Screening process

We reviewed articles in three stages: title screening, abstract review and full-text review. Two independent reviewers (CEB and MK) evaluated articles at each stage. Disagreements were resolved by a third independent reviewer (ETA).

In the title screening stage, articles were excluded if they were obviously irrelevant. In the abstract review stage, articles were excluded if they met any of the following criteria: (1) the writing or publication intervention described in the article was not implemented in an LMIC; (2) the paper did not examine an actual intervention (eg, reviews, editorials, frameworks); (3) the intervention described in the article focused on healthcare delivery or a clinical intervention. Articles focusing on healthcare delivery or clinical interventions were specifically excluded during the abstract review stage because reviewers were confident that such articles did not include interventions pertaining to writing or publishing. Articles on other topics sometimes included information about writing or publishing in the body of the paper but not in the abstract, so they were evaluated during full-text review to minimise the risk of improper exclusion. If the reviewer was unsure of whether the intervention included information about scientific writing or publishing, they voted to include it in the full-text screening stage.

In the full-text review stage, an article was excluded if it met any of the following criteria: (1) it met any of the exclusion criteria from the abstract review stage; (2) the article described an intervention that did not provide instruction specific to scientific writing or publishing; (3) the article described a research capacity strengthening intervention but it did not describe a writing intervention component (citing publications as a desired or measured outcome alone was not considered a writing intervention); (4) the article described an individual mentoring programme (rather than a group programme) or (5) the article described a writing and/or publication intervention embedded within a degree granting programme. Writing retreats were included if they offered a structured writing or publishing intervention that met all inclusion criteria. Individual mentoring programmes that connect researchers with mentors without a structured intervention protocol were excluded; however, interventions that met our inclusion criteria that included a one-on-one mentorship component were included. Interventions embedded within degree granting programmes were excluded because they are not broadly accessible to researchers.

During full-text screening, we determined that articles describing 41 interventions provided insufficient detail to enable further analysis. For example, several articles simply stated that manuscript workshops were held, without describing the content of the intervention or its evaluation.

During data abstraction and analysis, the intervention was considered the unit of analysis. In some cases, more than one paper was published about a particular intervention. When this occurred, we included all of the relevant papers identified through our search in our analytical sample and information from each article was synthesised to describe that single intervention. The six papers describing Structured Operational Research Training Initiative (SORT-IT) interventions were a notable exception to this rule. Though the overall intervention was similar among the six papers, we considered each paper to be a separate intervention because each described an intervention implemented in a distinct setting and context.

This systematic review process identified 64 articles for inclusion. The final analytical sample contained 23 articles describing 20 interventions (figure 1). If a paper in our main sample cited an additional reference that described details about the intervention, pertinent details from that article were included in our analysis, but the article was not formally added to our analytical sample. Such additional references are cited in our tables and labelled ‘supporting papers.’

Process for identifying articles for inclusion in systematic review of writing and publishing interventions.