Objective To explore the prevalence of Helicobacter pylori (Hp) infection and its association with dietary and lifestyle habits among the adult physical examination population in Xuzhou area. Methods Retrospectively selected the physical examination population who underwent HP testing at our hospital's physical examination center from May 2021 to December 2023 as the research object. The prevalence of Hp infection in the population was analyzed based on the physical examination results. A questionnaire survey was used to collect information on the eating and living habits of all study subjects. Logistic regression was used to analyze the relationship between eating and living habits and Hp infection. Results A total of 1 354 physical examination people were included in the study, and the Hp infection rate was 37.30% (505/1354). The difference in Hp infection rates among people of different age groups is statistically significant (P<0.05), with the middle-aged population (41-59 years old) having the highest Hp positive infection rate (45.38%).High salt (41.11%), hot diet (40.56%), history of smoking (45.23%) and drinking (43.80%), less consumption of fruits and vegetables (43.73%), irregular exercise (41.29%), irregular diet People who frequently eat out (43.56%) and eat out frequently (42.57%) have a higher Hp infection rate (P<0.05).After adjusting for demographic factors such as gender, age, place of residence and education level, multivariate Logistic regression results showed that high-salt diet (OR=3.975, 95%CI: 2.670-5.917) and hot diet (OR=3.357, 95%CI: 2.291-4.919), smoking (OR=1.458, 95%CI: 1.082-1.964), drinking alcohol (OR=1.654, 95%CI: 1.279-2.138), eating fruits and vegetables (OR=1.759, 95%CI: 1.345-2.301), regular exercise (OR=1.822, 95%CI: 1.371-2.421), regular diet (OR=1.893, 95%CI: 1.391-2.575), eating out (OR=1.690, 95%CI: 1.277-2.237) were associated with the risk of Hp infection (P<0.05). Conclusion The positive infection rate of Hp among the physical examination population in Xuzhou is slightly lower than the average epidemic level in China. Cultivating healthy eating and living habits can effectively reduce the risk of Hp infection.

OBJECTIVE To investigate the actual construction and operation status of established and under-construction centralized （cloud） prescription review centers （shortened for “prescription review center”） of close-knit county-level medical consortium in a certain city， so as to provide reference for improving the construction quality of the prescription review center. METHODS An online questionnaire survey was conducted to collect the data from 51 established and under-construction prescription review center in the city， covering basic information， funding sources， talent management， system construction， review rule maintenance， prescription review practices， prescription evaluation， data utilization， and current challenges. The collected data were summarized and analyzed. RESULTS A total of 51 valid questionnaires were retrieved， covering 32 established and 19 under-construction prescription review center. Among the 32 established prescription review centers， the main funding sources for their construction came from government financial allocations， accounting for 56.25%. Only 25.00% of prescription review center had review pharmacists who fully met national qualification requirements， and just 55.00% updated more than 10 review rule entries per month on average. Outpatient prescription verification realized full coverage， but 37.50% of prescription review centers only supported rationality verification of single prescriptions， and 50.00% could not retrieve laboratory and examination results to assist in prescription review. Additionally， 40.62% of prescription review center had not regularly conducted prescription evaluations for primary care institutions. The data from prescription review center was mainly used to support medication monitoring. Among the 19 prescription review centers currently in the planning stage， 63.16% had no identified funding sources. CONCLUSIONS The operation and construction of prescription review center in the city face challenges， such as funding shortages， absence of collaborative incentive mechanisms， and insufficient manpower.It is suggested that the state should issue a unified standard for the construction of the prescription review center as soon as possible， and local health administrative departments should formulate supporting policies and clarify assessment indicators in combination with the actual situation of the region.

Environmental pollution is a significant public health challenge worldwide, and investigating the causal relationship between environmental exposure and population health outcomes is a key objective of environmental epidemiology research. In recent years, the complexity of environmental exposures has increasingly come to the forefront, making it challenging for observational studies that dominate environmental epidemiology to accurately estimate causal effects. Causal inference methods are particularly advantageous in controlling for confounding factors, thus holding great potential in environmental epidemiology research. Researchers can use appropriate causal inference methods to simulate the process of randomization, providing strong support for revealing the causal relationship between environmental exposure and health outcomes. However, there is a lack of reviews on the application of causal inference methods in environmental epidemiology studies in China. Therefore, this study introduced the basic principles of common causal inference statistical methods in environmental epidemiology, summarized the applicable conditions, advantages and disadvantages of various methods, and provided R software implementation codes for these methods, aiming to offer guidance for optimizing research design and practicing causal inference statistical methods.

