50th Annual Regional Anesthesiology and Acute Pain Medicine Meeting

Classification and stratification of patient pain archetypes following total knee arthroplasty: a machine learning approach

Justin Chew, MD, PhD; Junying Wang; Alexandra Sideris, PhD; Victoria Xu; Daniel Maalouf, MD, MPH; Stavros Memtsoudis, MD, PhD, MBA; Jashvant Poeran, MD, PhD; Jiabin Liu, MD, PhD
Department of Anesthesiology and Pain Prevention Research Center, Hospital for Special Surgery, New York, NY

-Heterogeneity in pain courses following total knee arthroplasty (TKA) remains poorly characterized

-We utilize machine learning to identify pain archetypes following TKA, determine important features for predicting pain outcomes, and classify patients by risk of severe post-op pain

-This is a retrospective, single center study including all patients undergoing TKA from April 1, 2021 through October 31, 2024 (n = 17,200)

-Clustered patients into two groups with markedly different pain trajectories after TKA (by pain scores, MMEs consumed, and pain consult rate), stratifying into patients with either well controlled or poorly controlled post-op pain

-Identified the most important predictive factors as age, inpatient status, genicular block, OR duration, BMI, prior opioid or gabapentin use, and EBL

-Overall, achieved a 61% accuracy rate of classification on testing data subset

50th Annual Regional Anesthesiology and Acute Pain Medicine Meeting

CLASSIFICATION AND STRATIFICATION OF PATIENT PAIN ARCHETYPES FOLLOWING TOTAL KNEE ARTHROPLASTY: A MACHINE LEARNING APPROACH

Poster Presenter

Authors

Affiliations

Part of Topic