Robust Mixed Modelling Approaches in the Analysis of ICU Patient Data

Date: Wednesday 03 May 2023, 1.00PM
Location: Hybrid
Queen's University Belfast, University Road, Belfast, Northern Ireland, BT7 1NN MAPTC/0G/005, a building, across the road and facing the McClay library.
Local Group Meeting


Share this event

Last talk in RSSNI Spring programme
 
 
I am pleased to announce our final talk of the Spring session on Wednesday,  May  3rd., at 1pm.  Dr. Lisa McFetridge of the Department of Applied Mathematics  & Statistics, Queen's University of Belfast, NI, UK will speak on  Robust Mixed Modelling Approaches in the Analysis of ICU Patient Data.
This is a hybrid event which you may attend in person (MAPTC/0G/005, building across the road facing the McClay library) or on-line using the MS Teams link below

Abstract:
While invasive mechanical ventilation (MV) is a life-saving therapy for critically ill patients, it does pose substantial risks. Ventilator-associated events (VAE) extend the duration of patients on MV, prolong ICU stays placing pressure on a scarce critical care resource, and have an estimated attributable mortality of 10% (Klompas, 2022). Unfortunately, this has been exacerbated by the Covid-19 pandemic and rise in incidence of nosocomial pneumonia. Early intervention and surveillance strategies can decrease the rate of VAEs and thus improve the prognosis of such patients (Wolffers, 2021).

By leveraging a new intensive care unit database within the Belfast Health & Social Care Trust, this talk will explore the use of robust mixed models to uncover the factors relating to changes in patients’ longitudinal profiles, allowing for the early identification of outlying individuals and those with a greater propensity for outlying observations. The prevalence of outliers in medical settings has been demonstrated in a wide variety of scenarios necessitating more robust approaches to maximise the discriminative ability of the models (McFetridge, 2021). The statistical theoretical developments discussed in this talk will provide a better representation of this situation, counteracting the detrimental impact of longitudinal outliers.

References
M. Klompas et al. 2022. DOI: https://doi.org/10.1017/ice.2022.88
O. Wolffers et al. 2021. DOI: https://doi.org/10.1038/s41598-021-01402-3
 L.M. McFetridge et al. 2021. DOI: https://doi.org/10.1002/bimj.202000253

All welcome!

MS Teams Link:
https://teams.microsoft.com/l/meetup-join/19%3ameeting_OWEwZTk1MTctMmVmOS00YWJmLTlmNTctYTViY2Y4NTIyYjkz%40thread.v2/0?context=%7b%22Tid%22%3a%22eaab77ea-b4a5-49e3-a1e8-d6dd23a1f286%22%2c%22Oid%22%3a%22a3367a32-021b-4f90-958b-d42e2498b6c6%22%7d


Gilbert MacKenzie &
Felicity Lamrock
RSSNI
 
Dr. Lisa McFetridge of the Department of Applied Mathematics  & Statistics, Queen's University of Belfast, NI, UK.
 
Contact Gilbert MacKenzie