Search

Background: Experiencing a hyperglycaemic crisis is associated with a short- and long-term increased risk of mortality. We aimed to develop an explainable machine learning model for predicting 3-year mortality and providing individualized risk factor assessment of patients with hyperglycaemic crisis after admission. Methods: Based on five representative machine learning algorithms, we trained prediction models on data from patients with hyperglycaemic crisis admitted to two tertiary hospitals between 2016 and 2020. The models were internally validated by tenfold cross-validation and externally validated using previously unseen data from two other tertiary hospitals. A SHapley Additive exPlanations algorithm was used to interpret the predictions of the best performing model, and the relative importance of the features in the model was compared with the traditional statistical test results. Results: A total of 337 patients with hyperglycaemic crisis were enrolled in the study, 3-year mortality was 13.6% (46 patients). 257 patients were used to train the models, and 80 patients were used for model validation. The Light Gradient Boosting Machine model performed best across testing cohorts (area under the ROC curve 0.89 [95% CI 0.77–0.97]). Advanced age, higher blood glucose and blood urea nitrogen were the three most important predictors for increased mortality. Conclusion: The developed explainable model can provide estimates of the mortality and visual contribution of the features to the prediction for an individual patient with hyperglycaemic crisis. Advanced age, metabolic disorders, and impaired renal and cardiac function were important factors that predicted non-survival. Trial Registration Number: ChiCTR1800015981, 2018/05/04. [ABSTRACT FROM AUTHOR]

Mortality prediction based on electronic medical records is crucial for treatment decisions of shock patients in the ICU. Although clinical data on such patients often contain many missing values, the multi-view property of medical data could compensate for such missing information. Traditionally, mortality prediction models are built as two-stage approaches with additional data imputation steps, leading to issues in which the local optimal model at each step may not necessarily obtain a globally optimal solution. To overcome this problem, we conducted a multi-centre study using real-world data and aimed to develop an end-to-end mortality prediction model for shock patients. A Multi-task Oriented Diffusion Model (MODM) is proposed to fill in missing values and predict mortality simultaneously. Specifically, the model incorporates label information from different tasks to guide the optimal direction and effectively reduce uncertainty in the diffusion process. In addition, we propose a self-adjusting training strategy that balances the convergence rates among different tasks. The two largest well-known ICU datasets were used in this study, where 14,278 shock patients from eICU-CRD (2018) were included in the internal experiment, and 5,310 shock patients from MIMIC-IV (2012) were used as an external test. Compared with 14 state-of-the-art methods, the proposed model achieved the best performance with an AUC of 0.7998 in mortality prediction and notably good performance in terms of RMSE (0.0058, 0.0034, 0.0030, 0.0027) and MAE (0.3959, 0.4358, 0.4975, 0.5435) at random missing rates (10%, 30%, 50%, 70%, respectively) in the data imputation stage. The experimental results indicate the superiority of the proposed end-to-end MODM algorithm in handling real-world data. We released our code at. • The designed multi-task oriented diffusion model simultaneously fills in the missing data and predicts the mortality of shock patients via an end-to-end approach. • The developed self-adjusting strategy enables a stable and optimised convergence training process for multi-task learning, which avoids the problem of local optima. • The proposed method could cope with incomplete data in the real world to build a prediction model, which augments the number of training data. [ABSTRACT FROM AUTHOR]

Hand infection is a rare complication in patients with diabetes. Its clinical outcomes depend on the severity of hand infection caused by bacteria, but the difference in bacterial species in the regional disparity is unknown. The purpose of this study was to explore the influence of tropical and nontropical regions on bacterial species and clinical outcomes for diabetic hand. A systematic literature review was conducted using PubMed, EMBASE, Web of Science, and Google Scholar. Moreover, the bacterial species and clinical outcomes were analyzed with respect to multicenter wound care in China (nontropical regions). Both mixed bacteria (31.2% vs. 16.6%, p = 0.014) and fungi (7.5% vs. 0.8%, p = 0.017) in the nontropical region were significantly more prevalent than those in the tropical region. Staphylococcus and Streptococcus spp. were dominant in gram-positive bacteria, and Klebsiella, Escherichia coli, Proteus and Pseudomonas in gram-negative bacteria occupied the next majority in the two regions. The rate of surgical treatment in the patients was 31.2% in the nontropical region, which was significantly higher than the 11.4% in the tropical region (p = 0.001). Although the overall mortality was not significantly different, there was a tendency to be increased in tropical regions (6.3%) compared with nontropical regions (0.9%). However, amputation (32.9% vs. 31.3%, p = 0.762) and disability (6.3% vs. 12.2%, p = 0.138) were not significantly different between the two regions. Similar numbers of cases were reported, and the most common bacteria were similar in tropical and nontropical regions in patients with diabetic hand. There were more species of bacteria in the nontropical region, and their distribution was basically similar, except for fungi, which had differences between the two regions. The present study also showed that surgical treatment and mortality were inversely correlated because delays in debridement and surgery can deteriorate deep infections, eventually leading to amputation and even death. [ABSTRACT FROM AUTHOR]

Background: Previous observational studies have indicated a correlation between the gut microbiota and influenza; however, the exact nature of the bidirectional causal connection remains uncertain. Method: A two-way, two-sample Mendelian randomization (MR) study was conducted to evaluate the possible causal connection between the gut microbiota and the two outcomes of influenza (pneumonia without influenza and influenza pneumonia). The statistical analysis of gut microbiota is derived from the information of the most extensive meta-analysis (GWAS) conducted by the MiBioGen Alliance, encompassing a sample size of 18,340.The summary statistical data for influenza (not pneumonia, n = 291,090) and influenza pneumonia (n = 342,499) are from GWAS data published by FinnGen consortium R8.Estimate and summarize Single-nucleotide polymorphisms (SNPs) using Inverse variance weighted (IVW), MR Egger, and Weighted median (WM) in bidirectional MR analysis. To assess the heterogeneity, horizontal pleiotropy, and stability of SNPs, we employed Cochran's Q test, MR Egger intercept test, and sensitivity analysis. Result: The IVW analysis indicated that there was a significant association between influenza infection and five bacterial taxa. Additionally, the abundance changes of seven gut microbiota were found to be causally related to influenza infection. In addition, seven bacterial taxa showed a significant association with the occurrence of influenza pneumonia. The findings from the WM analysis largely support the outcomes of IVW, however, the results of MR egger analysis do not align with IVW. Furthermore, there is no proof to substantiate the cause-and-effect relationship between influenza pneumonia and the composition of gut microbiota. Conclusion: This analysis demonstrates a possible bidirectional causal connection between the prevalence of particular gut microbiota and the occurrence of influenza infection. The presence of certain gut microbiota may potentially contribute to the development of pneumonia caused by influenza. Additional investigation into the interaction between particular bacterial communities and influenza can enhance efforts in preventing, monitoring, and treating influenza. [ABSTRACT FROM AUTHOR]