Shaoyang Cui

Background: This single-center cross-sectional study examined health information-seeking behavior (HISB) among hospitalized patients with stroke, a population with substantial information needs during rehabilitation and discharge planning. Evidence remains limited on the joint associations of social support, quality of life, and multidimensional health literacy with HISB. Objective: To characterize overall and domain-specific HISB and examine its associations with social support, quality of life, and health literacy. Methods: We recruited 231 patients from the acupuncture, neurology, and rehabilitation departments of a tertiary hospital in Shenzhen, China, between October 2025 and March 2026. Participants completed the Chinese 43-item HISB Scale, Social Support Rating Scale (SSRS), WHOQOL-BREF, and a 24-item chronic-disease health-literacy scale. Pearson correlations and two-block hierarchical multiple linear regression were used. The WHOQOL-BREF composite was a study-specific sum of four domain scores, not an official WHOQOL-BREF total score. Results: Mean HISB was 3.27 (SD 0.20; possible range, 1– 5). Attitude (4.08 [SD 0.53]) and information needs (3.89 [SD 0.30]) exceeded source use (2.84 [SD 0.60]). Model 2 explained 22.2% of HISB variance (adjusted R²=0.183). Social support (B=0.012; beta=0.312; 95% CI, 0.007– 0.018; P< 0.001) and quality of life (B=0.005; beta=0.197; 95% CI, 0.001– 0.009; P=0.015) were positively associated with HISB. Total health literacy was not independently associated with HISB (P=0.155). Some sociodemographic coefficients were sensitive to the standard-error estimator, supporting cautious interpretation. Conclusions: The findings identify cross-sectional associations rather than causal effects. Social support and quality of life were consistently associated with more active HISB, whereas total health literacy was not independently associated after adjustment. Sociodemographic findings were less stable and require replication. Multicenter longitudinal studies are needed to confirm these associations.

Background: Effective options for managing chemotherapy-induced peripheral neuropathy (CIPN) remain limited. Evidence supporting acupuncture for CIPN continues to accumulate, warranting a comprehensive synthesis to clarify its effects. Methods: We searched eight databases and extracted study characteristics from the included studies. We assessed risk of bias using the Cochrane RoB 2 tool and graded the certainty of evidence using the GRADE framework. Meta-analyses evaluated CIPN symptoms, clinical effectiveness rate (ER), nerve conduction study (NCS) parameters, and quantitative sensory testing (QST) outcomes. Prespecified subgroup, sensitivity, publication-bias, and trial sequential analyses (TSA) were conducted. Results: The review included 40 RCTs involving 2496 participants. Compared with controls, acupuncture interventions, used alone or with cointerventions, improved several CIPN symptom outcomes, including FACT-Ntx scores (MD = − 1.43, 95% CI [− 1.91, − 0.95]) and increased the ER (RR = 1.74, 95% CI [1.57, 1.93]). Acupuncture also improved nerve conduction velocity and amplitude; however, no significant effects were found for worst pain scores, onset latency, or QST outcomes. GRADE rated the certainty of evidence for FACT-Ntx, EORTC QLQ-CIPN20, and ER as very low. TSA crossed the monitoring boundaries for FACT-Ntx, EORTC QLQ-CIPN20, and ER. Subgroup findings varied by acupuncture type, intervention purpose, and chemotherapy regimen. Conclusion: Both manual acupuncture and electroacupuncture were associated with higher ERs, whereas their effects on specific outcomes varied by intervention type, including whether acupuncture was combined with other treatments. LI4 and ST36 were the most frequently used acupoints, and distal hand and foot acupoints were also commonly selected. Future studies should assess how acupuncture modality, intervention purpose, and chemotherapy regimen influence outcomes. Because the certainty of evidence was very low, these findings should be interpreted cautiously.

Background: Small sample dataset and heterogeneous distributions limit the robustness and implementability of machine learning models for structured clinical data. To evaluate the value of moderate data augmentation for cardiovascular risk modeling and propose an interpretable and deployable solution within a “continuous risk to thresholding” framework. Methods: The heart disease classification dataset was randomly divided into training and validation sets with an 8:2 ratio. Constrained feature space augmentation was performed within the training set, and the effects of four thresholds (0×, 1×, 2×, and 3×) on support vector regression (SVR), random forest (RF), extreme gradient boosting (XGBoost), light Gradient Boosting Machine (LightGBM), and multi-layer perceptron (MLP) were compared. Continuous risk scores were evaluated using mean absolute error (MAE), root mean squared error (RMSE), and R ² . The output confusion matrix was then thresholded. Interpretative analysis was performed using SHapley Additive exPlanations (SHAP), and partial dependence plot (PDP). Results: 2 × augmentation achieved the favorable compromise between error (reduced MAE and RMSE) and goodness of fit (increased R ² ). Under the 2 × condition, RF performed favorable overall, achieving an accuracy of 94.0%, an F2 of 94.4%, a sensitivity of 95.9%, and a specificity of 91.8% after thresholding. SHAP/PDP shows that oldpeak, num major vessels, chest pain type, thal, exang, and max hr. are the primary driving factors, exhibiting stable patterns such as monotonic, step, and threshold. Moderate data augmentation (preferably 2×) can significantly improve robustness in small sample settings; RF achieves the favorable balance between accuracy, stability, and interpretability. Conclusion: Combining SHAP/PDP’s multi-layered interpretation and thresholding approach, this study provides a reusable multiplication guidance and risk stratification scheme, providing a methodological basis for deploying interpretable cardiovascular risk models.

