Association of the intergenerational structure with the onset of depressive symptoms among middle-aged and older adults in China.

BACKGROUND: Chinese family structure has undergone tremendous changes over the past few decades. Moreover, the association of the intergenerational structure with depression remains controversial. AIMS: This study aimed to find out the association of the intergenerational structure and the onset of depressive symptoms among Chinese middle-aged and older adults. METHODS: This study included 4,868 participants of the China Health and Retirement Longitudinal Study (CHARLS), who were enrolled in 2011 without depressive symptoms and followed up at least once later in 2013, 2015, 2018, and 2020. Taking the time-varying confounding effect into account, the time-dependent Cox regression models were used to estimate the association of the intergenerational structure and the onset of depressive symptoms. RESULTS: Among the studied middle-aged and older adults, compared to one-generation households, higher hazard ratios (HR) of developing depressive symptoms were found in three-generation households in the study population (HR = 1.21, 95% CI [1.08, 1.36]). Further, for female participants, skipping-generation households (HR = 1.38, 95% CI [1.05, 1.83]) and three-generation lineal households (HR = 1.21, 95% CI [1.02, 1.43]) were found to be significantly associated with new-onset depressive symptoms compared to empty-nest couples. For male participants, living alone (HR = 1.65, 95% CI [1.30, 2.11]), living in standardized nuclear households (HR = 1.27, 95% CI [1.06, 1.54]), impaired nuclear households (HR = 1.80, 95% CI [1.18, 2.76]), or three-generation lineal households (HR = 1.34, 95% CI [1.12, 1.60]) were found to have a significant association with the onset of depressive symptoms. CONCLUSIONS: This study found that males living alone, with unmarried children, or in three-generation lineal households, and females living with grandchildren were more likely to suffer from depressive symptoms. Therefore, special attention should be paid to people in these intergenerational structure subtypes.

Association between ambient PM1 and the prevalence of chronic kidney disease in China: A nationwide study.

Particulate of diameter ≤ 1 µm (PM1) presents a novel risk factor of adverse health effects. Nevertheless, the association of PM1 with the risk of chronic kidney disease (CKD) in the general population is not well understood, particularly in regions with high PM1 levels like China. Based on a nationwide representative survey involving 47,204 adults and multi-source ambient air pollution inversion data, the present study evaluated the association of PM1 with CKD prevalence in China. The two-year average PM1, particulate of diameter ≤ 2.5 µm (PM2.5), and PM1-2.5 values were accessed using a satellite-based random forest approach. CKD was defined as estimated glomerular filtration rate < 60 ml/min/1.73 m2 or albuminuria. The results suggested that a 10 µg/m3 rise in PM1 was related to a higher CKD risk (odds ratio [OR], 1.13; 95% confidence interval [CI] 1.08-1.18) and albuminuria (OR, 1.11; 95% CI, 1.05-1.17). The association between PM1 and CKD was more evident among urban populations, older adults, and those without comorbidities such as diabetes or hypertension. Every 1% increase in the PM1/PM2.5 ratio was related to the prevalence of CKD (OR, 1.03; 95% CI, 1.03-1.04), but no significant relationship was found for PM1-2.5. In conclusion, the present study demonstrated long-term exposure to PM1 was associated with an increased risk of CKD in the general population and PM1 might play a leading role in the observed relationship of PM2.5 with the risk of CKD. These findings provide crucial evidence for developing air pollution control strategies to reduce the burden of CKD.

Graph-guided deep hashing networks for similar patient retrieval.

