Advancing Underlying Cause of Death Inference Through Wide and Deep Model.

Introduction: Accurately filling out death certificates is essential for death surveillance. However, manually determining the underlying cause of death is often imprecise. In this study, we investigate the Wide and Deep framework as a method to improve the accuracy and reliability of inferring the underlying cause of death. Methods: Death report data from national-level cause of death surveillance sites in Fujian Province from 2016 to 2022, involving 403,547 deaths, were analyzed. The Wide and Deep embedded with Convolutional Neural Networks (CNN) was developed. Model performance was assessed using weighted accuracy, weighted precision, weighted recall, and weighted area under the curve (AUC). A comparison was made with XGBoost, CNN, Gated Recurrent Unit (GRU), Transformer, and GRU with Attention. Results: The Wide and Deep achieved strong performance metrics on the test set: precision of 95.75%, recall of 92.08%, F1 Score of 93.78%, and an AUC of 95.99%. The model also displayed specific F1 Scores for different cause-of-death chain lengths: 97.13% for single causes, 95.08% for double causes, 91.24% for triple causes, and 79.50% for quadruple causes. Conclusions: The Wide and Deep significantly enhances the ability to determine the root causes of death, providing a valuable tool for improving cause-of-death surveillance quality. Integrating artificial intelligence (AI) in this field is anticipated to streamline death registration and reporting procedures, thereby boosting the precision of public health data.

Healthy lifestyles and rapid progression of carotid plaque in population with atherosclerosis: A prospective cohort study in China.

Background: Healthy lifestyles are effective means to reduce major cardiovascular events. However, little is known about the association of healthy lifestyles with development of carotid atherosclerosis at the early stage of cardiovascular diseases (CVDs). Methods: We enrolled participants from Fujian province in the China PEACE MPP project. We calculated a healthy lifestyle score by adherence to non-smoking, sufficient physical activity, healthy diet and healthy body mass index. Cox proportional hazards regression models and restricted cubic splines (RCS) were used to explore the association between the healthy lifestyles and rapid progression of carotid plaque. Results: 8379 participants were included (mean age: 60.6 ± 8.3 years, 54.6 % female), with a median follow-up of 1.2 years (inter quartile range: 1.0-1.6). RCS showed a significant inverse association between the healthy lifestyle score and progression of carotid plaque. Participants with "intermediate" (HR: 0.72 [95 % confidence interval (CI): 0.65-0.80]) or "ideal" (HR: 0.68 [0.59-0.78]) adherence to healthy lifestyles had a lower risk of progression of carotid plaque compared to those with "poor" adherence. Age, sex, occupation, income, residence type and metabolic status were significant factors influencing the relationship. Farmers benefited more in non-smoking and sufficient physical activity compared to non-farmers, and participants with lower income or without dyslipidaemia benefited more in sufficient physical activity and healthy diet compared to their counterparts (p-for-interaction < 0.05). Conclusions: Healthy lifestyles were associated with lower risk of progression of carotid plaque in populations with atherosclerosis. Promotion of healthy lifestyles from the early stage of carotid atherosclerosis could reduce the burden of CVDs in China.

Long-term exposure to PM2.5 chemical constituents and diabesity: evidence from a multi-center cohort study in China.

Background: Long-term exposure to PM2.5 is known to increase the risks for diabetes and obesity, but its effects on their coexistence, termed diabesity, remain uncertain. This study aimed to investigate the associations of long-term exposure to PM2.5 and its chemical constituents with the risks for diabesity, diabetes, and obesity. Methods: This cross-sectional study used the baseline data of a multi-center cohort, consisting of three provincially representative cohorts comprising a total of 134,403 participants from the eastern (Fujian Province), central (Hubei Province), and western (Yunnan Province) regions of China. Obesity and diabetes, and diabesity were identified by a body mass index (BMI) ≥28 kg/m2 and fasting plasma glucose (FPG) ≥126 mg/dL. The average concentrations of PM2.5 and five chemical constituents (NO3 -, SO4 2-, NH4 +, organic matter, and black carbon) over participants' residence during the past three years were estimated using machine learning models. Logistic regression models with double robust estimators, Bayesian kernel machine regression, and weighted quantile sum regression were employed to estimate independent and joint effects of PM2.5 chemical constituents on the risks for diabesity, diabetes, and obesity, as well as the differences from the effects on obesity. Stratified analyses were performed to examine effect modification of sociodemographic and lifestyle factors. Findings: There were 129,244 participants with a mean age of 54.1 ± 13.8 years included in the study. Each interquartile range increase in PM2.5 concentration (8.53 µg/m3) was associated with an increased risk for diabesity (OR = 1.23 [1.17, 1.30]), diabetes only (OR = 1.16 [1.13, 1.19]), and obesity only (OR = 1.03 [1.00, 1.05]). Long-term exposure to each PM2.5 chemical constituent was associated with an increased risk for diabesity, where organic matter exposure, with maximum weight (48%), was associated with a higher risk for diabesity (OR = 1.21 [1.16, 1.27]). Among those with obesity, black carbon contributed most (68%) to the joint effect of PM2.5 chemical constituents on diabesity (OR = 1.16 [1.11, 1.22]). Physical activity reduced adverse effects of PM2.5 on diabesity. Also, additive rather than multiplicative effects of obesity on the PM2.5-diabetes association were observed. Interpretation: Long-term exposure to PM2.5 and its chemical constituents was associated with an increased risk for diabesity, stronger than associations for diabetes and obesity alone. The main constituents associated with diabesity and obesity were black carbon and organic matter. Funding: National Natural Science Foundation of China (42271433, 723B2017), National Key R&D Program of China (2023YFC3604702), Fundamental Research Funds for the Central Universities (2042023kfyq04, 2042024kf1024), the Science and Technology Major Project of Tibetan Autonomous Region of China (XZ202201ZD0001G), Science and technology project of Tibet Autonomous Region（XZ202303ZY0007G), Key R&D Project of Sichuan Province (2023YFS0251), Renmin Hospital of Wuhan University (JCRCYG-2022-003), Jiangxi Provincial 03 Special Foundation and 5G Program (20224ABC03A05), Wuhan University Specific Fund for Major School-level Internationalization Initiatives (WHU-GJZDZX-PT07).