Improved mortality prediction among intensive care unit (ICU) inpatients is a valuable and challenging task. Limited clinical data, especially with appropriate labels, are an important element restricting accurate predictions. Generative adversarial networks (GANs) are excellent generative models and have shown great potential for data simulation. However, there have been no relevant studies using GANs to predict mortality among ICU inpatients. In this study, we aim to evaluate the predictive performance of a variant of GAN called conditional medical GAN (c-med GAN) compared with some baseline models, including simplified acute physiology score II (SAPS II), support vector machine (SVM), and multilayer perceptron (MLP). Data from a publicly available intensive care database, the Medical Information Mart for Intensive Care III (MIMIC-III) database (v1.4), were included in this study. The area under the precision-recall curve (PR-AUC), area under the receiver operating characteristic curve (ROC-AUC), and F1 score were used to evaluate the predictive performance. In addition, the size of the dataset was artificially reduced, and the performance of the c-med GAN was compared in different size datasets. The results showed that c-med GAN achieves the best PR-AUC, ROC-AUC, and F1 score compared with SAPS II, SVM, and MLP when training in the full MIMIC-III dataset. When the size of the dataset was reduced, the prediction performances of both MLP and c-med GAN were affected. However, the c-med GAN still outperformed MLP on smaller datasets and had less degradation. The prediction of in-hospital mortality based on the c-med GAN for ICU patients showed better performance than the baseline models. Despite some inadequacies, this model may have a promising future in clinical applications which will be explored by further research.

Accurately predicting length of stay (LOS) is considered a challenging task for health care systems globally. In previous studies on LOS range prediction, researchers commonly pre-classified the LOS ranges, which were the same for all patients in the same classification, and then utilized a classifier for prediction. In this study, we innovatively aimed to predict the specific LOS range for each patient (the LOS range was different for each patient). In the modified deep neural network (DNN), the overall sample error (root mean square error (RMSE) method), the estimated sample error (ERR_pred method), the probability distribution with different loss functions (Dis_pred_Loss1, Dis_pred_Loss2, and Dis_pred_Loss3 method), and the generative adversarial networks (WGAN-GP for LOS method) are used for LOS range prediction. The Medical Information Mart for Intensive Care III (MIMIC-III) database is used to validate these methods. The RMSE method is convenient for LOS range prediction, but the predicted ranges are all consistent in the same batch of samples. The ERR_pred method can achieve better prediction results in samples with low errors. However, the prediction effect is worse in samples with larger errors. The Dis_pred_Loss1 method encounters a training instability problem. The Dis_pred_Loss2 and Dis_pred_Loss3 methods perform well in making predictions. Although WGAN-GP for LOS method does not show a substantial advantage over other methods, this method might have the potential to improve the predictive performance. The results show that it is possible to achieve an acceptable accurate LOS range prediction through a reasonable model design, which may help physicians in the clinic.

Background: There are few reports about the effect of glucocorticoids in the treatment of acute pancreatitis in humans. This study aims to evaluate the effect of glucocorticoids in the treatment of acute pancreatitis by propensity score matching analysis. Results: Acute pancreatitis patients admitted between 2014 and 2019 were collected from the database and analyzed. Included patients were divided into the glucocorticoids-used group (GC group) and the non-glucocorticoids-used group (NGC group) according to whether glucocorticoids were used. A total of 818 eligible patients were included in the final analysis. Seventy-six patients were treated with glucocorticoids, and 742 patients were treated without glucocorticoids. Before propensity score matching, the triglyceride levels (38.2 ± 18.5 vs. 20.2 ± 16.8, P < 0.05) and Acute Physiology and Chronic Health Evaluation II (APACHE II) scores (7.1 ± 2.5 vs. 4.5 ± 2.1, P < 0.05) at admission were significantly higher in the GC group than in the NGC group. The incidence of multi-organ failure (33.3% vs. 11.9%, P < 0.05) was significantly higher in the GC group than in the NGC group. Patients in the GC group showed a positive balance of fluid intake and output over 72 h. After 1:1 propensity score matching, 59 patients from each group (GC and NGC) were included in the analysis. There were no significant differences in age, sex, body mass index, triglycerides, or APACHE II scores between the two groups (P > 0.05), and the patients’ clinical outcomes were reversed. The proportion of patients with organ failure (40.7% vs. 52.5%, p < 0.05) and multi-organ failure (35.0% vs. 67.7%, P < 0.05) was significantly lower in the GC group than in the NGC group. Furthermore, patients in the GC group had significantly shorter lengths of hospital stay (12.9 ± 5.5 vs. 16.3 ± 7.7, P < 0.05) and costs (25,348.4 ± 2512.6vs. 32,421.7 ± 2813.3, P < 0.05) than those in the NGC group. Conclusions: This study presents preliminary confirmation of the beneficial effect of glucocorticoids in the treatment of acute pancreatitis. More high-quality prospective studies are needed in the future.

