Comparative analysis of two artificial intelligence models for the diagnosis and treatment guidelines of primary osteoporosis and the interpretation of clinical bone mineral density examination

Objective To provide empirical evidence for the application of artificial intelligence model in the clinical diagnosis and treatment of osteoporosis. Methods The core chapters of the guidelines for the diagnosis and treatment of primary osteoporosis were selected，and 10 orthopedic surgeons compared the answers of ChatGPT and DeepSeek with the guidelines from the four dimensions of accuracy，comprehensiveness，interpretability，and clinical applicability. Combined with the results of clinical bone mineral density test，the accuracy and scientificity of the model decision-making were evaluated. Results The scores of 10 orthopedic surgeons were consistent. The accuracy score of DeepSeek was significantly higher than that of ChatGPT，and DeepSeek was more accurate in drug management and parameter analysis，which conformed to the key points of guideline update. Although there was no statistical difference in comprehensiveness and interpretability score，DeepSeek built a multidimensional evaluation framework by integrating drug leave management，FLS collaboration mode and OSTA screening tools，and used charts to improve the efficiency of information presentation. The clinical application score of DeepSeek was better than that of ChatGPT，and its decision-making path and full-cycle monitoring system effectively shortened the time consumption of clinical decision-making. Clinical bone mineral density analysis showed that ChatGPT only provided a basic diagnosis and treatment framework，while DeepSeek showed a more complete guideline compliance and clinical applicability. Conclusion AI can improve the efficiency and quality of clinical decision-making in osteoporosis，and the performance of DeepSeek interpretation is better than that of ChatGPT.