With the rapid growth and widespread application of electronic health records (EHRs), similar patient retrieval has become an important task for downstream clinical decision support such as diagnostic reference, treatment planning, etc. However, the high dimensionality, large volume, and heterogeneity of EHRs pose challenges to the efficient and accurate retrieval of patients with similar medical conditions to the current case. Several previous studies have attempted to alleviate these issues by using hash coding techniques, improving retrieval efficiency but merely exploring underlying characteristics among instances to preserve retrieval accuracy. In this paper, drug categories of instances recorded in EHRs are regarded as the ground truth to determine the pairwise similarity, and we consider the abundant semantic information within such multi-labels and propose a novel framework named Graph-guided Deep Hashing Networks (GDHN). To capture correlation dependencies among the multi-labels, we first construct a label graph where each node represents a drug category, then a graph convolution network (GCN) is employed to derive the multi-label embedding of each instance. Thus, we can utilize the learned multi-label embeddings to guide the patient hashing process to obtain more informative and discriminative hash codes. Extensive experiments have been conducted on two datasets, including a real-world dataset concerning IgA nephropathy from Peking University First Hospital, and a publicly available dataset from MIMIC-III, compared with traditional hashing methods and state-of-the-art deep hashing methods using three evaluation metrics. The results demonstrate that GDHN outperforms the competitors at different hash code lengths, validating the superiority of our proposal.

Hospitalizations of Chronic Dialysis Patients: A National Study in China.

Background: Patients receiving chronic dialysis are usually with multiple comorbidities and at high risk for hospitalization, which lead to tremendous health care resource utilization. This study aims to explore the characteristics of hospitalizations among chronic dialysis patients in China. Methods: Hospital admissions from January 2013 to December 2015 were extracted from a national inpatient database in China. Chronic dialysis, including hemodialysis and peritoneal dialysis, was identified according to inpatient discharge records and International Classification of Diseases-10 (ICD-10) codes. The primary kidney disease, causes of admissions, modalities of dialysis, and comorbidities were analyzed. Multivariable logistic regression model was used to assess the association of patient characteristics with multiple hospitalizations per year. Results: Altogether, 266,636 hospitalizations from 124,721 chronic dialysis patients were included in the study. The mean age was 54.46 ± 15.63 years and 78.29% of them were receiving hemodialysis. The leading cause of hospitalizations was dialysis access-related, including dialysis access creation (25.06%) and complications of access (21.09%). The following causes were nonaccess surgery (1.89%), cardiovascular disease (1.66%), and infectious diseases (1.43%). One-fourth of the patients were hospitalized more than once per year. Multivariate logistic regression models indicated that the primary kidney disease of diabetic kidney disease (odds ratio [OR]: 1.16, 95% confidence interval [CI]: 1.11-1.22) or hypertensive nephropathy (OR: 1.33, 95% CI: 1.27-1.40), coronary heart disease (OR: 1.09, 95% CI: 1.05-1.14), cancer (OR: 1.21, 95% CI: 1.13-1.30), or modality of peritoneal dialysis (OR: 2.67, 95% CI: 2.59-2.75) was risk factors for multiple hospitalizations. Conclusion: Our study described characteristics and revealed the burden of hospitalizations of chronic dialysis patients in China. These findings highlight the importance of effective and efficient management strategies to reduce the high burden of hospitalization in dialysis population.

China trauma treatment statistics 2019: A national retrospective study based on hospitalized cases.

Objective: Trauma is China's fifth leading cause of death and ranked first among youths. Trauma databases have been well-established in many countries to announce the current state of trauma rescue, treatment and care. Nevertheless, China hasn't yet established a comparable database. This paper included two national-level databases in China to describe the current situation of trauma treatment and the epidemiological characteristics of trauma incidence, which sought to provide data support for decision-making, resource allocation, trauma prevention, trauma management, and other aspects. Methods: This study used the diagnosis and treatment data from the Hospital Quality Monitoring System (HQMS) and the China Trauma Rescue and Treatment Association (CTRTA) in 2019. A descriptive analysis was conducted to explore the demographic characteristics, trauma causes, injury degrees of trauma patients, disease burden and mortality rates in the abstracted hospitalized cases. Results: A total of 4,532,029 trauma patients were included, of which 4,436,653 were from HQMS and 95,376 from CTRTA respectively. The age group with the highest proportion is 50-54 years old (493,320 [11.12%] in HQMS and 12,025 [12.61%] in CTRTA). Fall was the most frequent cause of trauma hospitalization, accounting for 40.51% of all cases, followed by traffic injuries, accounting for 25.22%. However, for trauma patients aged between 20 and 24 years old, the most common cause of injury was traffic accidents (28.20%). Hospital expenses for trauma patients in 2019 exceeded 100.30 billion yuan, which increases significantly with age, and fall costs the most. The mortality rate of trauma inpatients was 0.77%, which gradually increased with age after 30-year-old, and was the highest in the age group above 85 (1.86%). Conclusion: This paper summarizes the demographic characteristics, trauma causes distribution, disease burden, mortality rate, and other relative data of inpatients in 2019, which can now be used as an up-to-date clinical evidence base for national healthcare prevention and management in China.

