Perioperative medicine & outcomes
Directed acyclic graphs to minimise bias and optimise causal inference in Sprint National Anaesthesia Project (SNAP-3): an observational cohort study of frailty, multimorbidity, and delirium in older surgical patients
Authors' conclusion
DAGs provide a transparent framework for statistical decision-making in observational research.
Study context
- Population
- Older surgical population (SNAP-3)
Methodological detailsdirected acyclic graphs, 'Evidence synthesis for constructing directed acyclic graphs' approach, independent clinical expert input
- Directed acyclic graphs
- 'Evidence synthesis for constructing directed acyclic graphs' approach
- Independent clinical expert input
Main finding
As reported in the abstract
- Directed acyclic graphs (DAGs) created for exposure-outcome relationships between frailty, multimorbidity and delirium (as exposure) and postoperative outcomes (length of hospital stay, delirium, morbidity, mortality)
- DAGs provide visual representations of assumptions made and an objective approach to appropriate statistical adjustments
- Key nodes in all DAGs: age, dementia, genetic predisposition, hearing and visual impairment, LOS, malignancy, operative severity, polypharmacy, postoperative perioperative medicine service, preoperative clinic review, sex, social deprivation, urgency; delirium, frailty, multimorbidity, interaction, morbidity as exposures or outcomes in certain DAGs
Observational data: associations, not effects.
Original publication
Directed acyclic graphs to minimise bias and optimise causal inference in SNAP-3: an observational cohort study of frailty, multimorbidity, and delirium in older surgical patients. Br J Anaesth. Published 27 May 2025. doi:10.
Composed by AI from the published abstract and checked against it. How Briefs are made →