The importance of early supervised exposure to undergraduate medical research and publishing is well-established.1 Various efforts have been made to increase medical student participation in research. A recent review of the performance of a summer research programme for medical students over seven years found 42% of the 346 projects were published, with 30% of all participating medical students being co-authors of at least one peer-reviewed PubMed-indexed article.2 Several other mandatory and elective research training programmes have been introduced around the world.1
Although publication of medical student research findings in peer-reviewed journals is often favourably looked upon by prospective employers and specialty programme directors,1 relatively little is known about factors that influence successful publication by medical students. Recent research from the UK indicates that students with clinical academic supervisors (defined as researchers with a medical degree with or without research degree(s); ie, researchers with MB ChB or both MB ChB and PhD degrees) were more likely to obtain first class honours and produced significantly more peer-reviewed publications compared to students supervised by non-clinical academic supervisors.3 Potential reasons to explain this observation were, however, not explored by the authors, and the study relied on self-reporting of academic outcomes (degree grades and publication success) by medical students. Furthermore, it is not known how generalisable this finding might be to other contexts.
We have recently shown that around a third of Bachelor of Medical Sciences with Honours (BMedSc(Hons)) theses at the University of Otago, New Zealand, are published.4 Although encouraged, publishing project findings in peer-reviewed journals is not required for the award of the degree.5 Assuming these medical students were similarly capable and motivated, we queried the significance of the supervisory milieu on influencing the rate of publication. This prompted us to examine the impact of having a clinical academic supervisor (compared with full-time non-clinical academic supervisors) on the publication rate and degree grade of BMedSc(Hons) theses at the University of Otago, New Zealand.
To examine the impact of supervisor-related factors on publication rates of undergraduate medical student research, we conducted a secondary analysis of an existing dataset. A detailed description of the methods used to generate this dataset has been previously published.4
In brief, a list of accepted BMedSc(Hons) research projects along with students’ and supervisors’ names was identified from the electronic Otago University Research Archive.6 BMedSc(Hons) theses accepted between 1 January 2002 and 31 December 2011 were included in the analysis. Using students’ and supervisors’ first and last names, PubMed and Google Scholar databases were searched in October 2015. Using October 2015 as the cut-off date allowed for a minimum of nearly four years from thesis submission to full-text journal publication, in keeping with findings from previous research on undergraduate medical student thesis publication.4,7 For a publication to be considered relevant, the student had to be a co-author, and the article title matched the topic of the thesis. Degree grade was obtained using the publicly available ‘University of Otago Graduate Search database’.8
Data were presented as number (percentage), mean (± standard deviation) and median (interquartile range). An independent-samples Student t-test was used to determine differences between variables. Logistic regression analysis was used to test for associations. A p value of ≤0.05 was considered statistically significant. All statistical analyses were performed using the Statistical Package for Social Sciences software (SPSS Statistics®, version 184.108.40.206). Power analysis was calculated using G*Power software.9
Over a 10-year period (2002–2011), a total of 89 theses were submitted and accepted (mean 8.9 theses per year, range 4–14). Overall, 36 theses (40.4%) resulted in 55 publications in peer-reviewed journals. There was a total of 137 supervisors (median 1 supervisor per student, range 1–3), 32.1% of whom were clinical academics (n=44).
There were no statistically significant differences in the number of publications (t87=1.19, p=0.1) or degree grade (t86=-1.36, p=0.49) between students who were supervised by clinical academic researchers and those who were not. The mean number of publications supervised by clinical academics was 0.41 publications per student, compared with 0.64 for the others. First-class theses supervised by clinical academics were 53% compared with 41% for the others; there was no correlation found between the number of supervisors per thesis and the number of publications (r=0.13, p=0.3) or the grade with which the degree is awarded (r=0.11, p=0.22).
Given the negative result, post hoc power analysis was conducted to investigate a type II error. With 89 theses and a type I error set at 0.05, a two-tailed t-test achieved a power of 0.83 to detect an effect size of 0.3.
To our knowledge, this is the first study from New Zealand (and anywhere outside of the UK) to explore some of the academic sequelae of medical student research projects supervised by clinical academic supervisors. Findings from our study imply that having a clinical academic as a supervisor does not appear to confer additional benefits to the degree grade or scientific output. This differs from findings from a recent large study from the UK, which found students supervised by clinical academic researchers were more likely to obtain first class honours and produce significantly more poster presentations and publications compared with students supervised by full-time non-clinical academic researchers.3 However, that study relied on self-reporting by students, whereas this current study relies on publications that are identifiable through standard databases.
Publishability of research relates to many factors, particularly the topic, the importance of the findings and academic rigour. In addition, it can be argued that publication success by medical students is dependent on student-factors (eg, is the student motivated to pursue publishing their findings?), supervisor-factors (eg, does he/she allocate sufficient time and enthusiasm to help the student-author?) and programme-factors (eg, does the research program adequately prepare students and reward supervisors for successful publication?).1,10 For example, applicants to competitive residency programmes are often well-aware of what “academic factors” (eg, peer-reviewed publications) are needed to increase their odds of acceptance.10 Supervisor-factors, on the other hand, cannot be easily extrapolated since, for example, the selection process of the mentors (eg, prior experience with medical students) is not obtainable. Furthermore, academic clinicians often face competing demands11 of clinical care, teaching and/or research, which could potentially affect their effectiveness in supervising the student’s project to fruition (ie, successful publication). It is possible there are interacting factors—for example, clinical academic supervisors may suggest topics that are inherently more publishable, they may attract more motivated students or they may be more motivated to see a project through from inception to publication. These are areas that could be explored in future research. In the meantime, we suggest the claim that academic supervisors are more likely to produce peer-reviewed publications compared to non-clinical academics3 is not one that is backed up by our findings.
Clinical academic supervisors did not provide additional benefit or disadvantage in terms of degree grade and research output. A number of reasons could explain this finding. First, the University of Otago is a research-intensive university, and often ranks highly for research quality in New Zealand.12 This implies that University staff members (clinical and non-clinical) are equally likely to be actively involved in medical student research supervision and publication. Furthermore, a small number of students in our cohort were co-supervised by clinical and non-clinical researchers, which could potentially confound our findings.
Clinical academic supervisors appear just as effective as non-clinical full-time researchers in supervising medical students involved in undergraduate research. Future research should focus on examining the association between clinical academic supervisors and publication rates from other medical student research projects both curricular and extra-curricular, and could focus on reasons behind our observed association.