Prevalence of psychological symptoms and associated risk factors among nurses in 30 provinces during the COVID-19 pandemic in China.

Background: With the outbreak of the coronavirus disease 2019 (COVID-19), nurses have won well-deserved recognition for their indispensable roles in providing humane and professional healthcare for patients. However, by the nature of their role working at the forefront of patient care, nurses are prone to experiencing mental health consequences. Therefore, we pay attention to measuring the magnitude of psychological symptoms and identifying associated factors among nurses in China. Methods: We launched a nationwide, cross-sectional survey of nurses who worked in secondary or tertiary hospitals and public or private hospitals from 30 provinces in China. The prevalence and severity of symptoms of burnout, depression, and anxiety were investigated, respectively. Multivariable logistic regression analyses were performed to identify factors associated with each psychological symptom. Findings: A total of 138 279 respondents who worked in 243 hospitals completed this survey. A substantial proportion of nurses reported symptoms of burnout (34%), depression (55·5%), and anxiety (41·8%). In line with the disproportionality of economic development, we noted that the middle or western region was an independent risk factor for depression and anxiety. Compared with those working in the secondary hospital, nurses who worked in tertiary hospitals were associated with a higher likelihood of burnout and depression. Interpretation: Nurses are experiencing emotional, physical, and mental exhaustion during the COVID-19 epidemic. Governments and health policymakers need to draw attention to reinforcing prevention and ameliorating countermeasures to safeguard nurses' health. Funding: The strategic consulting project of the Chinese Academy of Engineering [2021-32-5]. Advanced Institute of Infomation Technology, Peking University, Zhejiang Province [2020-Z-17].

Structure-aware siamese graph neural networks for encounter-level patient similarity learning.

Patient similarity learning has attracted great research interest in biomedical informatics. Correctly identifying the similarity between a given patient and patient records in the database could contribute to clinical references for diagnosis and medication. The sparsity of underlying relationships between patients poses difficulties for similarity learning, which becomes more challenging when considering real-world Electronic Health Records (EHRs) with a large number of missing values. In the paper, we organize EHRs as a graph and propose a novel deep learning framework, Structure-aware Siamese Graph neural Networks (SSGNet), to perform robust encounter-level patient similarity learning while capturing the intrinsic graph structure and mitigating the influence from missing values. The proposed SSGNet regards each patient encounter as a node, and learns the node embeddings and the similarity between nodes simultaneously via Graph Neural Networks (GNNs) with siamese architecture. Further, SSGNet employs a low-rank and contrastive objective to optimize the structure of the patient graph and enhance model capacity. The extensive experiments were conducted on two publicly available datasets and a real-world dataset regarding IgA nephropathy from Peking University First Hospital, in comparison with multiple baseline and state-of-the-art methods. The significant improvement in Accuracy, Precision, Recall and F1 score on the patient encounter pairwise similarity classification task demonstrates the superiority of SSGNet. The mean average precision (mAP) of SSGNet on the similar encounter retrieval task is also better than other competitors. Furthermore, SSGNet's stable similarity classification accuracies at different missing rates of data validate the effectiveness and robustness of our proposal.

