EUVIMEDEUVIMED entdecken
Crossref

Redundancy in Evidence Synthesis: Overlap, Waste, and Distortion in the Secondary Evidence Ecosystem

Chen Chengxun, Jianghong Yao, Yi Zhou

JMIR Metascience and Research Integrity · 02.11.2026 · 1 · e104150

DOI: 10.2196/104150 ↗

Kurz gefasst

Fachartikel in der Zeitschrift JMIR Metascience and Research Integrity, online erschienen am 02.11.2026. Verfasst von Chen Chengxun und 2 weiteren Autor:innen. Die Arbeit nennt 60 Literaturangaben.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Systematic reviews and meta-analyses are cornerstones of evidence-based medicine, yet the proliferation of overlapping and redundant syntheses has become a critical meta-research concern: it wastes finite research resources, confuses clinicians and guideline developers, and particularly in tertiary syntheses such as overviews and umbrella reviews, risks distorting the perceived weight of clinical evidence through double-counting of primary data. Drawing on 3 decades of meta-epidemiological, methodological, and policy literature, and treating the COVID-19 pandemic as an illustrative extreme case, this Viewpoint proposes a 3D conceptual framework comprising production-side drivers, propagation-side distortions, and governance-side failures that integrates previously fragmented observations into a unified systems model. We argue that redundancy is best understood not as a failure of individual review teams but as a structural pathology of the evidence ecosystem, sustained by misaligned incentives and by feedback loops through which weak governance enables excess production whose downstream distortions further erode oversight. The framework explains why single-dimension interventions have limited effect and maps coordinated intervention levers to each stakeholder group: review methodologists and meta-researchers, journal editors and peer reviewers, registration platform operators, and research funders and policymakers. The framework also locates the boundary between warranted replication and wasteful duplication—a distinction that current overlap metrics quantify without judging. We propose that a synthesis be treated as justified when its authors can name the decision it would alter and the element of the question existing syntheses leave unresolved, and we indicate how that judgment could be applied prospectively. We further specify a technically concrete proposal for an AI-assisted evidence adequacy assessment system, combining semantic similarity modelling, PICO (population, intervention, comparison, outcome) entity extraction, and evidence gap quantification, with a 3-tier graded output and integration points at PROSPERO preregistration, journal editorial triage, and funder proposal screening; we emphasize that this system is a specification awaiting validation—not a deployable tool. Against this diagnosis, we assess 4 intervention strategies against feasibility, evidence, scalability, and unintended consequences, and identify mandated preregistration scoping combined with AI-assisted adequacy assessment as the most tractable near-term pairing. We state plainly what bounds the argument: the framework is a conceptual synthesis rather than a systematic synthesis, the quantitative estimates cited are illustrative rather than exhaustive and derive from studies with heterogeneous definitions and sampling frames, and the causal relationships posited among the 3 dimensions remain to be tested empirically. Consistent with our conclusion, we contend that reform should aim not at minimizing redundancy per se but at calibrating the evidence ecosystem, preserving the epistemically valuable function of independent replication while suppressing structural waste. Since production, propagation, and governance are linked as a system, durable change requires coordinated action across registries, journals, funders, and academic institutions rather than isolated reform by any single stakeholder.

Zitieren

APA

Chengxun, C., Yao, J., & Zhou, Y. (2026). Redundancy in Evidence Synthesis: Overlap, Waste, and Distortion in the Secondary Evidence Ecosystem. JMIR Metascience and Research Integrity, 1, e104150. https://doi.org/10.2196/104150

Vancouver

Chengxun C, Yao J, Zhou Y. Redundancy in Evidence Synthesis: Overlap, Waste, and Distortion in the Secondary Evidence Ecosystem. JMIR Metascience and Research Integrity. 2026;1:e104150. doi:10.2196/104150

BibTeX
@article{Chengxun2026,
  author = {Chengxun, Chen and Yao, Jianghong and Zhou, Yi},
  title = {Redundancy in Evidence Synthesis: Overlap, Waste, and Distortion in the Secondary Evidence Ecosystem},
  journal = {JMIR Metascience and Research Integrity},
  year = {2026},
  volume = {1},
  pages = {e104150},
  doi = {10.2196/104150}
}
Kostenlos starten →