Background

Sub-Saharan Africa accounted for 1.7% of the world’s total emergency care research output from 2010 to 2015, hugely disproportional to its 12% share of the global population. Sub-Saharan Africa research output suffers from a lack of both quality and quantity, despite international collaboration on approximately 40% of all scientific publications. Increasing evidence shows that African research institutions do not provide the degree of support needed by individual scientists to autonomously produce and disseminate high-quality academic literature. This is due to a host of factors, including barriers to access of scientific materials, lack of focus on research training, and poor mentorship infrastructure. A 2014 report by the World Bank—focusing on research output and citation impact in Sub-Saharan Africa—recommends that African stakeholders accelerate support to research and research-based education to bolster the local knowledge economy.

Importance

This study describes a journal-initiated platform to improve publication quantity and quality in Sub-Saharan Africa emergency care research. Since its launch in 2011, the African Journal of Emergency Medicine ( AfJEM ) has strived to provide context-specific, open-access, academic publishing within the region. Editors were faced from the outset with the systemic issues of low submission quantity and poor quality described above. Very few initiatives exist globally to address these issues in low-resource settings, and none that the authors found is specific to a single journal or medical specialty. Collaboration was identified in the run-up to the launch of the journal as a potential way to specifically improve regional submissions; Author Assist was subsequently offered from the journal’s very first issue (2011). Author Assist is a peer-to-peer mentorship service that pairs early-career African authors with an experienced, published, volunteer assistant. Through one-on-one mentoring, the assistant helps the author improve his or her article to a publishable standard for resubmission.

Goals of This Investigation

The aim of this study was to describe the role of Author Assist as a publication mentorship program by evaluating the publication work flow of the AfJEM . Additional objectives included describing the proportion of articles that were accepted, rejected, or withdrawn subsequent to author assistance; the average length of time from rejection to resubmission; and the number of publications in the 2 years after publication in AfJEM with author assistance.

Study Design and Setting

A descriptive report of the journal work flow including the AfJEM ’s Author Assist program between January 2011 and December 2015 was applied as a service evaluation. The report involved details of the editorial process of the journal, as well as anonymized details of authors and assistants involved in the Author Assist program. The editorial process of the AfJEM , including Author Assist, is detailed in Figure 1 .

Editorial process of the AfJEM .

The first step after submission of a new article to the AfJEM includes a brief quality check by the editor in chief. This check includes a cursory review of the overall quality of the content, as well as whether the topic falls within the scope of the journal. As with most journals, the article can be rejected at this pre–peer review stage should its content be judged to be of poor quality or the theme deemed outside the journal scope. Articles judged to be of a reasonable to good quality, falling within the journal’s scope, are accepted for peer review. A handling editor is assigned to select and assign the peer reviewers. Peer review is double blinded: this means that the identity of the author(s) is withheld from peer reviewers and vice versa. Once the peer reviewers have provided their feedback to the handling editor, the handling editor makes the decision about whether the article should be considered for publication, revised, or rejected. This decision, along with the handling editor’s and peer reviewers’ feedback, is conveyed to the editor in chief for a final publication decision. The editor in chief then makes the publication decision that is provided to the author (accept, revise, or reject).

An article is considered for Author Assist when it has been rejected (either in pre– or post–peer review assessment) according to the overall poor quality of the content but still falls within the scope of the journal. The author(s) of these articles are informed in the publication decision letter that the article has been rejected because of the quality of the content, but—should the author(s) choose to make use of it—that the journal will provide free author assistance to improve the quality and will allow a single resubmission, as a new submission, at a later stage (the comments from the peer reviewers and handling editor, if any, are also included). Rejected articles (those that were out of journal scope or otherwise deemed inappropriate for Author Assist) are not offered resubmission. If the author agrees, an appropriate assistant, matched by interest and availability, is chosen from a database of experienced, published, volunteer assistants. All assistants are peer reviewers for the journal and, before joining Author Assist, must have published at least 3 articles in an emergency care topic, acting as first author on at least one. Currently, the majority of authors who meet this criterion practice in North America and Europe. Prospective assistants currently have to submit a brief publication résumé and complete a short online evaluation before inclusion consideration. Initially, assistants were included according to reviewer performance and recommendation from the editorial board or existing assistants. Assistance ranges from simple language editing for non–native-English-speaking authors to substantial reviewing, reanalyzing, and reinterpreting of the original data reported in the article. A memorandum of understanding outlining the expectations is agreed to by the author(s) and assistant. Authorship for the assistant may be included, but only if the International Committee of Medical Journal Editors’ authorship criteria are satisfied. Of the 12 articles accepted on resubmission, 9 have assistants included as an author. The memorandum of understanding also allows authors to withdraw their article from the program at any time, or to petition a replacement assistant.

The reworked article is then resubmitted as a new article and the submission process described above repeats itself, with one exception: if the article is accepted for peer review, it is assigned to a new handling editor and peer reviewers. To minimize bias, handling editors are not informed that an assistant was involved or that the submission is a reworked resubmission. From their perspective, the resubmission is the first submission. Handling editors do not have access to the Author Assist database and do not have knowledge of the identities of assistants. Detailed information related to Author Assist is known only to the Author Assist coordinator, technical editor, and editor in chief, who oversees the process and ensures that peer reviewers from the first round are not reselected. As a rule, the journal’s editor in chief does not overturn reject decisions made by handling editors.

Selection of Participants

All authors offered Author Assist between January 2011 and December 2015, their assistants, and their articles were included. This former unintentionally included only first authors (who happened to coincidentally also be the corresponding authors in all instances). The involvement of coauthors in the Author Assist process is not known. Given the limitations of the retrospective sample, we included only the first authors in the analysis.

Methods of Measurement

Data were obtained through the Author Assist coordination database. This was supplemented by the first authors’ publication numbers and rates from Scopus (Elsevier, Amsterdam, The Netherlands). The number of Author Assist referrals and rejection type, as well as reviewer comments, were extracted from the back end of the Elsevier Editorial System (Elsevier), the article submission platform for AfJEM . Authors’ and assistants’ names, as well as article titles, were used to link articles from the Author Assist coordination database to both Scopus and the Elsevier Editorial System. Names for authors, assistants, and article titles were removed from the sample for analysis. Affiliations and country of origin data are not presented. The Elsevier Editorial System was also used to obtain journal decision detail. The study was approved by the research ethics committee of the authors’ institution.

Outcome Measures

The primary outcome of this study was describing the proportion of articles that were accepted subsequent to author assistance. The secondary outcomes were describing the proportion of articles that were rejected or withdrawn subsequent to author assistance, the mean length of time (days) from first rejection to subsequent resubmission, and the number of publications (by the first author) in the 2 years after publication in AfJEM with author assistance.

Primary Data Analysis

The main results are expressed as proportions of the journal’s work flow decisions. Articles that were accepted, rejected, withdrawn from Author Assist, or in progress at the writing of this article are presented. To contrast, the outcome (accept or reject) of general resubmissions from articles that did not require Author Assist are provided. Publications in the 2 years subsequent to publication after Author Assist are provided as all publications and first-author publications, with the total and median provided. Given the small size of the sample, more complex statistical tests were not included. Microsoft Excel (version 15.0; Microsoft, Redmond, WA) was used for analysis. In an attempt to describe the critical issues that led to rejection in the first instance (eg, design, statistical, quality and clarity of writing, grammatical), an informal review was made of reviewer comments of pre– and post–Author Assist reviews.