Bidirectional associations of vision loss, hearing loss, and dual sensory loss with depressive symptoms among the middle-aged and older adults in China.

BACKGROUND: Most studies on the relationship between sensory loss and depression focus on the unidirectional association between sensory loss and the risk of depression based on cross-sectional designs. The present study aimed to explore the bidirectional longitudinal associations of vison loss (VL), 1 hearing loss (HL), 2 and dual sensory loss (DSL)3 with depressive symptoms among Chinese population. METHODS: A longitudinal study was conducted among 13,690 participants aged 45 years and older over four years. VL, HL, and DSL were identified through self-reporting, and depressive symptoms were assessed using a 10-item Center for Epidemiologic Studies Depression Scale. Multivariable Cox proportional hazards regression models were constructed to estimate the bidirectional associations of VL, HL, and DSL with depressive symptoms. RESULTS: Participants with self-reported VL (HR: 1.14, 95%CI: 1.04-1.24), HL (HR: 1.22, 95%CI: 1.07-1.37), and DSL (HR: 1.27, 95%CI: 1.08-1.49) were associated with higher risk of developing depressive symptoms, compared with those without VL, HL, and DSL, respectively. In comparison with those without depressive symptoms, participants with depressive symptoms in the baseline had higher risk of developing VL (HR: 1.43, 95%CI: 1.33-1.54), HL (HR: 1.49, 95%CI: 1.36-1.63), and DSL (HR: 1.76, 95%CI: 1.59-1.95). LIMITATIONS: Sensory loss was defined only based on participants' self-report. CONCLUSIONS: Significant bidirectional associations exist between self-reported VL, HL, or DSL and depressive symptoms. The mental health of people with VL and HL should be focused on, and regular assessments of vision and hearing in people with depressive symptoms are recommended.

Association between pulse pressure, systolic blood pressure and the risk of rapid decline of kidney function among general population without hypertension: results from the China health and retirement longitudinal study (CHARLS).

BACKGROUND: Association between blood pressure (BP) and kidney function among the middle and old aged general population without hypertension remains unclear. METHODS: Participants aged ≥ 45 years, with complete data in 2011 and 2015 interviews of the China Health and Retirement Longitudinal Study(CHARLS), and without pre-existing hypertension were included. Systolic BP (SBP) was categorized as low (< 120 mmHg), medium (120-129 mmHg), and high (120-139 mmHg). Diastolic BP (DBP) was categorized as low (< 60 mmHg), medium (60-74 mmHg), and high (75-89 mmHg). Pulse pressure (PP) was categorized as normal (< 60 mmHg) and high (≥ 60 mmHg). The outcome was defined as rapid decline of estimated glomerular filtration rate(eGFR, decline ≥ 4 ml/min/1.73 m2/year). BP combination was designed according to the category of SBP and PP. The association between BP components, types of BP combination, and the risk of rapid decline of eGFR was analyzed using multivariate logistic regression models, respectively. Age-stratified analyses were conducted. RESULTS: Of 4,534 participants included, 695(15.3%) individuals were recognized as having rapid decline of eGFR. High PP[odds ratio(OR) = 1.34, 95%confidence interval(CI) 1.02-1.75], low SBP (OR = 1.28, 95%CI 1.03-1.59), and high SBP (OR = 1.32, 95% CI 1.02-1.71) were significantly associated with the risk of eGFR decline. Low SBP were associated with 65% increment of the risk of eGFR decline among participants aged < 55 years. The combination of high SBP and high PP (OR = 1.79, 95% CI 1.27-2.54) and the combination of low SBP and high PP (OR = 3.07, 95% CI 1.24-7.58) were associated with the increased risk of eGFR decline among the middle and old aged general population. CONCLUSION: Single and combination of high PP and high SBP could be the risk indicators of eGFR decline among the middle and old aged general population.

Predicting Prolonged Length of ICU Stay through Machine Learning.

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.

Influence of doctors' perception on the diagnostic status of chronic kidney disease: results from 976 409 individuals with electronic health records in China.

