Defining and detecting poor outcome in heterogeneous surgical cohorts: a proof-of-concept composite distribution- and MCID-based method for study-arm-level evidence synthesis
MethodsX · 2026-12 · 17 · 104232
DOI: 10.1016/j.mex.2026.104232 ↗Kurz gefasst
Fachartikel in der Zeitschrift MethodsX, 2026-12. Verfasst von Kieran Dey und 2 weiteren Autor:innen. Beteiligte Einrichtung: Trauma and Orthopaedic Department, Nottingham University Hospitals NHS Trust, Nottingham, United Kingdom. Die Arbeit nennt 25 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).
Abstract (Originalsprache)
Surgical outcome studies are often too heterogeneous to pool and rarely report individual patient data, so which patients do badly, and why, cannot be answered directly. Aggregate data can show which study arms did badly and what those arms had in common, but only if poor outcome is defined in advance and applied the same way to every study. We describe a reproducible method for doing this at study-arm level when meta-analysis is inappropriate. It was developed in a single systematic review (17 study arms from 13 studies) and is presented as a proof of concept that needs testing in other datasets. In the worked example, three of four classified arms held under every alternative return-time threshold, but classification was sensitive to the scoring rule, the minimal clinically important differences (MCIDs) used and missing return times.•A composite poor-outcome definition combines distribution-based thresholds (the mean plus one standard deviation across arms) with published MCIDs, borrowed only under explicit criteria.•Each classification is traceable to the events behind it and is tested against alternative cut-offs, thresholds and leave-one-study-out derivations.•A Synthesis Without Meta-analysis (SWiM) moderator analysis reports each moderator's direction, size and confidence using explicit rating rules, at study-arm level only.
Schlagwörter
Stichwörter der Autor:innen
Autor:innen und Institutionen
- Trauma and Orthopaedic Department, Nottingham University Hospitals NHS Trust, Nottingham, United Kingdom.
- Academic Unit of Injury, Recovery and Inflammation Sciences, School of Medicine, NIHR Nottingham Biomedical Research Centre, University of Nottingham, Queen's Medical Centre, Lenton Road, Nottingham, NG7 2UH, United Kin…
Zitieren
APA
Dey, K., Sangha, B. K. S., & Norrish, A. (2026). Defining and detecting poor outcome in heterogeneous surgical cohorts: a proof-of-concept composite distribution- and MCID-based method for study-arm-level evidence synthesis. MethodsX, 17, 104232. https://doi.org/10.1016/j.mex.2026.104232
Vancouver
Dey K, Sangha BKS, Norrish A. Defining and detecting poor outcome in heterogeneous surgical cohorts: a proof-of-concept composite distribution- and MCID-based method for study-arm-level evidence synthesis. MethodsX. 2026;17:104232. doi:10.1016/j.mex.2026.104232
BibTeX
@article{Dey2026,
author = {Dey, Kieran and Sangha, Brandon Karamveer Singh and Norrish, Alan},
title = {Defining and detecting poor outcome in heterogeneous surgical cohorts: a proof-of-concept composite distribution- and MCID-based method for study-arm-level evidence synthesis},
journal = {MethodsX},
year = {2026},
volume = {17},
pages = {104232},
doi = {10.1016/j.mex.2026.104232}
}
EUVIMED