Causal inference in perioperative medicine observational research: part 1, a graphical introduction

Vijay Krishnamoorthy; Danny J N Wong; Matt Wilson; Karthik Raghunathan; Tetsu Ohnuma; Duncan McLean; S Ramani Moonesinghe; Steve K Harris

doi:10.1016/j.bja.2020.03.031

Causal inference in perioperative medicine observational research: part 1, a graphical introduction

Br J Anaesth. 2020 Sep;125(3):393-397. doi: 10.1016/j.bja.2020.03.031. Epub 2020 Jun 27.

Authors

Vijay Krishnamoorthy¹, Danny J N Wong², Matt Wilson³, Karthik Raghunathan⁴, Tetsu Ohnuma⁵, Duncan McLean⁶, S Ramani Moonesinghe³, Steve K Harris³

Affiliations

¹ Critical Care and Perioperative Epidemiologic Research (CAPER) Unit, Department of Anesthesiology, Duke University Hospital, Durham, NC, USA. Electronic address: [email protected].
² Department of Anaesthesia, Guy's and St Thomas' NHS Foundation Trust, London, UK.
³ Critical Care, University College London Hospitals NHS Foundation Trust, London, UK.
⁴ Critical Care and Perioperative Epidemiologic Research (CAPER) Unit, Department of Anesthesiology, Duke University Hospital, Durham, NC, USA; Anesthesiology Service, Durham Veterans Affairs Medical Center, Durham, NC, USA.
⁵ Critical Care and Perioperative Epidemiologic Research (CAPER) Unit, Department of Anesthesiology, Duke University Hospital, Durham, NC, USA.
⁶ Department of Anesthesiology, University of North Carolina at Chapel Hill, Chapel Hill, NC, USA.

PMID: 32600803
DOI: 10.1016/j.bja.2020.03.031

Abstract

Graphical models have emerged as a tool to map out the interplay between multiple measured and unmeasured variables, and can help strengthen the case for a causal association between exposures and outcomes in observational studies. In Part 1 of this methods series, we will introduce the reader to graphical models for causal inference in perioperative medicine, and set the framework for Part 2 of the series involving advanced methods for causal inference.

Keywords: causal inference; confounding; epidemiology; graphical models; observational research.

Publication types

Review

MeSH terms

Biomedical Research / methods*
Biomedical Research / statistics & numerical data
Humans
Models, Statistical*
Observational Studies as Topic / methods*
Observational Studies as Topic / statistics & numerical data
Perioperative Medicine / methods*
Perioperative Medicine / statistics & numerical data