BACKGROUND: The diagnostic status of chronic kidney disease (CKD) and its underlying reasons provide evidence that can improve CKD management. However, the situation in developing countries remains under-investigated. METHODS: Adults with electronic health records (EHRs; 2008-19) in Yinzhou, China were included. The gold standard for CKD was defined as having persistently reduced estimated glomerular filtration rate (eGFR), albuminuria/proteinuria, haematuria or a history of CKD. CKD stages (G1-G5) were defined by eGFR. Clinical diagnosis of CKD in the real world setting was evaluated using International Classification of Diseases (ICD)-10 codes related to primary cause or stages of CKD. The specialty of doctors who administered the serum creatinine (SCr) tests and who made the primary-cause/CKD-staging diagnoses was analysed. The accuracy of CKD-staging codes was assessed. RESULTS: Altogether, 85 519 CKD patients were identified from 976 409 individuals with EHRs. Of them, 10 287 (12.0%) having persistent urinary abnormalities or labelled with CKD-related ICD codes did not receive SCr tests within 12 months before or after the urine tests. Among 75 147 patients who received SCr tests, 46 150 (61.4%) missed any CKD-related codes, 6857 (35.7%) were merely labelled with primary-cause codes, and only 2140 (2.9%) were labelled with CKD-staging codes. The majority of CKD patients (51.6-91.1%) received SCr tests from non-nephrologists, whereas CKD-staging diagnoses were mainly from nephrologists (52.3-64.8%). Only 3 of 42 general hospitals had nephrologists. The CKD-staging codes had high specificity (>99.0%) but low sensitivity (G3-G4: <10.0%). CONCLUSIONS: Under-perception of CKD among doctors, rather than unsatisfactory health-seeking behaviour or low detection rates, was the main cause of under-diagnosis of CKD in China. Intensification of CKD education among doctors with different specialties might bring about immediate effective improvement in the diagnosis and awareness of CKD.

Age- and Gender-Specific Prevalence of Frailty and Its Outcomes in the Longevous Population: The Chinese Longitudinal Healthy Longevity Study.

Background: Frailty is an epidemic age-related syndrome addressing heavy burden to the healthcare system. Subject to the rarity, age-, and gender-specific prevalence of frailty and its prognosis among the longevous population remains under-investigated. Methods: Based on the Chinese Longitudinal Healthy Longevity Study (CLHLS, 2008-2018), individuals aged ≥ 65 years having complete data of frailty were recruited. Modified Fried criteria (exhaustion, shrink, weakness, low mobility, and inactivity) were adopted to define pre-frailty (1-2 domains) and frailty (≥3 domains), respectively. The association between pre-frailty/frailty and adverse outcomes (frequent hospitalization, limited physical performance, cognitive decline, multimorbidity, and dependence) was analyzed using logistic regression models. The association between pre-frailty/frailty and mortality was analyzed using Cox proportional hazards models. Age- and gender-stratified analyses were performed. Results: Totally, 13,859 participants aged 85.8 ± 11.1 years, including 2,056 centenarians, were recruited. The overall prevalence of pre-frailty and frailty were 54.1 and 26.3%, respectively. Only 5.0% of centenarians were non-frailty whereas 59.9% of the young-old (65-79 years) showed pre-frailty. Both pre-frailty and frailty were associated with the increased risk of multiple adverse outcomes, such as incident limited physical performance, cognitive decline and dependence, respectively (P < 0.05). Frail males were more vulnerable to the risk of mortality (hazard ratio [HR] = 2.3, 95% confidence interval [CI], 2.1-2.6) compared with frail females (HR = 1.9, 95%CI, 1.7-2.1). The strongest association between frailty and mortality was observed among the young-old (HR = 3.6, 95%CI, 2.8-4.5). Exhaustion was the most common domain among patients with pre-frailty (74.8%) or frailty (83.2%), followed by shrink (32.3%) in pre-frailty and low mobility (83.0%) in frailty. Inactivity among females aged 65-79 years showed the strongest association with the risk of mortality (HR = 3.50, 95%CI, 2.52-4.87). Conclusion: A huge gap exists between longer life and healthy aging in China. According to the age- and gender-specific prevalence and prognosis of frailty, the strategy of frailty prevention and intervention should be further individualized.

