Automated Pain Assessment Increases Nurses’ Working Efficiency in elderly with hip fractures

Abstract

Abstract The aim of this study was to design a modified ResNet 50 algorithm to achieve automatic classification model for pain expressions by elderly patients with hip fractures. We built a dataset by combining the advantages of deep learning in image recognition, using a hybrid of the Multi-Task Cascaded Convolutional Neural Networks (MTCNN). ResNet 50-based network framework utilized transfer learning to implement model function. We performed Bayesian optimization on the hyperparameters in the learning process. We calculated intraclass correlation between visual analog scale scores provided by nurses independently and those provided by pain expression evaluation assistant.The automatic pain expression recognition model for elderly patients with hip fractures, constructed using the algorithm, achieved an accuracy of 99.6% on the training set, 98.7% on the validation set, and 98.2% on the test set. The substantial kappa coefficient of 0.683 confirmed the PEEA's efficacy in pain assessment.This study demonstrates that the modified ResNet 50 algorithm with Bayesian optimization can be used to construct an automatic pain expression recognition model for elderly patients with hip fractures with higher accuracy.