This study aimed to investigate whether mobile health (mHealth)-based transitional care, which utilized WeChat, could benefit patients undergoing hepatectomy. From January 2019 to December 2020, 807 patients who underwent hepatectomy in a tertiary care hospital were included in the study. Patients were divided into a transitional care group (TC group) ( n = 238) and a control group ( n = 569) based on whether they participated in mHealth-based transitional care, which utilized the WeChat application. Depending on the type of variables, t-tests, χ² tests, and other statistical methods were used to compare differences between groups. Propensity score matching analysis was conducted for factors that differed in basic characteristics. After 1:1 matching, 238 patients were included in each group. Less readmission within 90 days of discharge, better quality of life, and higher satisfaction were found in the TC group both before and after matching ( p < .05). This study showed the potential of applying mHealth-based transitional care among post-hepatectomy patients.

This study aimed to develop a nomogram for predicting the risk of pressure injury (PI) in adult patients undergoing abdominal surgery and validate its effectiveness among these patients. This study retrospectively included 11,247 adult patients, who underwent abdominal surgery and postoperative supervision in ICU, in a tertiary care hospital in western China between January 2017 and December 2020. All datasets were extracted from the patient’s medical records and randomly divided into the training cohort (8,997) and the validation cohort (2,250) by 8:2. The univariable logistic regression was used to select potentially relevant features. Then, multivariable logistic regression was also conducted and utilized to establish the nomogram. The nomogram was compared with the Braden scale for predicting PI in the validation cohort through the area under the curve (AUC) of the receiver operator characteristic (ROC) curve, Hosmer-Lemeshow (H-L) test, and decision curve analysis (DCA). 873 (7.8%) patients suffered PIs. Logistic regression analysis showed that time of operation, weight, type of operation, albumin, and Braden scale score were independent risk factors for PI. A nomogram integrating five selected characteristics was constructed. The AUC of the ROC curve for the nomogram was 0.831, with a specificity of 85.2% and sensitivity of 63.7%. The AUC of the ROC curve for the Braden scale was 0.567, with a specificity of only 33.0%. The P-values of the H-L test were 0.45 (nomogram) and 0.22 (Braden scale), both indicating good calibration. The DCA also displayed that the nomogram had better predictive validity. Compared with the Braden scale, the nomogram showed a better predictive performance. This nomogram is informative and has the potential to better guide caregivers for risk stratification and prevention of PI, although it requires further validation.

Aim: Nurse‐led transitional care programmes (TCPs) have been shown to facilitate patient recovery in several illness settings, but its role among patients discharged with T‐tubes remains uncertain. The aim of the study was to investigate the effects of a nurse‐led TCP among patients discharged with T‐tubes.Design: This retrospective cohort study was conducted at a tertiary medical centre.Methods: From January 2018 to December 2020, a total of 706 patients discharged with T‐tubes after biliary surgery were included in the study. Patients were divided into a TCP group (n = 255) and a control group (n = 451) based on whether they participated in a TCP. The baseline characteristics, discharge readiness, self‐care ability, transitional care quality and quality of life (QoL) were compared between the groups.Results: Self‐care ability and transitional care quality were significantly higher in the TCP group. Patients in the TCP group also exhibited improved QoL and satisfaction. The results suggest that the incorporation of a nurse‐led TCP among patients discharged with T‐tubes after biliary surgery is feasible and effective.No Patient or Public Contribution.