Anatomical bias in extrapulmonary tuberculosis research: Implications for knowledge structures and research evaluation

Abstract: -

Background: Bibliometric analyses in biomedical research often treat diseases as homogeneous entities, overlooking internal heterogeneity. In extrapulmonary tuberculosis (EPTB), diverse anatomical manifestations are typically aggregated, potentially obscuring structural imbalances in knowledge production. This study introduces the concept of anatomical bias, referring to systematic disparities in research attention and visibility across anatomical sites. 

Methods: We conducted a comprehensive bibliometric analysis of 3,963 Scopus-indexed publications on EPTB from 2000 to 2025. Publications were classified into seven anatomical categories. Analytical approaches included co-citation analysis, keyword co-occurrence, thematic evolution, collaboration network analysis, and ordered logistic regression to assess citation-based outcomes. Bibliometric impact was measured using normalized citation scores and journal prestige indicators. 

Results: The distribution of research output was highly skewed, with lymphatic (59.5%) and skeletal (17.7%) tuberculosis dominating the literature, while other anatomically significant forms each accounted for less than 6%. Knowledge structures were centered on disease-level themes such as diagnosis, treatment, and HIV co-infection, with anatomical topics remaining peripheral and weakly integrated. Regression analysis showed no significant citation advantage for dominant anatomical sites (OR=0.967), indicating that their prominence is driven by publication volume rather than higher impact. In contrast, international collaboration (OR=2.701) and larger author teams (OR=1.806) were strong predictors of higher citation outcomes. 

Conclusions: Anatomical bias in EPTB research is structural rather than performance-based, reflecting disparities in research attention and visibility rather than intrinsic scholarly impact. These findings highlight limitations of aggregated bibliometric indicators and support the need for anatomy-sensitive evaluation frameworks in heterogeneous biomedical fields.