Predicting Prolonged Length of Hospital Stay for Peritoneal Dialysis-Treated Patients Using Stacked Generalization: Model Development and Validation Study.

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.

Association of cardiovascular disease with 30-day hospital readmission in Chinese patients receiving maintenance dialysis.

BACKGROUND: Previous studies have shown cardiovascular disease (CVD) to be a risk factor in the prediction of 30-day hospital readmission among patients receiving dialysis. However, studies of Asian populations are limited. In the present study, we examined the association between CVD and 30-day hospital readmission in Chinese patients receiving maintenance dialysis. METHODS: Patients receiving maintenance dialysis were identified by searching a national claims database, the China Health Insurance Research Association (CHIRA) database, using the International Classification of Diseases revision 10 (ICD-10) and items of medical service claims. Patients aged ≥18 years who were discharged after index hospitalization between January 2015 and December 2015 were included in our retrospective analysis. CVD-related diagnoses were divided into three categories: coronary heart disease (CHD), heart failure (HF), and stroke. Thirty-day hospital readmission was defined as any hospital readmission within the 30 days following discharge. Logistic regression models adjusted for logit of propensity scores (PS) were used to assess the association of CVD with 30-day hospital readmission. RESULTS: Of 4,700 patients receiving dialysis, the 30-day hospital readmission rate was 10.4%. Compared with patients without CVD, there was an increased risk of 30-day hospital readmission among maintenance dialysis patients with total CVD [odds ratio (OR): 1.33, 95% confidence interval (CI): 1.06-1.66]. Patients with HF (OR: 1.77, CI: 1.27-2.47) and stroke (OR: 2.14, 95% CI: 1.53-2.98) had a greater risk of 30-day hospital readmission. The fully adjusted OR of CHD for the risk of 30-day hospital readmission was 1.22 (95% CI: 0.97-1.55). CONCLUSIONS: CVDs, especially stroke and HF, are independent predictors of 30-day hospital readmission in Chinese patients receiving dialysis, and could help to guide interventions to improve the quality of care for these patients.

Long-Term Exposure to Ambient PM2.5, Sunlight, and Obesity: A Nationwide Study in China.

Background: Accumulated researches revealed that both fine particulate matter (PM2.5) and sunlight exposure may be a risk factor for obesity, while researches regarding the potential effect modification by sunlight exposure on the relationship between PM2.5 and obesity are limited. We aim to investigate whether the effect of PM2.5 on obesity is affected by sunlight exposure among the general population in China. Methods: A sample of 47,204 adults in China was included. Obesity and abdominal obesity were assessed based on body mass index, waist circumference and waist-to-hip ratio, respectively. The five-year exposure to PM2.5 and sunlight were accessed using the multi-source satellite products and a geochemical transport model. The relationship between PM2.5, sunshine duration, and the obesity or abdominal obesity risk was evaluated using the general additive model. Results: The proportion of obesity and abdominal obesity was 12.6% and 26.8%, respectively. Levels of long-term PM2.5 ranged from 13.2 to 72.1 µg/m3 with the mean of 46.6 µg/m3. Each 10 µg/m3 rise in PM2.5 was related to a higher obesity risk [OR 1.12 (95% CI 1.09-1.14)] and abdominal obesity [OR 1.10 (95% CI 1.07-1.13)]. The association between PM2.5 and obesity varied according to sunshine duration, with the highest ORs of 1.56 (95% CI 1.28-1.91) for obesity and 1.66 (95% CI 1.34-2.07) for abdominal obesity in the bottom quartile of sunlight exposure (3.21-5.34 hours/day). Conclusion: Long-term PM2.5 effect on obesity risk among the general Chinese population are influenced by sunlight exposure. More attention might be paid to reduce the adverse impacts of exposure to air pollution under short sunshine duration conditions.

