Abstract

Hip fracture patients have a high risk of mortality after surgery, with 30-day postoperative rates as high as 10%. This study aimed to explore the predictive ability of preoperative characteristics in traumatic hip fracture patients as they relate to 30-day postoperative mortality using readily available variables in clinical practice. All adult patients who underwent primary emergency hip fracture surgery in Sweden between 2008 and 2017 were included in the analysis. Associations between the possible predictors and 30-day mortality was performed using a multivariate logistic regression (LR) model; the bidirectional stepwise method was used for variable selection. An LR model and convolutional neural network (CNN) were then fitted for prediction. The relative importance of individual predictors was evaluated using the permutation importance and Gini importance. A total of 134,915 traumatic hip fracture patients were included in the study. The CNN and LR models displayed an acceptable predictive ability for predicting 30-day postoperative mortality using a test dataset, displaying an area under the ROC curve (AUC) of as high as 0.76. The variables with the highest importance in prediction were age, sex, hypertension, dementia, American Society of Anesthesiologists (ASA) classification, and the Revised Cardiac Risk Index (RCRI). Both the CNN and LR models achieved an acceptable performance in identifying patients at risk of mortality 30 days after hip fracture surgery. The most important variables for prediction, based on the variables used in the current study are age, hypertension, dementia, sex, ASA classification, and RCRI.

Highlights

  • Hip fractures are one of the most common type of fractures in elderly people with an incidence that has doubled in the last 20 years and is expected to keep growing [1,2,3,4,5]

  • The current study aims to explore the predictive values of preoperative characteristics in traumatic hip fracture patients for 30-day postoperative mortality by using a convolutional neural network (CNN) and a logistic regression (LR) model, in addition to comparing the performance of the two algorithms

  • Because the area under the receiver operating characteristic (ROC) curve (AUC), a thresholdindependent metric, has been widely adopted by the medical society as a global measure of the predictive ability of the models [29], which avoids the potential arbitrariness associated with the selection of the threshold and is prevalence-independent [30], we reported AUC values of the models based on the model-predicted probabilities

Read more

Summary

Introduction

Hip fractures are one of the most common type of fractures in elderly people with an incidence that has doubled in the last 20 years and is expected to keep growing [1,2,3,4,5]. Hip fracture patients have a high risk of mortality after surgery, with 30-day postoperative rates as high as 10% [9]. With the goal of reducing the mortality rates in this patient population, it is vital to explore and understand the factors that potentially could predict postoperative mortality. This has previously been attempted using both logistic regression models and neural networks [10,11,12]. Outcome prediction models using neural networks could be useful tools in attempting to reduce mortality after hip fracture surgery by identifying modifiable risk factors. The current study aims to explore the predictive values of preoperative characteristics in traumatic hip fracture patients for 30-day postoperative mortality by using a convolutional neural network (CNN) and a LR model, in addition to comparing the performance of the two algorithms

Objectives
Methods
Results
Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call