Baig19} Three of these studies recruited participants through their applications to otolaryngology residency or national selection, subjecting their results to membership bias.^{Reference Nguyen, Liu and Church13,Reference McCaffrey14,Reference Bhutta, Mandavia, Syed, Qureshi, Hettige and Wong16} Similarly, four studies surveyed medical students with a specific interest in otolaryngology.^{Reference Takeda, Morio, Snell, Otaki, Takahashi and Kai7,Reference Nellis, Eisele, Francis, Hillel and Lin15–Reference Newton, Grayson and Thompson17}

Two studies were only available as conference abstracts.^{Reference Acharya, Kuo, Raithatha, Haywood, Sharma and Kothari11,Reference Nguyen, Liu and Church13} As a result, we obtained limited information on these studies with no details of the sampling methodology^{Reference Acharya, Kuo, Raithatha, Haywood, Sharma and Kothari11} or the method of statistical analyses.^{Reference Nguyen, Liu and Church13} This limited our ability to appraise their data collection methods.

All included studies relied on questionnaires to collect data, with the inherent limitation that this restricts the opportunity for participants to express their views. Six studies used questionnaires that consisted entirely of closed questions which further compounds this disadvantage.^{Reference Takeda, Morio, Snell, Otaki, Takahashi and Kai7,Reference Powell, Cooles, Carrie and Paleri10,Reference Doshi and Carrie12,Reference McCaffrey14,Reference Newton, Grayson and Thompson17,Reference Qorban, Al-Khatib, Howldar, Allinjawi, Jawa and Baig19} Response rates showed wide variation, with one study that had only a 12.5 per cent response rate.^{Reference Powell, Cooles, Carrie and Paleri10} One study did not specify whether responses to their questionnaire were anonymous.^{Reference Nellis, Eisele, Francis, Hillel and Lin15} If participants were not promised confidentiality their responses may have been altered through response bias or the Hawthorne effect, where participants' awareness of being observed leads to alterations in their behaviour, potentially undermining the results. Three studies had notably small sample sizes leaving results prone to sampling error and limiting the power to detect significance where statistical analysis was applied.^{Reference Doshi and Carrie12,Reference Nellis, Eisele, Francis, Hillel and Lin15,Reference Newton, Grayson and Thompson17}

Reliability and validity

Copies of questionnaires could not be obtained for five studies, prohibiting a thorough assessment of their reliability and validity.^{Reference Acharya, Kuo, Raithatha, Haywood, Sharma and Kothari11,Reference Nguyen, Liu and Church13,Reference Nellis, Eisele, Francis, Hillel and Lin15–Reference Newton, Grayson and Thompson17} Six studies provided copies of their questionnaire, five of which appeared to have good face validity.^{Reference Takeda, Morio, Snell, Otaki, Takahashi and Kai7,Reference Powell, Cooles, Carrie and Paleri10,Reference Doshi and Carrie12,Reference McCaffrey14,Reference Garg, Arora, Kumar and Singh18} The questionnaire by Qorban et al. was poorly designed: it could not be ascertained whether a large proportion of responses were in the context of attracting or deterring the students from choosing otolaryngology as a career.^{Reference Qorban, Al-Khatib, Howldar, Allinjawi, Jawa and Baig19} These results were neither acknowledged nor discussed by the authors. Instead, responses from the remainder of the questionnaire were used to draw conclusions. Thus, we feel this study has poor face and internal validity. Three studies analysed results based on self-reported specialty preferences of medical students. It was not demonstrated whether these specialty preferences had predictive validity for application to otolaryngology training.^{Reference Takeda, Morio, Snell, Otaki, Takahashi and Kai7,Reference Bhutta, Mandavia, Syed, Qureshi, Hettige and Wong16,Reference Newton, Grayson and Thompson17}

The external validity of the findings is limited. There was wide variation in the geographical location of the studies. None of the included studies showed an equal representation of medical schools or teaching hospitals within the country where they were completed. One study included only female participants.^{Reference Garg, Arora, Kumar and Singh18}

Implications for clinical practice

A key finding in our review was the importance of exposure as an influential factor in medical students’ and junior doctors’ decision to pursue a career in otolaryngology. In the UK, otolaryngology is under-represented in medical school, typically comprising less than 1 per cent of the curriculum.^{Reference Davies and Elhassan20} The average duration of undergraduate otolaryngology placement in the UK varies between 8 to 13.7 days, with most UK medical graduates feeling they have not received enough otolaryngology undergraduate training.^{Reference Ferguson, Bacila and Swamy21} Similar findings are observed in Canada, with mean placement duration of 4.6 days, and in the United States, with placement durations showing wide variance between 4 hours and 4 weeks. Like the UK, a significant proportion of medical schools lack mandatory otolaryngology placements.^{Reference Ishman, Stewart, Senser, Stewart, Stanley and Stierer22–Reference Campisi, Asaria and Brown24} Furthermore, this appears to result in important omissions from the curriculum. A recent review of UK otolaryngology undergraduate curricula showed only two thirds of medical schools include taking an otolaryngology history, and only 58 per cent included tonsillitis.^{Reference Steven, Mires, Lloyd and McAleer25}

