Many high quality studies have emerged from public databases, such as Surveillance, Epidemiology, and End Results (SEER), National Health and Nutrition Examination Survey (NHANES), The Cancer Genome Atlas (TCGA), and Medical Information Mart for Intensive Care (MIMIC); however, these data are often characterized by a high degree of dimensional heterogeneity, timeliness, scarcity, irregularity, and other characteristics, resulting in the value of these data not being fully utilized. Data-mining technology has been a frontier field in medical research, as it demonstrates excellent performance in evaluating patient risks and assisting clinical decision-making in building disease-prediction models. Therefore, data mining has unique advantages in clinical big-data research, especially in large-scale medical public databases. This article introduced the main medical public database and described the steps, tasks, and models of data mining in simple language. Additionally, we described data-mining methods along with their practical applications. The goal of this work was to aid clinical researchers in gaining a clear and intuitive understanding of the application of data-mining technology on clinical big-data in order to promote the production of research results that are beneficial to doctors and patients.

Objective. The hemoglobin‐to‐red cell distribution width ratio (HRR) is associated with the prognosis of sepsis‐associated encephalopathy (SAE). This study aimed to determine the relationship between HRR and SAE and to clarify the possible mechanism of HRR as a prognostic factor for SAE. Methods. A multivariate Cox proportional‐hazards regression model was used to assess the correlation between HRR and all‐cause mortality. Piecewise linear regression and smooth‐curve Cox proportional‐hazards regression models were used to observe whether there was a nonlinear relationship between HRR and all‐cause mortality in SAE. Results. This study included 8853 patients with SAE. A nonlinear relationship between HRR and SAE was observed through a two‐segment regression model. The left inflection point for the HRR threshold was calculated to be 15.54, which was negatively correlated with all‐cause mortality (HR = 0.83, 95% CI = 0.76–0.91, p < 0.001). Subgroup analyses revealed significant interactions between white blood cell count, glucose, and patients who received dialysis and HRR. The inverse correlation between HRR and SAE was more pronounced in patients who did not receive vasopressin (HR = 0.91, 95% CI = 0.87–0.96, p < 0.001) than in those who did receive vasopressin (HR = 0.94, 95% CI = 0.88–1.02, p = 0.152) and was significantly more pronounced in patients without myocardial infarction (HR = 0.91, 95% CI = 0.88–0.96, p < 0.001) than in those with myocardial infarction (HR = 0.94, 95% CI = 0.87–1.02, p < 0.114). Conclusion. This large retrospective study found a nonlinear relationship between all‐cause mortality and HRR in patients with SAE in intensive care units, with low HRR being inversely associated with increased all‐cause mortality in patients with SAE.

Concurrent engineering has obtained increasing attention to solve the design problems of multidisciplinary systems. In practical engineering, there are epistemic uncertainties during whole design cycle of complex systems. Especially in earlier design phases, the effects of epistemic uncertainties are not usually easy to be quantified. It is because design information is insufficient. Furthermore, commonly used probability theory is also not suitable to be utilized. In this situation, epistemic uncertainties will be introduced unavoidably by mathematical models or simulation tools and may affect the performance of complex system significantly. To solve this problem, evidence theory is introduced and combined with the collaborative optimization method in this study. An evidence-based collaborative reliability optimization method is also proposed. Evidence theory is a powerful approach to handle epistemic uncertainties by Plausibility and Belief. Meanwhile, collaborative optimization is widely utilized in the concurrent design of complex systems. An aircraft conceptual design problem is utilized to show the application of the proposed method.

Background. The prognosis is poor when acute pancreatitis (AP) progresses to sepsis; therefore, it is necessary to accurately predict the probability of sepsis and develop a personalized treatment plan to reduce the disease burden of AP patients. Methods. A total of 1295 patients with AP and 43 variables were extracted from the Medical Information Mart for Intensive Care (MIMIC) IV database. The included patients were randomly assigned to the training set and to the validation set at a ratio of 7 : 3. The chi‐square test or Fisher’s exact test was used to test the distribution of categorical variables, and Student’s t‐test was used for continuous variables. Multivariate logistic regression was used to establish a prognostic model for predicting the occurrence of sepsis in AP patients. The indicators to verify the overall performance of the model included the area under the receiver operating characteristic curve (AUC), calibration curves, the net reclassification improvement (NRI), the integrated discrimination improvement (IDI), and a decision curve analysis (DCA). Results. The multifactor analysis results showed that temperature, phosphate, calcium, lactate, the mean blood pressure (MBP), urinary output, Glasgow Coma Scale (GCS), Charlson Comorbidity Index (CCI), sodium, platelet count, and albumin were independent risk factors. All of the indicators proved that the prediction performance and clinical profitability of the newly established nomogram were better than those of other common indicators (including SIRS, BISAP, SOFA, and qSOFA). Conclusions. The new risk‐prediction system that was established in this research can accurately predict the probability of sepsis in patients with acute pancreatitis, and this helps clinicians formulate personalized treatment plans for patients. The new model can reduce the disease burden of patients and can contribute to the reasonable allocation of medical resources, which is significant for tertiary prevention.

