This study aimed to construct machine learning (ML) models for predicting prolonged length of stay (pLOS) in intensive care units (ICU) among general ICU patients. A multicenter database called eICU (Collaborative Research Database) was used for model derivation and internal validation, and the Medical Information Mart for Intensive Care (MIMIC) III database was used for external validation. We used four different ML methods (random forest, support vector machine, deep learning, and gradient boosting decision tree (GBDT)) to develop prediction models. The prediction performance of the four models were compared with the customized simplified acute physiology score (SAPS) II. The area under the receiver operation characteristic curve (AUROC), area under the precision-recall curve (AUPRC), estimated calibration index (ECI), and Brier score were used to measure performance. In internal validation, the GBDT model achieved the best overall performance (Brier score, 0.164), discrimination (AUROC, 0.742; AUPRC, 0.537), and calibration (ECI, 8.224). In external validation, the GBDT model also achieved the best overall performance (Brier score, 0.166), discrimination (AUROC, 0.747; AUPRC, 0.536), and calibration (ECI, 8.294). External validation showed that the calibration curve of the GBDT model was an optimal fit, and four ML models outperformed the customized SAPS II model. The GBDT-based pLOS-ICU prediction model had the best prediction performance among the five models on both internal and external datasets. Furthermore, it has the potential to assist ICU physicians to identify patients with pLOS-ICU risk and provide appropriate clinical interventions to improve patient outcomes.

Background: Patients receiving maintenance hemodialysis (HD) and peritoneal dialysis (PD) are frequently hospitalized. Reducing unplanned 30-day hospital readmissions is a key priority for improving the quality of health care. The purpose of this study was to assess the association between the Charlson Comorbidity Index (CCI), which has been used to evaluate multi-comorbidities status, and 30-day readmission in patients on HD and PD therapy. Methods: The Hospital Quality Monitoring System (HQMS), a national administrative database for hospitalized patients in China was used to extract dialysis patients admitted from January 2013 to December 2015. The outcome was the unplanned readmission following the hospital discharge within 30 days. For patients with multiple hospitalizations, a single hospitalization was randomly selected as the index hospitalization. A cause-specific Cox proportional hazard model was utilized to assess the association of CCI with readmission within 30 days. Results: Of the 124,721 patients included in the study, 19,893 patients (16.0%) were identified as experiencing unplanned readmissions within 30 days. Compared with patients without comorbidity (CCI = 2, scored for dialysis), the risk of 30-day readmission increased with elevated CCI score. The hazards ratio (HR) for those with CCI 3–4, 5–6 and > 6 was 1.01 (95% confidence interval [CI] 0.98–1.05), 1.09 (95% CI 1.05–1.14), and 1.14 (95% CI 1.09–1.20), respectively. Conclusions: Our study indicated that CCI was independently associated with the risk of 30-day readmission for patients receiving dialysis including HD and PD, and could be used for risk-stratification.

Background: The increasing number of patients treated with peritoneal dialysis (PD) and their consistently high rate of hospital admissions have placed a large burden on the health care system. Early clinical interventions and optimal management of patients at a high risk of prolonged length of stay (pLOS) may help improve the medical efficiency and prognosis of PD-treated patients. If timely clinical interventions are not provided, patients at a high risk of pLOS may face a poor prognosis and high medical expenses, which will also be a burden on hospitals. Therefore, physicians need an effective pLOS prediction model for PD-treated patients. Objective: This study aimed to develop an optimal data-driven model for predicting the pLOS risk of PD-treated patients using basic admission data. Methods: Patient data collected using the Hospital Quality Monitoring System (HQMS) in China were used to develop pLOS prediction models. A stacking model was constructed with support vector machine, random forest (RF), and K-nearest neighbor algorithms as its base models and traditional logistic regression (LR) as its meta-model. The meta-model used the outputs of all 3 base models as input and generated the output of the stacking model. Another LR-based pLOS prediction model was built as the benchmark model. The prediction performance of the stacking model was compared with that of its base models and the benchmark model. Five-fold cross-validation was employed to develop and validate the models. Performance measures included the Brier score, area under the receiver operating characteristic curve (AUROC), estimated calibration index (ECI), accuracy, sensitivity, specificity, and geometric mean (Gm). In addition, a calibration plot was employed to visually demonstrate the calibration power of each model. Results: The final cohort extracted from the HQMS database consisted of 23,992 eligible PD-treated patients, among whom 30.3% had a pLOS (ie, longer than the average LOS, which was 16 days in our study). Among the models, the stacking model achieved the best calibration (ECI 8.691), balanced accuracy (Gm 0.690), accuracy (0.695), and specificity (0.701). Meanwhile, the stacking and RF models had the best overall performance (Brier score 0.174 for both) and discrimination (AUROC 0.757 for the stacking model and 0.756 for the RF model). Compared with the benchmark LR model, the stacking model was superior in all performance measures except sensitivity, but there was no significant difference in sensitivity between the 2 models. The 2-sided t tests revealed significant performance differences between the stacking and LR models in overall performance, discrimination, calibration, balanced accuracy, and accuracy. Conclusions: This study is the first to develop data-driven pLOS prediction models for PD-treated patients using basic admission data from a national database. The results indicate the feasibility of utilizing a stacking-based pLOS prediction model for PD-treated patients. The pLOS prediction tools developed in this study have the potential to assist clinicians in identifying patients at a high risk of pLOS and to allocate resources optimally for PD-treated patients.