Objective: To develop a depression recognition model by integrating the spirit-expression diagnostic framework of traditional Chinese medicine (TCM) with machine learning algorithms. The proposed model seeks to establish a TCM-informed tool for early depression screening, thereby bridging traditional diagnostic principles with modern computational approaches. Methods: The study included patients with depression who visited the Shanghai Pudong New Area Mental Health Center from October 1, 2022 to October 1, 2023, as well as students and teachers from Shanghai University of Traditional Chinese Medicine during the same period as the healthy control group. Videos of 3 – 10 s were captured using a Xiaomi Pad 5, and the TCM spirit and expressions were determined by TCM experts (at least 3 out of 5 experts agreed to determine the category of TCM spirit and expressions). Basic information, facial images, and interview information were collected through a portable TCM intelligent analysis and diagnosis device, and facial diagnosis features were extracted using the Open CV computer vision library technology. Statistical analysis methods such as parametric and non-parametric tests were used to analyze the baseline data, TCM spirit and expression features, and facial diagnosis feature parameters of the two groups, to compare the differences in TCM spirit and expression and facial features. Five machine learning algorithms, including extreme gradient boosting (XGBoost), decision tree (DT), Bernoulli naive Bayes (BernoulliNB), support vector machine (SVM), and k-nearest neighbor (KNN) classification, were used to construct a depression recognition model based on the fusion of TCM spirit and expression features. The performance of the model was evaluated using metrics such as accuracy, precision, and the area under the receiver operating characteristic (ROC) curve (AUC). The model results were explained using the Shapley Additive exPlanations (SHAP). Results: A total of 93 depression patients and 87 healthy individuals were ultimately included in this study. There was no statistically significant difference in the baseline characteristics between the two groups (P > 0.05). The differences in the characteristics of the spirit and expressions in TCM and facial features between the two groups were shown as follows. (i) Quantispirit facial analysis revealed that depression patients exhibited significantly reduced facial spirit and luminance compared with healthy controls (P < 0.05), with characteristic features such as sad expressions, facial erythema, and changes in the lip color ranging from erythematous to cyanotic. (ii) Depressed patients exhibited significantly lower values in facial complexion L, lip L, and a values, and gloss index, but higher values in facial complexion a and b, lip b, low gloss index, and matte index (all P < 0.05). (iii) The results of multiple models show that the XGBoost-based depression recognition model, integrating the TCM “spirit-expression” diagnostic framework, achieved an accuracy of 98.61% and significantly outperformed four benchmark algorithms—DT, BernoulliNB, SVM, and KNN (P < 0.01). (iv) The SHAP visualization results show that in the recognition model constructed by the XGBoost algorithm, the complexion b value, categories of facial spirit, high gloss index, low gloss index, categories of facial expression and texture features have significant contribution to the model. Conclusion This study demonstrates that integrating TCM spirit-expression diagnostic features with machine learning enables the construction of a high-precision depression detection model, offering a novel paradigm for objective depression diagnosis.

Objective To evaluate the short-term effects of comprehensive health management interventions for stroke high-risk population screening on the prevention and treatment of ischemic stroke, and to provide reference and basis for improving and exploring health management and prevention strategies for stroke high-risk population. Methods From 2018 to 2022, 13 community health service centers in Jiading District, Shanghai were selected in the present study. Based on information push platform, stroke risk assessment and health intervention follow-up were conducted for community residents through convenience sampling. The residents were divided into a full course intervention group (intervention group) and a routine intervention group (control group) according to different health intervention measures and forms. The incidence of ischemic stroke in the two groups of survey subjects was tracked within 36 months. Results A total of 52144 subjects were included in the study. The total number of patients in the full course intervention group was 14227, with an incidence density of 577.32/100 000 (556.49/100 000-598.12/100 000), which was lower than that of the conventional intervention group (37 917), with an incidence density of 1 485.47/100 000 (1 464.99/100 000-1 505.94/100 000) (χ²=2490.212, P<0.001). The relative risk of the full course intervention group was 0.39, and the relative risk of stroke risk factors in the full course intervention group from low to high was 0.33, 0.43, 0.45, and 0.49, respectively. The incidence density of males in the full course intervention group was 660.76 (627.46/100 000 - 694.05/100 000), with a relative risk of 0.43, and the incidence density of female patients was 509.71/100 000 (483.37/100 000 - 536.05/100 000), with a relative risk of 0.35. The overall incidence density of the population under 62 years old gourp, 62-75 years old group and over 75 years old group was 197.45/100 000 (173.09/100 000 -221.80/100 000), 608.36/100 000 (580.19/100 000-636.54/100 000), and 1 025.06/100 000 (958.51/100 000-1 091.61/100 000), with relative risks of 0.51, 0.44, and 0.38, respectively. Conclusion Comprehensive health management measures can effectively reduce the short-term risk of ischemic stroke, and should be further promoted and improved to enhance the effectiveness of stroke prevention and control.

