Shouchuang Zhang
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Shouchuang Zhang has a total of 9 co-authors. Here are the top co-authors by number of shared publications:
Jiayi Weng
1 shared publicationLanyue Zhang
1 shared publication- Qiaosheng Li
Department of Health Policy and Management, School of Public Health, Peking University, Beijing 100191, China
1 shared publication - Ting Wang
National Center for Infectious Diseases, Beijing Ditan Hospital, Capital Medical University, Beijing 100015, China
1 shared publication Weiyan Jian
1 shared publication- Weiyan Jian
Department of Health Policy and Management, School of Public Health, Peking University, Beijing 100191, China
1 shared publication
Latest Publications
- Open Access
Urban-rural disparities in the association between social trust patterns and changes in depressive symptoms: longitudinal evidence from an elderly Chinese population
Objective To examine the relationship between social trust and depressive symptoms among China’s elderly, placing special emphasis on the disparities between urban and rural settings. Design We employed latent profile analysis to categorise individual patterns of social trust. Subsequently, multiple linear regression analysis was employed to determine if there was an association between these identified social trust patterns and depressive symptoms. Additionally, we examined the potential interactive effects between urban-rural differences and patterns of social trust on depressive symptoms. Setting The China Family Panel Studies (CFPS) database. Participants The data was sourced from the CFPS for the years 2018 and 2020, encompassing a total of 5645 participants aged 60 and above. Outcome measures Depressive symptoms were evaluated employing an eight-item adaptation of the Centre for Epidemiologic Studies Depression Scale. The scores from these eight items were aggregated to create an index of depressive symptoms, which was used to quantify the severity of depressive symptoms. Results The findings demonstrate a significant link between patterns of social trust and depressive symptoms, with individuals manifesting high social trust (HST) showing a lower propensity for depressive symptoms (Beta=−2.26, 95% CI=−2.62, –1.92). Furthermore, a marked association is apparent between social trust patterns and the changes in depressive symptoms. Additionally, urban dwellers (Beta=−0.67, 95% CI=−1.23, –0.11) demonstrate a more pronounced correlation between patterns of social trust and depressive symptoms, particularly within the HST group. Conclusion Our findings highlight a strong link between social trust patterns and depressive symptoms, particularly regarding their changes. Urbanites, notably within the HST group, show a lower risk of experiencing depressive symptoms. There is an urgent requirement to establish social trust-specific interventions to decrease susceptibility to depressive symptoms among the rural populace.
Open AccessUrban-rural disparities in the association between social trust patterns and changes in depressive symptoms: longitudinal evidence from an elderly Chinese population
Objective To examine the relationship between social trust and depressive symptoms among China’s elderly, placing special emphasis on the disparities between urban and rural settings. Design We employed latent profile analysis to categorise individual patterns of social trust. Subsequently, multiple linear regression analysis was employed to determine if there was an association between these identified social trust patterns and depressive symptoms. Additionally, we examined the potential interactive effects between urban-rural differences and patterns of social trust on depressive symptoms. Setting The China Family Panel Studies (CFPS) database. Participants The data was sourced from the CFPS for the years 2018 and 2020, encompassing a total of 5645 participants aged 60 and above. Outcome measures Depressive symptoms were evaluated employing an eight-item adaptation of the Centre for Epidemiologic Studies Depression Scale. The scores from these eight items were aggregated to create an index of depressive symptoms, which was used to quantify the severity of depressive symptoms. Results The findings demonstrate a significant link between patterns of social trust and depressive symptoms, with individuals manifesting high social trust (HST) showing a lower propensity for depressive symptoms (Beta=−2.26, 95% CI=−2.62, –1.92). Furthermore, a marked association is apparent between social trust patterns and the changes in depressive symptoms. Additionally, urban dwellers (Beta=−0.67, 95% CI=−1.23, –0.11) demonstrate a more pronounced correlation between patterns of social trust and depressive symptoms, particularly within the HST group. Conclusion Our findings highlight a strong link between social trust patterns and depressive symptoms, particularly regarding their changes. Urbanites, notably within the HST group, show a lower risk of experiencing depressive symptoms. There is an urgent requirement to establish social trust-specific interventions to decrease susceptibility to depressive symptoms among the rural populace.