Background: Pharmacological treatment of motor disorders in early Parkinson (PD) has reached a bottleneck, and acupuncture as an important component of non-pharmacological therapy has demonstrated therapeutic potential. However, current evidence is limited in both quantity and quality, highlighting the need for further clarification of its efficacy. Methods: A multi-center, investigator and participant blinded, randomized parallel group design was employed. 136 eligible participants were randomly assigned in a 1:1 ratio to either the acupuncture or sham acupuncture group via a central randomization system, received 12 sessions over 4 weeks. The primary outcome was motor function, assessed using the Unified Parkinson’s Disease Rating Scale-III (UPDRS-III). Secondary outcomes were objective gait metrics captured with the Ready system, disease stage assessed by the Hoehn and Yahr scale (HY), and quality of life measured by the Activities of Daily Living scale (ADL) and the Parkinson’s Disease Questionnaire 39 (PDQ 39). Levodopa dosage was recorded before and after treatment. Assessments were conducted at four time points (weeks 1, 2, 3, and 4), with 8 weeks follow up. Blood samples were collected before and after treatment for LC-MS/MS untargeted metabolomics and deep proteomics analysis. Discussion: This study will aim to clarify the efficacy of acupuncture for early PD motor symptoms through an evidence-based randomized controlled trial (RCT), and explore the underlying mechanisms of acupuncture using multi-omics approaches.

Background/objectives: Acupuncture is recommended for the treatment of cancer-related fatigue (CRF), but the strength of evidence remains limited. This study aims to comprehensively update the evaluation of the effectiveness of acupuncture for CRF and clarify the latest strength of evidence. Methods: Eight databases were searched from inception to April 2025 to identify eligible randomized controlled trials (RCTs). The quality of the study was assessed using the Cochrane risk of bias tool version 2.0(ROB 2.0). Meta-analyses were performed using R studio 4.4 software and RevMan software (version 5.4). Subgroup analyses by acupuncture type were conducted to explore sources of heterogeneity and further elucidate the efficacy of different acupuncture modalities. Publication bias was assessed when applicable. Finally, evidence quality was rated using the GRADE system. Results: This meta-analysis included 28 RCTs. Risk of bias assessment indicated low overall study quality. The meta-analysis of the primary outcome(Piper Fatigue Scale, PFS) revealed that acupuncture intervention significantly improved fatigue levels compared to the control group (MD = -0.56, 95% CI: -0.74 to -0.38, I2 = 45%, P < 0.01). Subgroup analyses showed that different acupuncture type had distinct advantages, suggesting distinct targeted advantages. Other factors showed significant differences. Sensitivity analyses for multiple outcome measures yielded unstable results, with indications of publication bias. According to GRADE criteria, the most outcomes were rated as low or very low quality. Conclusions: Current evidence suggests that acupuncture may alleviate CRF. However, conclusive evidence supporting its therapeutic efficacy remains limited. Given the methodological concerns and low certainty of the available evidence, further high-quality studies are needed to confirm these findings. Systematic review registration: https://www.crd.york.ac.uk/prospero/ , identifier CRD42024603184.

Background: Stroke is one of the leading causes of death and disability worldwide, making early screening and risk prediction crucial. Traditional methods have limitations in handling nonlinear relationships between variables, class imbalance, and model interpretability. Methods: Logistic regression (LR), random forest (RF), extreme gradient boosting (XGBoost), categorical boosting (CatBoost), multi-layer perceptron (MLP) neural network, and ensemble models were constructed and compared. Their performance in stroke risk prediction was systematically evaluated, and feature contributions were interpreted using SHapley Additive exPlanations (SHAP). Confusion matrices and Precision-Recall (PR) curves were used to compare the differences in recognition of the positive class (stroke patients) among the models, and training time was calculated to quantify resource consumption. Results: The ensemble model and neural network demonstrated superior overall predictive ability to traditional algorithms, with the MLP performing particularly well in terms of recall. SHAP results revealed that “hypertension,” “average blood glucose level,” and “age” were key influencing factors. Confusion matrices and PR curves indicated differences in positive classification among the models. Training time analysis provided a basis for resource assessment for subsequent deployment. Conclusion: Machine learning methods have advantages in stroke risk prediction. Incorporating interpretability analysis can enhance the clinical credibility of the models, providing data and methodological reference for stroke risk stratification management and early warning.

