Detecting and treating older adults who fall in an environment without others is essential. Millimeter-wave radar sensors do not have the disadvantage of invading user privacy like cameras, nor do they require users to wear them in real-time like wearable devices. Actual samples of older adults fall are difficult to collect, and it is unethical to require older adults to fall repeatedly to collect data. In addition, different body types and action patterns will inevitably reduce the model’s performance when new users use the model. In this paper, we constructed a fall detection model based on anomaly detection. The model is trained only using non-fall samples and detects falls as abnormal actions. The proposed model uses a domain generalization architecture based on domain feature alignment to extract domain-invariant features of the model, thereby improving the model’s generalization ability. In addition, we introduced the idea of denoising learning into the feature extractor and feature predictor to improve the model’s anti-interference ability. We conducted sufficient experiments to explore the effectiveness of the proposed method. When tested with new domain data, the proposed model has a true positive rate of 96.12%, a false positive rate of 0.97%, and an area under the receiver operating characteristic of 0.9979.
- Article type
- Year
- Co-author
Open Access
Research Article
Issue
To build a prediction model of the in-hospital death of patients with cardiac arrest.
This study is a retrospective analysis based on the medical information mart for intensive care-Ⅳ (MIMIC-Ⅳ)2.0. We gathered the information of patients above 18 years old, with cardiac arrest and intensive care unit (ICU) experience. A stepwise multi-variate logistic regression analysis was performed to filter variables, variables with P values < 0.05 were kept and enter as predictors of in-hospital death of patients with cardiac arrest. The model was evaluated with receiver operating characteristic (ROC) curve for discriminative power and with calibration curve for consistency. Finally, an online dynamic nomogram calculator was built to calculate the risk of in-hospital death.
This study included 1772 patients with cardiac arrest. The mean age of those patients was (64.93±16.52) years old, and 963 (54.3%) patients suffered in-hospital death. The factors of the prediction model for in-hospital death of cardiac arrest patients constructed based on multi-variate logistic regression included: potential cardiac disease diagnosis, age adjusted Chalson comorbidity index(CCI), body mass index (BMI), vital signs, lowest lactic acid and lowest Glasgow coma scale (GCS) during the first 24 hours after entering ICU, cardiac ultrasound examination, invasive mechanical ventilation and vasopressin utilization. The sensitivity and specificity of the prediction model were 73.1%(95% CI: 0.702-0.759) and 71.6%(95% CI: 0.683-0.745), respectively. Area under the ROC curve was 0.806(95% CI: 0.786-0.826).
The prediction model built in this study can properly predict the in-hospital death of patients with cardiac arrest.
京公网安备11010802044758号