One possible way to increase exposure is through more widespread use of e-learning. E-learning carries several advantages: learners have control over the content, pace, time and place in which learning occurs.^{Reference Zehry, Halder and Theodosiou26} A wealth of otolaryngology e-learning resources exists online, illustrated by the recent development of a validated tool to assess the quality of educational websites in otolaryngology (the Modified Education in Otolaryngology Website assessment tool).^{Reference Yang, Hosseini, Mascarella, Young, Posel and Fung27} So how does e-learning compare to traditional teaching methods? A recent systematic literature review by Tarpada et al. assessed the efficacy of e-learning for otolaryngology education. This demonstrated that, when compared with standard teaching methods, the use of e-learning improved objective performance in academic or clinical measures or resulted in no difference in performance but higher student satisfaction.^{Reference Tarpada, Hsueh and Gibber28} But despite a growing body of evidence to support e-learning in otolaryngology, only a minority of UK medical schools are using it as a teaching method.^{Reference Steven, Mires, Lloyd and McAleer25}

Optional clerkships, or student selected components, can influence students’ career prospects.^{Reference Peel, Schlachta and Alkhamesi29} One study in our review demonstrated that a student clerkship in otolaryngology not only positively influenced a participant's decision to pursue otolaryngology as a career but also had predictive validity for application to otolaryngology residency.^{Reference Nellis, Eisele, Francis, Hillel and Lin15} In the UK, around two thirds of medical schools offer student selected components in otolaryngology, which is lower than that of Canada and USA where around 81 per cent and 89 per cent, respectively, offer selective clerkships in otolaryngology.^{Reference Boscoe and Cabrera-Muffly23,Reference Campisi, Asaria and Brown24} Wider availability of otolaryngology student selected components in the UK, particularly for those in the early years of medical school may help to increase interest in the field and application rates to higher surgical training posts. Student selected components also provide a clinical supervisor who can act as a mentor and positive role model. Medical students’ interest in surgery has been shown to significantly increase when partnered with a surgical mentor.^{Reference Peel, Schlachta and Alkhamesi29} Otolaryngology undergraduate leads should therefore be identifying motivated consultants and registrars who would make suitable mentors.

A recent systematic review of factors affecting surgery as a career choice demonstrated the critical importance of exposure during pre-clinical years. Peel et al. suggested that decisions about specialties of interest are formed during this period without dissuasion from considerations of work-life balance or lifestyle.^{Reference Peel, Schlachta and Alkhamesi29} A common way in which medical students are exposed to specialties at an early stage is through university societies, the significance of which should not be underestimated. A recent study of 276 Brazilian medical graduates assessed the impact of participating in a surgeon-led trauma society during medical school. This showed that over 90 per cent of participants believed that participating in the society provided knowledge and information that the medical school curriculum could not provide and had positively influenced their career choice.^{Reference Simões, Dorigatti, Silveira, Calderan, Rizoli and Fraga30} Although several undergraduate otolaryngology societies exist in the UK, not every medical school has one. Widespread availability of such societies and greater representation of otolaryngology within generic surgical societies should be facilitated by senior otolaryngologists in order to maximise undergraduate exposure.

Comparison with other reviews

In our review, neither a lack of same-sex role models or gender discrimination were identified as deterrents from pursuing a career in otolaryngology. In contrast, a gender-specific deterrence to a career in surgery for women has been clearly defined.^{Reference Peel, Schlachta and Alkhamesi29} This disparity may in part be explained by the larger consultant female workforce in otolaryngology, but also due to the greater propensity for less than full-time training within the specialty.^{Reference Wilmot, Davis and Carrie8,9} Although our review did not identify gender bias as a deterrent, there is evidence that it is present within otolaryngology.^{Reference Johnson31} Our study population was exclusively medical students and junior doctors not yet on higher surgical training programmes. Therefore, it may be that gender discrimination is only recognised and experienced during training within the specialty.

Lifestyle and work-life balance were identified as key influential factors for pursuing a career in otolaryngology. There is a lack of objective data from the UK regarding lifestyle and work-life balance in otolaryngology; however, a recent systematic review by Pulcrano et al. provided a comprehensive understanding of quality of life and burnout rates amongst surgeons practising in the USA. Analysis of eight studies specifically pertaining to quality of life in otolaryngology residents and attending physicians showed that they had some of the highest rates of career dissatisfaction amongst the surgical specialties.^{Reference Pulcrano, Evans and Sosin32} They found that 20 per cent of otolaryngologists reported low quality of life. This suggests there is perhaps a mismatch between medical students’ and junior doctors’ perceptions of otolaryngology and the reality of training and working within the specialty. It should be noted however that quality of life parameters were better for attending physicians than residents.^{Reference Pulcrano, Evans and Sosin32} In the USA, otolaryngology residents work around 71 hours per week on average.^{Reference Garcia-Rodriguez, Sanchez, Ko, Williams, Peterson and Yaremchuk33} In the UK, the European Working Time Directive together with the introduction of a new junior doctor's contract in 2016 limits average working hours to 48 per week, with a maximum of 72 hours in a 7-day period.³⁴ Further research is needed to ascertain whether the quality of life of otolaryngologists in the UK is comparable to that in the USA.

Implications for research

Given the limitations of our study, the results are not generalisable and no causation can be inferred. However, we hope that this review highlights where further research may be directed. Most of our review findings were based upon the responses of applicants to otolaryngology training programmes or students with intent to pursue a career in otolaryngology. More research to explore the views of those not considering otolaryngology as a career is required. Given the limited undergraduate exposure to otolaryngology, it is possible that we are losing potential otolaryngologists due to lack of awareness or knowledge about the specialty.