Authorship inequalities in global health research: the IeDEA Southern Africa collaboration

Abstract

Background: The International epidemiology Databases to Evaluate AIDS conducts research in several regions, including in Southern Africa. We assessed authorship inequalities for the Southern African region, which is led by South African and Swiss investigators.

Methods: We analysed authorships of publications from 2007 to 2020 by gender, country income group, time and citation impact. We used 2020 World Bank categories to define income groups and the relative citation ratio (RCR) to assess citation impact. Authorship parasitism was defined as articles without authors from the countries where the study was conducted. A regression model examined the probability of different authorship positions.

Results: We included 313 articles. Of the 1064 contributing authors, 547 (51.4%) were women, and 223 (21.0%) were from 32 low-income/lower middle-income countries (LLMICs), 269 (25.3%) were from 13 upper middle-income countries and 572 (53.8%) were from 25 high-income countries (HICs). Most articles (150/157, 95.5%) reporting data from Southern Africa included authors from all participating countries. Women were more likely to be the first author than men (OR 1.74; 95% CI 1.06 to 2.83) but less likely to be last authors (OR 0.63; 95% CI 0.40 to 0.99). Compared with HIC, LLMIC authors were less likely to publish as first (OR 0.21; 95% CI 0.11 to 0.41) or last author (OR 0.20; 95% CI 0.09 to 0.42). The proportion of women and LLMIC first and last authors increased over time. The RCR tended to be higher, indicating greater impact, if first or last authors were from HIC (p=0.06).

Conclusions: This analysis of a global health collaboration co-led by South African and Swiss investigators showed little evidence of authorship parasitism. There were stark inequalities in authorship position, with women occupying more first and men more last author positions and researchers from LLMIC being 'stuck in the middle' on the byline. Global health research collaborations should monitor, analyse and address authorship inequalities.

Keywords: Health policies and all other topics.

Figure 1

Proportion of female authorships across the range of standardised authorship position, with a weighted linear regression line (A) and proportion of authorships by country income level with weighted cubic splines (B). The size of the circles is proportional to the number of authorships in each position. Lower-income countries include World Bank lower middle-income countries and low-income countries.

Figure 2

Proportions of first, last and other authorships over time for female authors (A), authors from high-income countries (HICs) (B), upper middle-income countries (UMICs) (C) and lower-income countries (LICs) (D). The size of the circles is proportional to the number of authorships. Lower-income countries include World Bank lower middle-income countries and low-income countries.