Gastric ulcer （GU） is a high-incidence digestive system disease characterized pathologically by disruption of gastric mucosal integrity， with clinical features including a prolonged course and periodic recurrence. Modern medicine attributes its pathogenesis to the dynamic imbalance between aggressive and defensive factors，while traditional Chinese medicine （TCM） posits its development as closely linked to spleen deficiency. Current therapies combining acid suppressants and antibiotics face challenges such as high recurrence rates，poor mucosal healing，and adverse drug reactions. Long-term use may induce metabolic disturbances like hypergastrinemia and reduced intestinal microbiota diversity. Therefore，exploring safer and longer-lasting therapeutic strategies has become a critical focus. TCM has extensive clinical experience and unique advantages in GU prevention and treatment. Studies demonstrate that the classic formula Shenling Baizhu San exhibits therapeutic properties of "invigorating spleen and tonifying Qi to restore physiological balance and eliminating dampness and regulating middle energizer to unblock Qi movement"， enabling a holistic approach targeting both symptoms and root causes in GU with spleen deficiency as the core pathology by suppressing aggressive factors and strengthening defensive factors. Experimental research reveals its mechanisms involve enhancing the physicochemical barrier of the mucus layer，repairing epithelial barriers and microcirculation，modulating gastric acid secretion and gastrointestinal motility，and regulating microecological barriers and mucosal immunity. Clinical evidence confirms its synergistic effects in promoting ulcer healing，improving Helicobacter pylori eradication rates，and reducing recurrence risks. This review examined the etiology and pathogenesis of GU and systematically evaluated Shenling Baizhu San from three perspectives-clinical application，pharmacological effects， and experimental research-to provide insights for optimizing integrated traditional Chinese and Western medicine protocols and expanding its clinical applications.

Background Lead (Pb) is a highly toxic heavy metal that accumulates in the body, potentially leading to multi-systemic impairment. Compared with traditional statistical methods, machine learning techniques offer unique advantages, opening new avenues for occupational health risk assessment and the exploratory analysis of complex associations. Objective To examine the association between occupational lead exposure and multiple health indicators and to identify key risk factors for lead toxicity. Methods A cross-sectional study was conducted, integrating occupational hygiene investigation results from 16 lead-acid battery enterprises in Jiangsu Province with occupational health examination data from 1914 lead-exposed workers. Inter-group differences were analyzed using the χ² test or Fisher's exact test. Binary logistic regression and machine learning algorithms [CatBoost, Naive Bayes model (NBM), and random forest (RF)] were employed to evaluate the association between blood lead (PbB), urine lead (PbU), and health indicators including blood pressure (BP), red blood cell count (RBC), and alanine aminotransferase (ALT). Results The prevalence of abnormal PbB and PbU were 14.52% and 9.35%, respectively. The risks of abnormal BP, RBC, and ALT were significantly increased in the population with high lead levels (P<0.05). PbB abnormalities were closely associated with gender, environmental lead concentration, wearing masks, smoking, and alcohol consumption (P<0.05). Regarding occupational hazards, workers exposed to lead dust had a 1.98-fold risk of PbU abnormality compared to those exposed to lead fumes. The plate coating and acid leaching process posed the highest risk for both PbB (OR=8.81) and PbU (OR=5.46) abnormalities compared with assembly process. Furthermore, the risks of PbB and PbU abnormalities were significantly elevated among workers with abnormal BP, RBC or ALT (P<0.05). Among the models, CatBoost performed best in predicting RBC abnormality (accuracy: 95.8%; precision: 44.9%; F1 score: 0.952; AUC: 0.981). Feature importance analysis identified PbB and PbU as the core factors affecting abnormal RBC and ALT, while RBC and ALT abnormalities as key features for predicting the risk of PbB and PbU abnormalities. Conclusion By integrating traditional statistical methods with machine learning, this study reveals a complex bidirectional association between occupational lead exposure and multiple health indicators, and identifies gender, job category, and environmental Pb concentration as the key factors influencing PbB abnormalities. These findings provide a scientific foundation for the implementation of precision occupational health management models.

