Objective To identify the potential categories of self-awareness of falls in elderly patients with hip osteoarthritis (HOA), and to analyze the characteristics and influencing factors of different categories. Methods A cross-sectional study was conducted on elderly patients with HOA who were hospitalized in the orthopedic ward of West China Hospital of Sichuan University between February and April 2026 by convenience sampling. The participants were investigated by general information questionnaire, Self-awareness of Falls in Elderly scale, Fall Risk Self-rating Scale for the Elderly, Barthel index scale for self-care ability assessment, Family Care Index Questionnaire, Generalized Anxiety Disorder-7 self-rating scale and Berg Balance Scale. The potential categories of self-awareness of falls were divided by potential profile analysis, and the influencing factors of self-awareness of falls were explored by single factor analysis and multinomial logistic regression analysis. Results A total of 301 elderly patients with HOA were included. The self-awareness of falls of elderly patients with HOA could be divided into three potential categories: “low alertness and low cognition” “medium alertness and high cognition” and “high alertness and high cognition”. Compared with “high alertness and high cognition”, patients with lower education level [odds ratio (OR)=0.199, 95% confidence interval (CI) (0.097, 0.411), P<0.001] and fewer chronic diseases [OR=0.654, 95%CI (0.453, 0.945), P=0.024] were more likely to belong to “medium alertness and high cognition”; patients with good self-care ability [OR=0.346, 95%CI (0.192, 0.623), P<0.001], good balance ability [OR=0.119, 95%CI (0.048, 0.293), P<0.001], fewer chronic diseases [OR=0.478, 95%CI (0.292, 0.781), P=0.003], and low degree of pain [OR=0.448, 95%CI (0.276, 0.728), P=0.001] were more likely to belong to “low alertness and low cognition”. Conclusions There is group heterogeneity in the self-awareness of falls of elderly patients with HOA. In clinical work, medical staff should formulate personalized intervention strategies for different types of patients according to their different characteristics.
Objective To evaluate the performance of lightweight Chinese large language models (LLMs) in answering specialized lung cancer questions, and to explore the impact of retrieval-augmented generation (RAG) on model performance. Methods Eleven lightweight Chinese LLMs with parameter sizes ranging from 7B to 32B were included. A lung cancer-specific evaluation dataset consisting of 200 questions [100 A1-type (basic knowledge) and 100 A2-type (clinical case) questions], constructed based on clinical guidelines and thoracic surgery textbooks, was used for assessment. Model performance was evaluated under two conditions (with and without RAG). Accuracy was used to assess model performance, and response latency was recorded to reflect inference efficiency. An accuracy–latency scatter plot was constructed for descriptive analysis of overall model performance. Results All models successfully completed the evaluation. With the introduction of RAG, the overall average accuracy improved from 61.68% to 76.36%. Smaller models demonstrated the most significant improvement (e.g., the accuracy of DeepSeek-7B increased from 32.50% to 60.00%, P<0.001). The average response latency increased from 12.58 s to 13.80 s. The Qwen3 series showed the best overall performance, and Qwen3-32B achieved the highest accuracy under both conditions (76.50% and 84.00%, respectively). After RAG integration, performance differences among model families were markedly reduced. Based on the accuracy-latency trade-off, Qwen3-32B achieved the best balance between accuracy and response latency under the baseline condition, whereas Qwen3-14B demonstrated superior overall performance in terms of accuracy, latency, and computational cost after RAG integration. Conclusion The integration of RAG technology improves the ability of lightweight Chinese LLMs to answer specialized lung cancer questions. Under the dual practical constraints of limited computational resources and medical data security requirements, the "lightweight model+RAG" technical framework may represent a promising deployment solution.