Presentation Details
| Orthopedic Adaptation and Validation of the AWOL-S Delirium Risk Score Stephen Batter1, Dongmei Sun1, Huong Do1, Ahmed Deeb1, William Salerno1, Bella Mehta1, 2, Susan Goodman1, 2, Linda Russell1, 2. 1Hospital for Special Surgery, New York, NY, USA.2Weill Cornell Medicine, New York, NY, USA |
Abstract
BACKGROUND: Postoperative delirium (POD) is one of the most common complications following orthopedic surgery. The five-variable AWOL-S score (Age, American Society of Anesthesiologists Physical Status Classification System (ASA) level, spelling WORLD backwards, orientation, and surgery specific risk) predicts POD across general surgical populations with an area under the receiver operating characteristic curve (AUC-ROC) of 0.71. However, it has not been optimized for use at an orthopedic specialty hospital. We aimed to develop, validate, and adapt an orthopedic specific AWOL-S model into the electronic medical record (EMR) for real time delirium risk stratification.
METHODS: We performed a retrospective cohort study at a high-volume orthopedic center between June 2023 and January 2025. The five variables of the AWOL-S screen are entered into our EMR, which generates a score. The AWOL-S model was rederived via logistic regression using our institution’s data. Four features of the original model were kept; an orthopedic specific surgical risk score was newly implemented based on our institution’s historical rates of delirium. Delirium was cataloged via ICD-10 codes. New logistic regression coefficients were derived from a derivation set and tested on a validation set. A new probability cut point was designated using Youden’s J index to flag patients at elevated risk of POD.
RESULTS: A total of 26,042 surgical cases met inclusion criteria, with a mean age of 64.5 ± 13.9 years. There were 18,451 (70.8%) in the derivation set and 7,591 (29.2%) in the validation set. The model achieved an AUC-ROC of 0.828 in derivation, and 0.821 in validation. Patients were flagged at elevated risk of POD with a 1.80% probability of developing POD. There was a 1.6% delirium incidence across included cases.
CONCLUSIONS: The orthopedic-specific AWOL-S score demonstrates improved performance across a more homogenous orthopedic population. Utilization of this real-time tool can assist clinicians in optimizing patients’ perioperative management to reduce POD.
No part of this publication may be reproduced, distributed, or transmitted in any form or by any means, including photocopying, recording, or other electronic or mechanical methods, without the prior written permission of the author.
METHODS: We performed a retrospective cohort study at a high-volume orthopedic center between June 2023 and January 2025. The five variables of the AWOL-S screen are entered into our EMR, which generates a score. The AWOL-S model was rederived via logistic regression using our institution’s data. Four features of the original model were kept; an orthopedic specific surgical risk score was newly implemented based on our institution’s historical rates of delirium. Delirium was cataloged via ICD-10 codes. New logistic regression coefficients were derived from a derivation set and tested on a validation set. A new probability cut point was designated using Youden’s J index to flag patients at elevated risk of POD.
RESULTS: A total of 26,042 surgical cases met inclusion criteria, with a mean age of 64.5 ± 13.9 years. There were 18,451 (70.8%) in the derivation set and 7,591 (29.2%) in the validation set. The model achieved an AUC-ROC of 0.828 in derivation, and 0.821 in validation. Patients were flagged at elevated risk of POD with a 1.80% probability of developing POD. There was a 1.6% delirium incidence across included cases.
CONCLUSIONS: The orthopedic-specific AWOL-S score demonstrates improved performance across a more homogenous orthopedic population. Utilization of this real-time tool can assist clinicians in optimizing patients’ perioperative management to reduce POD.
No part of this publication may be reproduced, distributed, or transmitted in any form or by any means, including photocopying, recording, or other electronic or mechanical methods, without the prior written permission of the author.
Content Locked. Log into a registered attendee account to access this presentation.