ABSTRACT
The COVID-19 pandemic has made a dramatic impact on human life, medical systems, and financial resources. Due to the disease's pervasive nature, many different and interdisciplinary fields of research pivoted to study the disease. For example, deep learning (DL) techniques were employed early to assess patient diagnosis and prognosis from chest radiographs (CXRs) and computed tomography (CT) scans. While the use of artificial intelligence (AI) in the medical sector has displayed promising results, DL may suffer from lack of reproducibility and generalizability. In this study, the robustness of a pre-trained DL model utilizing the DenseNet-121 architecture was evaluated by using a larger collection of CXRs from the same institution that provided the original model with its test and training datasets. The current test set contained a larger span of dates, incorporated different strains of the virus, and included different immunization statuses. Considering differences in these factors, model performance between the original and current test sets was evaluated using area under the receiver operating characteristic curve (ROC AUC) [95% CI]. Statistical comparisons were performed using the Delong, Kolmogorov-Smirnov, and Wilcoxon rank-sum tests. Uniform manifold approximation and projection (UMAP) was used to help visualize whether underlying causes were responsible for differences in performance between test sets. In the task of classifying between COVID-positive and COVID-negative patients, the DL model achieved an AUC of 0.67 [0.65, 0.70], compared with the original performance of 0.76 [0.73, 0.79]. The results of this study suggest that underlying biases or overfitting may hinder performance when generalizing the model. © 2023 SPIE.
ABSTRACT
PURPOSE: The aim of this study was to evaluate changes in both mechanism and diagnoses of injuries presenting to the orthopaedic department during this lockdown period, as well as to observe any changes in operative case-mix during this time. METHODS: A study period of twelve weeks following the introduction of the nationwide "lockdown period", March 23rd - June 14th, 2020 was identified and compared to the same time period in 2019 as a "baseline period". A retrospective analysis of all emergency orthopaedic referrals and surgical procedures performed during these time frames was undertaken. All data was collected and screened using the 'eTrauma' management platform (Open Medical, UK). The study included data from a five NHS Foundation Trusts within North West London. A total of 6695 referrals were included for analysis. RESULTS: The total number of referrals received during the lockdown period fell by 35.3% (n=2631) compared to the same period in 2019 (n=4064). Falls remained proportionally the most common mechanism of injury across all age groups in both time periods. The proportion sports related injuries compared to the overall number of injuries fell significantly during the lockdown period (p<0.001), however, the proportion of pushbike related accidents increased significantly (p<0.001). The total number of operations performed during the lockdown period fell by 38.8% (n=1046) during lockdown (n=1732). The proportion of patients undergoing operative intervention for Neck of Femur (NOF) and ankle fractures remained similar during both study periods. A more non-operative approach was seen in the management of wrist fractures, with 41.4% of injuries undergoing an operation during the lockdown period compared to 58.6% at baseline (p<0.001). CONCLUSION: In conclusion, the nationwide lockdown has led to a decrease in emergency orthopaedic referrals and procedure numbers. There has been a change in mechanism of injuries, with fewer sporting injuries, conversely, there has been an increase in the number of pushbike or scooter related injuries during the lockdown period. NOF fractures remained at similar levels to the previous year. There was a change in strategy for managing distal radius fractures with more fractures being treated non-operatively.