Discussion
We conducted this systematic review to describe scientific writing and publishing interventions in LMICs and identify gaps in this literature to support capacity strengthening efforts. Research publications serve many critical purposes: they can inform health policy and health interventions, they are the currency for advancement in research careers, and they can enhance a country’s credibility and influence.12 39 Further, by publishing their scientific research, researchers in LMICs may advance professionally and thereby gain the funding and platform needed to set research priorities for their own country and region.12 40 41 Without publications, what is often a substantial financial investment in the research may be wasted.42–45
We identified only 20 writing and publishing interventions in LMICs that described their activities, evaluations and outcome measures in enough detail to enable analysis. Among these, we highlighted the approaches of interventions that reported the number of submissions and/or publications as outcomes because they offered quantifiable metrics of success. Notable approaches implemented by this subset of 14 interventions included an intervention length of approximately 5 days and the provision of one-on-one mentorship during and after the main intervention to support writing and publication. A study of mentorship at Makerere University College of Health Sciences in Uganda found that both mentors and mentees reported that lack of formal structure was a barrier to effective mentoring.46 The formal structure of writing and publishing interventions may be an ideal setting for focused, mutually beneficial mentoring relationships between experienced faculty and junior researchers or trainees.
The mentorship component of these interventions overcomes writing and publishing barriers reported in the literature, including poor access to scientific writing instruction, inadequate writing skills, a lack of dedicated time and opportunity for scientific writing, and lack of support from more experienced researchers.41 44 Several papers in our sample emphasised that their interventions benefited from a large number of facilitators with a high ratio of mentors to participants for writing and publishing.
The context in which interventions are implemented differs across settings, and writing and publishing interventions should be tailored to the preferences and needs of their intended participants. Needs assessments, which can help achieve this goal,47 were conducted in nearly half of interventions across our sample, and it is possible in some cases that they were performed but not reported. Papers in this analysis reported learning key information from needs assessments including why participants were motivated to publish and participate in the intervention; the experience and skill level of participants; and specific barriers to writing and publishing that participants faced. It is also valuable to learn about the resources available to participants, like budget, technology and library access.44 48 Needs assessments were reported less frequently in the Publications Reported group compared with the Other Interventions group; it is possible that reporting publications is related to conducting a needs assessment, but the mechanism by which these might be related is not clear. One of the Publications Reported interventions that did not report a needs assessment is the ongoing ‘eJCIndia’ journal club in India.17 A needs assessment may not be relevant for an ongoing journal club because the intervention can be modified as needs are identified across time, and because the group and their needs also likely change across time. Regardless, several papers identified the importance of conducting a needs assessment as part of the intervention planning process.
Our sample offered a rich collection of lessons learnt and suggestions for future interventions. Three lessons stood out: (1) the importance of keeping interventions short to allow researchers to balance work and family demands, (2) the value of a high ratio of mentors to mentees and (3) the need for plenty of time from senior researchers to provide detailed feedback on participants’ writing. Another recurring theme was the need for a budget to support open access fees and for accessing articles behind a paywall, as well as the need for high-quality internet connectivity. Limited time, lack of mentorship, and lack of a budget for open access fees, article access and technology infrastructure have been cited throughout the literature as barriers to publication.44 48
Though 64 articles were identified for inclusion in the analytical sample, 41 lacked sufficient detail about the content, structure and delivery of the intervention to allow for analysis. Future papers describing writing and publishing interventions should include details that will allow readers to understand and replicate an intervention. Key details include whether a needs assessment of scientific writing and publishing was conducted and how the findings informed programme design, length of the writing and publishing intervention, topics covered, whether one-on-one writing mentorship was provided, attendance and outcome measures. Articles should also include information about the residential affiliation of intervention leadership so that readers know whether leaders were local, from high-income countries, or if leadership transitioned from high-income countries to local leaders.
Careful evaluation that is contextualised to a group’s goals and particular setting will help shape scientific understanding of the power of writing and publishing interventions to strengthen capacity. Ideally, such research will offer a range of strategies to support efforts across different regions.49 In addition to providing details about the intervention’s approach and evaluation, tracking quantifiable metrics such as number of submissions and/or publications post intervention may offer insight into whether and which components of an intervention may have been most effective. Further, allocating enough follow-up time (eg, 2–3 years after an intervention) to track publications will allow greater capture of this key outcome.
Questions that will help inform future writing and publishing interventions include whether one comprehensive training can make a lasting impact in supporting individuals in greater overall productivity, including number of publications, grant applications and professional promotions.
Additionally, it would be helpful to understand whether training local senior researchers to lead, facilitate and otherwise participate in author trainings additionally supports their own professional productivity and advancement and local capacity.50 Finally, every intervention has a budget and reporting the cost of training a group of researchers can be helpful to others planning similar activities.
Writing and publishing a research paper occur in the later stages of the research process, building on a wide range of skills required to carry out research. These skills include the ability to search and understand the scientific literature in a given topic area, and to conceptualise and investigate a research question with an appropriate methodological approach. A research publication will not be successful unless these foundational skills are in place. Papers describing standalone writing and publishing interventions should describe the amount of prior research training participants had entering the programme and interventions that are part of a broader research capacity programme should describe the research training that the programme provided.
Limitations
Several limitations should be noted when interpreting our results. The analytical sample included five SORT-IT interventions19 21 32–34 and one precursor to SORT-IT (Union/MSF operational research training20); this core programme was repeated across time and in various settings. We considered these as separate interventions with the assumption that if they were published as original research, they were distinct. There was, however, overlap.
We did not capture every writing and publishing intervention conducted in LMICs, only those that were indexed in the five databases we searched and also provided sufficient detail to qualify for inclusion in our sample. Additionally, we did not capture interventions that were conducted but not published.
Our classification of interventions as Publications Reported or Other was based on whether they reported as outcomes submitted and/or published papers. It is possible that participants in interventions that did not report submissions and/or publications published papers on their own, and this was not captured in our classification. Additionally, publication can easily take 2 years or more and papers published after an intervention’s follow-up period were not included in our results.
Finally, our paper selection and data extraction process were subject to errors; for example, overlooking key details during data extraction. To minimise the risk of such errors, two authors independently reviewed papers at each stage, and a third author resolved any discrepancies.