Assessing language bias in pediatric surgical systematic reviews: A meta-epidemiological study

Linguistic bias limits insight from non-English regions and potentially skews conclusions toward Western contexts. We evaluated the extent and impact of language-based exclusions in systematic reviews (SRs) within pediatric surgery, assessing how such limitations threaten the comprehensiveness and generalizability of evidence syntheses.

With health librarian oversight, Medline, Embase, and Global Health (via Ovid) were searched from 2020 to 2024 for SRs related to pediatric surgery. A representative sample of 392 articles was analyzed for their country and publisher of origin, practice of language limitation and the transparency of the exclusion process.

Of 82 included SRs, 48 (58.5 %) restricted inclusion to English-language studies, and 34 (41.5 %) incorporated non-English studies in their search strategies. Transparent reporting of language exclusions was found in 25 SRs (31.7 %), primarily in PRISMA flowcharts. Among transparently reporting SRs, the proportion of studies excluded due to language ranged from 0.17 % to 98.4 % (mean 9.5 %). international journals (n = 20) had a mean language-based exclusion rate of 6.0 % (median 0.6 %, range 0.11-98.41 %), compared to 1.3 % (median 0.23 %, range 0.02-5.53 %) among North American journals. After removing one extreme outlier, the mean exclusion rates were 1.13 % and 1.31 %, respectively (p = 0.87, Cohen's d = -0.09), indicating no significant regional difference.

Our study demonstrates that nearly 60 % of pediatric surgery SRs exclude non-English language studies, often without transparent reporting. This language bias risks underrepresenting global research and potentially skews evidence towards Western contexts. Multilingual inclusion, greater transparency, and utilization of modern translation tools may improve the equity and validity of future pediatric surgical research.

Level 2 evidence.