Development of a scoring tool for predicting prolonged length of hospital stay in peritoneal dialysis patients through data mining.

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.

Using machine learning methods to predict in-hospital mortality of sepsis patients in the ICU.

BACKGROUND: Early and accurate identification of sepsis patients with high risk of in-hospital death can help physicians in intensive care units (ICUs) make optimal clinical decisions. This study aimed to develop machine learning-based tools to predict the risk of hospital death of patients with sepsis in ICUs. METHODS: The source database used for model development and validation is the medical information mart for intensive care (MIMIC) III. We identified adult sepsis patients using the new sepsis definition Sepsis-3. A total of 86 predictor variables consisting of demographics, laboratory tests and comorbidities were used. We employed the least absolute shrinkage and selection operator (LASSO), random forest (RF), gradient boosting machine (GBM) and the traditional logistic regression (LR) method to develop prediction models. In addition, the prediction performance of the four developed models was evaluated and compared with that of an existent scoring tool - simplified acute physiology score (SAPS) II - using five different performance measures: the area under the receiver operating characteristic curve (AUROC), Brier score, sensitivity, specificity and calibration plot. RESULTS: The records of 16,688 sepsis patients in MIMIC III were used for model training and test. Amongst them, 2949 (17.7%) patients had in-hospital death. The average AUROCs of the LASSO, RF, GBM, LR and SAPS II models were 0.829, 0.829, 0.845, 0.833 and 0.77, respectively. The Brier scores of the LASSO, RF, GBM, LR and SAPS II models were 0.108, 0.109, 0.104, 0.107 and 0.146, respectively. The calibration plots showed that the GBM, LASSO and LR models had good calibration; the RF model underestimated high-risk patients; and SAPS II had the poorest calibration. CONCLUSION: The machine learning-based models developed in this study had good prediction performance. Amongst them, the GBM model showed the best performance in predicting the risk of in-hospital death. It has the potential to assist physicians in the ICU to perform appropriate clinical interventions for critically ill sepsis patients and thus may help improve the prognoses of sepsis patients in the ICU.

Anemia and Kidney Function Decline among the Middle-Aged and Elderly in China: A Population-Based National Longitudinal Study.

Chronic kidney disease (CKD) is a public health burden, and anemia is common among patients with CKD. However, less is known regarding the longitudinal association between anemia and deterioration of kidney function among the general population. The China Health and Retirement Longitudinal Study is a nationally representative survey for households with members aged ≥ 45 years. Participants without creatinine and demographic data in 2011 and 2015 were excluded. Anemia was defined according to definitions of the World Health Organization. Rapid decline in kidney function was defined as a ≥16.9% (quartile 3) decline in estimated glomerular filtration rate (eGFR), calculated using the CKD-EPI equation during 2011-2015. Multivariate logistic regression and restricted cubic splines were used to explore their relationship. Altogether, 7210 eligible participants were included in the analysis, with a mean age of 58.6 ± 8.8 years. Rapid decline in kidney function occurred among 1802 (25.0%) participants. Those with kidney function decline were more likely to be older, male, and have anemia, lower eGFRs, hypertension, and cardiovascular disease (P < 0.05). Anemia, or hemoglobin, was independently associated with rapid decline in kidney function after adjusting for potential confounding factors (OR = 1.64, 95% CI, 1.32-2.04; OR = 0.90, 95% CI, 0.87-0.94, respectively). Restricted cubic splines showed a nonlinear relationship between hemoglobin and rapid decline in kidney function, especially for men with anemia (P < 0.05). In conclusion, anemia is an independent risk factor for progression of kidney function among the middle-aged and elderly population. Attentive management and intervention strategies targeting anemia could be effective to reduce the risk of kidney failure and improve the prognosis of the general population.