Background: Falls are a major public health problem in the older adults that can lead to poor clinical outcomes. There have been few reports on the short-term prognoses of older critically ill patients, and so we sought to determine the impact of falls on elderly patients in intensive care units (ICUs). Patients and Methods: This retrospective study of 4503 patients (aged 65 years or older) analyzed data in the Medical Information Mart for Intensive Care-III critical care database. Of those, 2459 (54.6%) older adults are males, and 2044 (45.4%) older adults are females. Based on information from the medical care record assessment forms, patients were classified into the following two groups based on whether they had a fall within the previous 3 months: falls (n=1142) and nonfalls (n=3361). The primary outcomes of this study were 30- and 90-day mortality. Associations between the results of the Kaplan-Meier (KM) survival analysis, Cox proportional-hazards regression models, and subgroup analysis and its outcomes were analyzed using stabilized inverse probability treatment weighting (IPTW). Results: KM survival curves with stabilized IPTW indicated that 30- and 90-day survival rates were significantly lower in elderly critically ill patients with a history of falls within the previous 3 months than in those patients without a history of falls (all p< 0.001). Multivariate Cox proportional-hazards regression analysis indicated that 30- and 90-day mortality rates were 1.35 times higher (95% confidence interval [CI]=1.16– 1.57, p< 0.001) and 1.36 times higher (95% CI=1.19– 1.55, p< 0.001), respectively, in elderly critically ill patients with a history of falls within the previous 3 months than in those patients without a history of falls. Conclusion: Experience of falls within the previous 3 months prior to ICU admission increased the risk of short-term mortality and affected the prognoses of elderly patients. Falls should therefore receive adequate attention from clinical healthcare providers and management decision-makers.

Background. A survival prediction model based on deep learning has higher accuracy than the CPH model in predicting the survival of CCU patients, and it also has a better discrimination ability. We collected information on patients with various diseases in coronary care units (CCUs) from the Medical Information Mart for Intensive Care III (MIMIC-III) database. The purpose of this study was to use this information to construct a neural-network model based on deep learning to predict the survival probabilities of patients with conditions that are common in CCUs. Method. We collected information on patients in the United States with five common diseases in CCUs from 2001 to 2012. We randomly divided the patients into a training cohort and a testing cohort at a ratio of 7 : 3 and applied a survival prediction method based on deep learning to predict their survival probability. We compared our model with the Cox proportional-hazards regression (CPH) model and used the concordance indexes (C-indexes), receiver operating characteristic (ROC) curve, and calibration plots to evaluate the predictive performance of the model. Results. The 3,388 CCU patients included in the study were randomly divided into 2,371 in the training cohort and 1,017 in the testing cohort. The stepwise regression results showed that the important factors affecting patient survival were the type of disease, age, race, anion gap, glucose, neutrophils, white blood cells, potassium, creatine kinase, and blood urea nitrogen ( ). We used the training cohort to construct a deep-learning model, for which the C-index was 0.833, or about 5% higher than that for the CPH model (0.786). The C-index of the deep-learning model for the test cohort was 0.822, which was also higher than that for the CPH model (0.782). The areas under the ROC curve for the 28-day, 90-day, and 1-year survival probabilities were 0.875, 0.865, and 0.874, respectively, in the deep-learning model, respectively, and 0.830, 0.843, and 0.806 in the CPH model. These values indicate that the survival analysis model based on deep learning is better than the traditional CPH model in predicting the survival of CCU patients. Conclusion. A survival prediction model based on deep learning has higher accuracy than the CPH model in predicting the survival of CCU patients, and it also has a better discrimination ability.