Background Ozone (O₃) is a major air pollutant. The existing monitoring system has uneven distribution of sites, insufficient coverage in underdeveloped areas, and low temporal resolution, making it difficult to obtain hourly data. This limits the dynamic identification of pollution and the formulation of prevention and control strategies. Objective To construct an hourly O₃ concentration estimation model based on ensemble machine learning, aiming to improve the accuracy of pollution exposure assessment and explore O₃ health impacts. Methods This study integrated land use regression modeling with modern machine learning techniques, employing random forest and XGBoost algorithms to construct base models, and stacking integration using non-negative least squares. The ensemble model was trained and validated across China using high-resolution, multi-source geographic data (e.g., meteorologicaldata, population density, land cover types, and aerosol optical thickness). It was tested in Taiyuan City, combined with a distributed lag non-linear model to analyze the association between O₃ and emergency admissions. Results The constructed ensemble model performed well in predicting O₃ concentration, with a higher coefficient of determination (R²) and a lower root-mean-square deviation (RMSE) compared to the single models. The R² improved from 0.90 to 0.92, and the RMSE decreased from 11.41 to 10.62, enhancing both prediction accuracy and generalization ability. In the application to Taiyuan City, the model successfully imputed the hourly-level data for the entire year. The distributed lag non-linear model analysis revealed that the relative risk (RR) values for the 6th to 8th days following O₃ exposure were 1.14 (95%CI: 1.01, 1.29), 1.16 (95%CI: 1.02, 1.31), and 1.14 (95%CI: 1.01, 1.29), respectively, which were significantly higher than 1, indicating a significant lagged association (lagged 6-8 d) between O₃ and the number of emergency room visits. Conclusion A high-precision, hourly-level O₃ concentration estimation model is successfully constructed by combining the land use regression model with an ensemble machine learning approach to provide a scientific basis for environmental policy formulation and public health intervention. The application of the model verifies its generalization ability and practical application value, which can provide a new technical framework for subsequent environmental health research.

Background Lead (Pb) is a highly toxic heavy metal that accumulates in the body, potentially leading to multi-systemic impairment. Compared with traditional statistical methods, machine learning techniques offer unique advantages, opening new avenues for occupational health risk assessment and the exploratory analysis of complex associations. Objective To examine the association between occupational lead exposure and multiple health indicators and to identify key risk factors for lead toxicity. Methods A cross-sectional study was conducted, integrating occupational hygiene investigation results from 16 lead-acid battery enterprises in Jiangsu Province with occupational health examination data from 1914 lead-exposed workers. Inter-group differences were analyzed using the χ² test or Fisher's exact test. Binary logistic regression and machine learning algorithms [CatBoost, Naive Bayes model (NBM), and random forest (RF)] were employed to evaluate the association between blood lead (PbB), urine lead (PbU), and health indicators including blood pressure (BP), red blood cell count (RBC), and alanine aminotransferase (ALT). Results The prevalence of abnormal PbB and PbU were 14.52% and 9.35%, respectively. The risks of abnormal BP, RBC, and ALT were significantly increased in the population with high lead levels (P<0.05). PbB abnormalities were closely associated with gender, environmental lead concentration, wearing masks, smoking, and alcohol consumption (P<0.05). Regarding occupational hazards, workers exposed to lead dust had a 1.98-fold risk of PbU abnormality compared to those exposed to lead fumes. The plate coating and acid leaching process posed the highest risk for both PbB (OR=8.81) and PbU (OR=5.46) abnormalities compared with assembly process. Furthermore, the risks of PbB and PbU abnormalities were significantly elevated among workers with abnormal BP, RBC or ALT (P<0.05). Among the models, CatBoost performed best in predicting RBC abnormality (accuracy: 95.8%; precision: 44.9%; F1 score: 0.952; AUC: 0.981). Feature importance analysis identified PbB and PbU as the core factors affecting abnormal RBC and ALT, while RBC and ALT abnormalities as key features for predicting the risk of PbB and PbU abnormalities. Conclusion By integrating traditional statistical methods with machine learning, this study reveals a complex bidirectional association between occupational lead exposure and multiple health indicators, and identifies gender, job category, and environmental Pb concentration as the key factors influencing PbB abnormalities. These findings provide a scientific foundation for the implementation of precision occupational health management models.

OBJECTIVE This study aims to establish a prompt engineering system for large language models in pharmaceutical popularization， and provide references for pharmacists to carry out efficient and standardized science popularization work. METHODS This study systematically expounded the principles and classifications of prompt engineering， as well as its effect on alleviating problems including model output hallucinations and poor interpretability. The design and optimization strategies of prompt engineering were defined for two core scenarios， namely text-to-text and text-to-image. Typical examples were adopted to compare the output effects before and after application. In addition， the limitations of prompt engineering applied in pharmaceutical popularization at the current stage were summarized. RESULTS In the two major scenarios of pharmaceutical popularization， well-designed prompt engineering improved the accuracy， readability and efficiency of outputs generated by large language models， and produced personalized popularization content adapted to clinical practice. CONCLUSIONS Prompt engineering can effectively improve the output quality of pharmaceutical popularization. The formulated standardized prompt engineering templates tailored for pharmaceutical popularization， can help pharmacists improve the efficiency and quality of popularization content creation.