Causal directed acyclic graph (DAG) to test the effect of treatment(s) in aneurysmal subarachnoid haemorrhage (ASAH).
No Thumbnail Available
All Authors
Berzuini C.
Anderson I.
Bulters D.
Toma A.
Walsh D.
Dulhanty L.
Galea J.
Patel H.
LTHT Author
Anderson, Ian
LTHT Department
Neurosciences
Neurosurgery
Neurosurgery
Non Medic
Publication Date
2024
Item Type
Conference Abstract
Language
Subject
BRAIN ISCHAEMIA , BIOMEDICAL RESEARCH , SUBARACHNOID HAEMORRHAGE , INTRACRANIAL PRESSURE
Subject Headings
Abstract
Background and aims: Randomized Controlled Trials are considered the most reliable method to assess the effect of medical interventions. Observational studies (OS) are considered less dependable becasue of selection and confounding. DAGs are useful in visualizing confounders and causal links between variables and can facilitate adjustments to estimate causal effects. Our study aimed to record and describe a DAG designed to assess the causal effects of treatment(s) in aSAH using OS. Method(s): Our DAG was collaboratively created by causal inference and aSAH experts. DAG describing how variables (presenting characteristics, interventions, patient responses) affect outcome for a typical aSAH patient was created. Each node of the DAG representing a one or more variables in the problem was associated with relevant definitions, and an explanation of reasons for inclusion and exclusion. The DAG was inspected for causality sequence and the biological plausibility of each variable. Result(s): Figure 1. DAG representing assumptions about how variables causally relate to Outcome acknowledging that each node is informed by everything known to that time. Unobserved variables and selection are acknowledged though (U-0) and (D-Admit). HTN- hypertension, WFNSWorld Federation of Neurological Sciences, mRS- modified Rankin Score, CVS- cardiovascular, GCS- Glasgow Coma Scale, Admit- Admission, H'cephalus- Hydrocephalus, ICP-intracranial pressure, DCI-delayed cerebral ischaemia, Rx-treatment, D-decision. Conclusion(s): The described DAG is designed to help guide data collection and analysis to evaluate the effect of treatments using observational data.
Journal
European Stroke Journal