Background: The hospital admission rate is high in patients treated with peritoneal dialysis (PD), and the length of stay (LOS) in the hospital is a key indicator of medical resource allocation. This study aimed to develop a scoring tool for predicting prolonged LOS (pLOS) in PD patients by combining machine learning and traditional logistic regression (LR). Methods: This study was based on patient data collected using the Hospital Quality Monitoring System (HQMS) in China. Three machine learning methods, classification and regression tree (CART), random forest (RF), and gradient boosting decision tree (GBDT), were used to develop models to predict pLOS, which is longer than the average LOS, in PD patients. The model with the best prediction performance was used to identify predictive factors contributing to the outcome. A multivariate LR model based on the identified predictors was then built to derive the score assigned to each predictor. Finally, a scoring tool was developed, and it was tested by stratifying PD patients into different pLOS risk groups. Results: A total of 22,859 PD patients were included in our study, with 25.2% having pLOS. Among the three machine learning models, the RF model achieved the best prediction performance and thus was used to identify the 10 most predictive variables for building the scoring system. The multivariate LR model based on the identified predictors showed good discrimination power with an AUROC of 0.721 in the test dataset, and its coefficients were used as a basis for scoring tool development. On the basis of the developed scoring tool, PD patients were divided into three groups: low risk (≤5), median risk [5–10], and high risk (>10). The observed pLOS proportions in the low-risk, median-risk, and high-risk groups in the test dataset were 11.4%, 29.5%, and 54.7%, respectively. Conclusions: This study developed a scoring tool to predict pLOS in PD patients. The scoring tool can effectively discriminate patients with different pLOS risks and be easily implemented in clinical practice. The pLOS scoring tool has a great potential to help physicians allocate medical resources optimally and achieve improved clinical outcomes.

Active mobility, encompassing walking and cycling for transportation, is a potential solution to health issues arising from inadequate physical activity in modern society. However, the extent of active mobility's impact on individual physical activity levels, and its association with health as mediated by physical activities, is not fully quantified. This study aims to clarify the direct relationship between active mobility usage and individual health, as well as the indirect relationship mediated by physical activity, with a focus on varying levels of physical activity intensity. Utilizing data from the 2017 U.S. National Household Travel Survey (NHTS), we employed Poisson regression to predict active mobility usage based on socio-demographic and household socio-economic characteristics. A Structural Equation Model (SEM) was then used to investigate the direct and indirect effects of active mobility on individual health, mediated by physical activity. We further segmented individuals according to their intensity of physical activity to examine how such effect differs between different levels of physical activity. The study demonstrates that active mobility usage positively correlates with both the amount and intensity of physical activity. The effect of active mobility on individual health includes a direct positive effect (29% for intensity, 67.7% for amount) and an indirect effect mediated by physical activity (71% for intensity, 32.3% for amount). Notably, the mediation effect of active mobility on health is more substantial in the context of vigorous physical activities compared to light or moderate activities. Our findings reveal a significant positive influence of active mobility on individual health, encompassing both direct and indirect effects mediated by physical activities. These results quantitatively underscore the health benefits of active mobility and suggest the importance of promoting active mobility as a strategy to improve public health.

Fine particulate matter (PM_2.5) from different source sectors might differ in toxicity. However, data from large-scale studies on vulnerable children in low- and middle-income countries (LMICs) are insufficient. To analyze the association of under-five death (U5D) with long-term exposure to PM_2.5 from different sources. We evaluated demographic and health survey data for 79,995 babies born in 2017 in 16 Asian and African LMICs (AA-LMICs) and a Latin America low-income country (i.e., Haiti). Long-term exposure to PM_2.5 was assessed by a well-established product that attributed the annual concentration to 20 source sectors in 2017. The associations of survival during < 5-year periods with each source-specific concentration of PM_2.5 were analyzed by Cox regression with multiple adjustments. We derived a multiple-pollutant ridge regression model to estimate the joint exposure-response function (JERF) between U5D and PM_2.5 mixtures. To evaluate how sources affected PM_2.5 toxicity, we evaluated the number of U5Ds attributable to PM_2.5 based on the source profiles for 88 AA-LMICs. According to the single-pollutant model, the risk of U5D increased by 7% (95% confidence interval [CI]: 5%, 9%) for each 10 μg/m³ increment in total PM_2.5 concentration. The model performance was lower than that of the multiple-pollutant ridge regression model. For each 10 μg/m³ increment in PM_2.5, the excess risk of U5D ranged from 6% (95% CI: 4%, 9%) in Nepal to 10% (95% CI: 6%, 14%) in Mauritania. Based on the JERF, PM_2.5 contributed to 817,647 (95% CI: 585,729, 1,050,439), i.e., 28.0% (95% CI: 20.1%, 35.8%), of all U5Ds across the 88 AA-LMICs. The PM_2.5-related U5Ds were mostly attributable to PM_2.5 produced by desert dust, followed by solid biofuel combustion and open fires. The average toxicity of PM_2.5 varied by source profile, which should be taken into consideration when planning public health interventions. For some AA LMICs, natural sources of PM_2.5 had the most significant health effects, and should not be ignored to ensure the protection of child health.