- Open Access
- 1 December 2024
- Published by Cambridge University Press (CUP) in International Journal of Technology Assessment in Health Care
- Vol. 40 (S1) , S63
PP23 Machine-Learning-Based Evaluation Of Community Resilience Through Social Media During The First Post-COVID-19 Reopening In China
Introduction: The Chinese government lifted most COVID-19 pandemic restrictions in December 2022, triggering a spike in confirmed cases and higher demand for medications. Consequently, a significant number of residents resorted to social media to seek assistance. This study aimed to evaluate community resilience by leveraging Weibo user datasets, coupled with interpretable machine learning (ML)-based techniques, to identify important resilience characteristics.Methods: Datasets geotagged from the Sina Weibo social media platform between 8 December 2022 and 7 January 2023 were crawled using search terms of “help-seeking” and the keywords of conventional drugs. This study utilized natural language processing (NLP) to label COVID-19-related posts to identify the type of posts, stakeholders’ behaviors, and other information. We built a comprehensive evaluation model, and five ML-based algorithms were compared for analyzing community resilience. Local interpretable model-agnostic explanations (LIME) was employed to verify five models and the XGBoost algorithm showed optimal effects. Shapley Additive Explanations (SHAP) elucidated the best model’s outputs and estimated contributions for key resilience characteristics.Results: For this study, 199,709 posts were collected. Out of these, 48,425 posts were identified as help-seeking posts, with more than two-thirds receiving responses from community level. The area under curve (AUC) of the XGBoost model was 0.82 (95% confidence interval [CI]: 0.82, 0.83), and the values of accuracy and F1 score were 0.72 and 0.80, respectively. This result demonstrated that the model can successfully evaluate community resilience and subsequently identify the features driving this outcome. Collective efficacy in providing aid, support from official rescue guidelines, and residents’ rapid response to rescue information were identified as the most important characteristics for evaluating community resilience.Conclusions: This study is the first to harness social media data to quantify community resilience in China based on a framework we developed. Five updated ML-based algorithms were developed to evaluate community resilience, and XGBoost showed optimal effects. Three characteristics of community resilience were found as potential predictors that can enhance decision-making support to reshape health emergency rescue activities.
- 1 December 2024
- Published by Cambridge University Press (CUP) in International Journal of Technology Assessment in Health Care
- Vol. 40 (S1) , S63
Open AccessPP23 Machine-Learning-Based Evaluation Of Community Resilience Through Social Media During The First Post-COVID-19 Reopening In China
Introduction: The Chinese government lifted most COVID-19 pandemic restrictions in December 2022, triggering a spike in confirmed cases and higher demand for medications. Consequently, a significant number of residents resorted to social media to seek assistance. This study aimed to evaluate community resilience by leveraging Weibo user datasets, coupled with interpretable machine learning (ML)-based techniques, to identify important resilience characteristics.Methods: Datasets geotagged from the Sina Weibo social media platform between 8 December 2022 and 7 January 2023 were crawled using search terms of “help-seeking” and the keywords of conventional drugs. This study utilized natural language processing (NLP) to label COVID-19-related posts to identify the type of posts, stakeholders’ behaviors, and other information. We built a comprehensive evaluation model, and five ML-based algorithms were compared for analyzing community resilience. Local interpretable model-agnostic explanations (LIME) was employed to verify five models and the XGBoost algorithm showed optimal effects. Shapley Additive Explanations (SHAP) elucidated the best model’s outputs and estimated contributions for key resilience characteristics.Results: For this study, 199,709 posts were collected. Out of these, 48,425 posts were identified as help-seeking posts, with more than two-thirds receiving responses from community level. The area under curve (AUC) of the XGBoost model was 0.82 (95% confidence interval [CI]: 0.82, 0.83), and the values of accuracy and F1 score were 0.72 and 0.80, respectively. This result demonstrated that the model can successfully evaluate community resilience and subsequently identify the features driving this outcome. Collective efficacy in providing aid, support from official rescue guidelines, and residents’ rapid response to rescue information were identified as the most important characteristics for evaluating community resilience.Conclusions: This study is the first to harness social media data to quantify community resilience in China based on a framework we developed. Five updated ML-based algorithms were developed to evaluate community resilience, and XGBoost showed optimal effects. Three characteristics of community resilience were found as potential predictors that can enhance decision-making support to reshape health emergency rescue activities.