Write-up and dissemination of undergraduate and postgraduate research at the University of Rwanda: a cross-sectional study

Introduction Research is essential in all areas of health development. However, medical students and residents frequently lack the time and training on performing research. This is especially prevalent in resource-limited settings. We aimed to compare the word counts of undergraduate and postgraduate dissertations with published projects in Rwanda, and to identify the proportion of postgraduate pediatric research projects that have been published since 2012. Methods Retrospective, cross-sectional study of undergraduate and postgraduate research dissertations at the University of Rwanda. Dissertations were then compared to randomly selected published papers of Rwandan research. Each IMRaD (Introduction, Methodology, Results and Discussion) section word count was compared using Student's t-test. Results 19/190 (10%) undergraduate dissertations and 22/41 (54%) postgraduate dissertations, were available in electronic format for word-count analysis. The mean total word count for postgraduate dissertations (5163 words) was significantly longer (p<0.001) than the randomly selected peer-reviewed journal articles (2959 words). Each section of the IMRaD structure of postgraduate dissertations was significantly longer than those of the control group. Undergraduates used a similar number of words to published papers, but used significantly more tables and figures. Of the 41 postgraduate dissertations, only four (10%), were published in peer-reviewed journals. Conclusion This is the first study to assess the writing style of Rwandan medical students and pediatric postgraduate residents. A simple step to increase dissemination of research findings would be for institutions to modify academic regulations so that students write-up in manuscript form rather than dissertation format.


Introduction
Research is vital in different avenues of health development and care delivery whilst also playing an essential role in the academic advancement of the researcher [1][2][3]. Research is defined as "critical and exhaustive investigation or experimentation, having for its aim the discovery of new facts and their correct interpretation, the revision of accepted conclusions, theories, or laws in the light of newly discovered facts, or the practical application of such new or revised conclusions, theories, or laws" [4]. It should, therefore, be part of the educational culture in medical training programs [5].
There are benefits to both medical students and residents, who shall be referred to as 'students' in this article. It is now evident that engaging students in research may lead to increased participation in research after completion of training [6,7]. However, due to the burdens of patient care, students are frequently too busy to find sufficient time for research activities [8,9]. Scholarly research activity programs are essential components of the modern undergraduate (UG) and postgraduate (PG) medical curriculum [10].
Research encourages students to develop the necessary competencies for future research activities and enter academic careers as physician-scientists [11]. Undertaking research can help develop transferable skills such as communication skills, time management, medical statistics, academic writing, systematic thinking, critical appraisal skills, information technology (IT) skills and how to practice evidence-based medicine [3,10]. In Rwanda, medical students and residents face various barriers to performing research, notably a lack of good mentorship and financial burdens; yet they have to undertake research in order to graduate from their training [5]. It is an aim of the University of Rwanda (UR) to improve the quality and quantity of their research output [12]. In 2014, thanks to intervention from the broader community, Rwanda was the most influential country for research in East Africa for published research output, with the UR being the second most influential University in East Africa [12].
Peer-reviewed published papers (control group) classically follow the IMRaD structure of Introduction, Methodology, Results and Discussion [13]. Undertaking a research dissertation is already a significant challenge to students. Therefore, progressing to the submission of the results in a peer-reviewed journal adds a new burden to the student's research journey and requires its own specific competencies. The language used in "writing-up for publication" can be foreign to the student and many students in this setting have had little opportunity or feedback in academic writing.
One study in Rwanda demonstrated that students needed formal training to develop these skills in order to improve the scholarly output of the participants [14]. To compound this, many academic institutions ask students to submit research in the form of a "dissertation" which may have other idiosyncrasies which are not consistent with manuscripts in peer-reviewed journals and dissertations may have specific requirements related to the institution. Publication is defined as "copies of a work or document distributed to the public by sale, rental, lease, or lending" [4]. In this study, we define publication as being articles that have gone through peer-review by a medical journal. Encouraging students to publish dissertation work is important as the future synthesis of research, in systematic reviews, is compromised if journal publications represent a biased selection of research projects conducted [15]. Many research projects, however, go unreported with one study finding that only 46% of US NIH-funded trials were published in a peer-reviewed, Medline-indexed journal within 30months of completion [16]. Furthermore, the research work represents not only the author but also the research institution, the funding body, and other researchers [17]. Submission for publication can equip the student in quality scientific writing, presentation skills and raising their knowledge and practice of scientific ethics [18]. Writing-up research in a concise manner is a fundamental step in the dissemination process. There is currently no  Table 1).

Discussion
The primary objective of this study was to compare the weighting of word counts for each IMRaD section between postgraduate (PG), undergraduate (UG), and published projects in Rwanda. We found that PG student dissertations used significantly more words than the published papers. This could be contributing, with other factors, to the low publication output of the PG students' research activities.
We found that only four (10%) PG pediatric research projects were published between 2012-2017. In Germany one study found that of 807 research projects approved by the IRB, 52% were published [21]. The authors of this study found that study design was not significantly associated with subsequent publication but that multicenter status, international collaboration, and a large sample size were positively associated with subsequent publication [21]. In our study, 73% of PG research projects were cross-sectional (Table 1).
However, in the German study [21], it was found that study methodology did not affect the proportion which were published, the right peer-reviewed journal, navigation of the submission process, English language, and "publishing language" [9]. This can then be followed by the disheartening rejection of a manuscript during the peer-review process. In addition, some journals require either submission, type editing and/or publication fees, which most students cannot afford in the resource-limited setting. These challenges are surmountable and findings from Rwanda have shown that confidence and skills in academic writing can improve through training and intervention [14]. All PG sections of the IMRAD were significantly longer in PG dissertations compared to peer-reviewed articles (control group).
We were unable to find any published articles from other institutions regarding this subject. We are postulating that one possible explanation for low rates of research dissemination in our population is that dissertations are not concise and therefore the final write-up is not an appropriate length for submission to a peer-reviewed journal. Dissertations involve a major educational element to the student's experience. The dissertation is submitted solely by the student as part of their assessment with supervisors giving "guidance" rather than tight editing. For publication, however, multiple authors in clear collaboration, according to a clearly defined journal format, edit the text. These skills of writing concisely also need to be learnt by students and they also need guidance in order to be able to do so. Being able to write in a concise, academic, publishable style, is an additional skill and should not weaken, but rather strengthen, the dissertation required for graduation. The primary focus of students is to undertake a piece of work as a requirement of their studies. Publication is a secondary opportunity that some students will exploit and it is not a requirement for graduation. Research authors publish for personal and altruistic reasons that are both valid. Personal reasons include increasing their personal academic profile, for personal development, to better understand a topic, to build their reputation and for career advancement. Altruistic reasons include raising awareness of the topic and research, to disseminate findings, and ultimately improve quality of care for patients. Regarding the ethics of non-publication; research is dependent on the willingness of participants to expose themselves to the risks involved [24]. The ethical justification for these risks is that society will eventually benefit from the knowledge

Conclusion
This is the first study to assess the writing style of Rwandan medical students and pediatric residents. The results should be used to strengthen research activities in Rwanda and further afield in the Sub-Saharan region so that dissemination of research findings will be increased.

Limitations:
We accessed 10% and 54% of UG and PG dissertations respectively and therefore our results may not be wholly representative of the whole cohort. We opportunistically selected the control group and this could have been subject to bias.
A potential source of bias and confounding was the change in  Some institutions may choose to modify academic regulations so that student's write-up in "manuscript" style and length rather than "dissertation" format. Tables and figure   Table 1: Baseline data on dissertations and peer-reviewed articles