Background: Acupuncture shows potential in treating nausea and vomiting (CINV) induced by highly emetogenic chemotherapy (HEC). However, the certainty of its efficacy evidence remains unclear, warranting a comprehensive evaluation. Method: Two independent reviewers systematically searched eight databases from inception to December 2024 to identify eligible randomized controlled trials (RCTs). Relevant data were extracted using a standardized form, and risk of bias was assessed using the Cochrane risk of bias tool version 2.0 (ROB 2.0). Meta-analysis was performed using R Studio 4.4 software. Subgroup analysis was conducted based on acupuncture type. Additionally, publication bias was detected using appropriate methods according to the heterogeneity of different outcomes, where appropriate. Finally, evidence quality was rated using the GRADE system. Result: A total of 58 RCTs were included in this meta-analysis. Rob 2.0 result indicated that most studies were at high risk of bias, with low methodological quality. For the primary outcome, acupuncture significantly improved the complete control rate during the overall (RR 1.54, 95% CI 1.36–1.75; P < 0.001; I ² = 36%) and the delayed phase (RR 1.56, 95% CI 1.32–1.86; P < 0.001; I ² = 0%). For other CINV outcomes, acupuncture demonstrated considerable therapeutic potential for vomiting-related outcomes, while uncertainty in alleviating nausea symptoms. Subgroup analyses showed that different acupuncture types had distinct advantages. Sensitivity analyses for several outcomes were unstable, and there were indications of publication bias. According to GRADE, only the acute vomiting duration score was rated as moderate quality; all other outcomes were rated as low or very low quality. Conclusion: Although acupuncture for HEC-induced CINV shows some positive effects, there are various limitations that render the current evidence insufficient to conclusively establish its efficacy; therefore, further high-quality studies are required. Systematic review registration: PROSPERO, identifier: CRD42024588165.

Background: Artificial intelligence (AI) technologies are increasingly being integrated into clinical nursing practice. However, nurses with different professional backgrounds may differ in their attitudes toward AI applications. Traditional Chinese medicine (TCM) nurses, whose practice integrates traditional Chinese nursing techniques, may have distinct perceptions and levels of acceptance of AI compared with general clinical nurses. Methods: A multicenter cross‐sectional survey was conducted from 10 January to 7 April 2026 in three hospitals and 30 community health service centers in South China. A total of 590 TCM nurses were included. Data were collected using a self‐designed general information questionnaire and standardized scales. Latent profile analysis was performed using the 15‐item scores of the Attitude Scale Toward the Use of Artificial Intelligence Technologies in Nursing as observed indicators. Associations with profile membership were examined using R3STEP to account for classification uncertainty, with simultaneous adjustment for 13 covariates. Results: The mean total score for attitudes toward AI applications among TCM nurses was 48.13 ± 13.49. Three latent profiles were identified: the low AI acceptance profile (25.14%), the moderate AI acceptance profile (39.68%), and the high AI acceptance profile (35.18%). R3STEP analyses identified profile‐specific associations of TCM specialist nursing experience, education, professional title, TCM nursing professional level, intelligent‐device use, self‐rated familiarity with AI, AI‐related training, and technology anxiety with profile membership. Conclusions: TCM nurses demonstrated three latent profiles of attitudes toward AI applications: low, moderate, and high AI acceptance. Profile membership was associated with several occupational and technology‐related characteristics, with patterns varying across profile comparisons. These findings may inform AI training and implementation that take differences in nurses’ acceptance into account.

Purpose: Acupuncture shows potential as a treatment for sciatica, but the credibility and consistency of supporting evidence remain unclear, warranting critical and comprehensive evaluation. This overview is aims to assess the reliability, adequacy, and limitations of current evidence on acupuncture for sciatica using a multidimensional approach and further examine its efficacy through a secondary meta-analysis. Methods: Systematic reviews and meta-analyses (SRs/MAs) meeting PICOS criteria were identified from eight databases by two independent reviewers. Evidence reliability was assessed using AMSTAR-2, ROBIS, PRISMA-A, and the GROOVE tool across four domains: methodological quality, bias risk, reporting accuracy, and study overlap. Duplicate randomized controlled trials (RCTs) were excluded based on the Corrected Covered Area (CCA) analysis, and a secondary meta-analysis was conducted. Sensitivity analyses and funnel plots assessed robustness and publication bias. Results: Seven SRs/MAs were included. AMSTAR-2 revealed significant methodological flaws, particularly due to a lack of protocol pre-registration. ROBIS assessments showed a high risk of bias, with most studies relying on single-database searches and lacking comprehensive strategies. PRISMA-A indicated generally low reporting quality, especially regarding descriptions of acupuncture sensation. The GROOVE tool yielded a CCA of 7.23%, reflecting moderate study overlap. The secondary meta-analysis showed that acupuncture significantly improved treatment effectiveness (RR = 1.23; 95% CI: 1.20– 1.26; P = 0.008), reduced pain intensity, and increased pain threshold. Sensitivity analyses confirmed the robustness of results, while funnel plots suggested some publication bias. Acupuncture was generally considered safe across studies. Conclusion: Although current evidence is limited by methodological flaws, publication bias, and poor reporting quality, acupuncture shows promising clinical potential for sciatica. High-quality, rigorously designed studies are needed to confirm its efficacy.