Perioperative medicine & outcomes

Dynamic Nomogram for Predicting the Risk of Perioperative Neurocognitive Disorders in Adults

Anesth AnalgPublished Prediction model development and validation study

Authors' conclusion

We developed a simple and rapid online tool to preoperatively screen patients' risk of PND using GLM based on multicenter data

Study context

Population
Surgical patients older than 45 years undergoing GA without history of PND, 3 academic hospitals in China; external validation on multiparameter intelligent monitoring in intensive care (MIMIC) Ⅳ1,051 multicentre + 2884 MIMIC Ⅳ analysed
Comparison
General linear model (GLM) vs artificial neural network (ANN) vs naive Bayes (NB) models for predicting PND
Outcome
Discrimination for PND (ROCAUC, PRAUC, Brier score, IPA, sensitivity, specificity)
Methodological detailsmulticentre, machine learning, case-control sampling 3:1, external validation, Online tool with 10 routine variables developed for preoperative screening
  • Multicentre
  • Machine learning
  • Case-control sampling 3:1
  • External validation
  • Online tool with 10 routine variables developed for preoperative screening

Performance

As reported in the abstract

GLM performance on validation-1 dataset
Test or modelROCAUCPRAUCSensitivitySpecificityBrier scoreIndex of prediction accuracy (IPA)
GLM0.874 (95% CI 0.833-0.915)0.685 (95% CI 0.584-0.786)72.6% (95% CI 61.4%-81.5%)84.4% (95% CI 79.3%-88.4%)0.13144.7%

GLM

ROCAUC0.874 (95% CI 0.833-0.915)
PRAUC0.685 (95% CI 0.584-0.786)
Sensitivity72.6% (95% CI 61.4%-81.5%)
Specificity84.4% (95% CI 79.3%-88.4%)
Brier score0.131
Index of prediction accuracy (IPA)44.7%

Further findings

GLM performance on validation-2 dataset (MIMIC Ⅳ)
0.761 (95% CI 0.712-0.809) · ROCAUC · PRAUC 0.475 (95% CI 0.370-0.581) · Brier score 0.053 · IPA 76.8%

Original publication

Dynamic Nomogram for Predicting the Risk of Perioperative Neurocognitive Disorders in Adults. Anesth Analg. Published 18 October 2023. doi:10.1213/ane.0000000000006746

Composed by AI from the published abstract and checked against it. How Briefs are made →

Get them every week

New studies on this topic in My Week and in your Monday e-mail.

We e-mail you a sign-in link and a code. No password, no spam.