The development of large language models in the field of Traditional Chinese Medicine(TCM)using artificial intelli-gence has significantly contributed to both the innovation and preservation of TCM.This paper outlines the current research status and process of large language models in TCM,focusing on three key tasks:collecting TCM data,fine-tuning models with specific instruc-tions,and using different methods to evaluate model performance.It also highlights cutting-edge techniques,such as prompt engineer-ing,retrieval-augmented generation,and reinforcement learning from human feedback,which have enhanced the models'adaptability in various TCM applications.The challenges faced by TCM language models,such as data privacy,ethical biases,model interpret-ability,technical difficulties,and evaluation standards,are also analyzed,indicating areas for further improvement.Looking ahead,the combination of these models with advanced AI techniques like deep learning,and the integration of multimodal information such as TCM diagnostic data and herbal images,can open new possibilities for large language models in TCM.This will enhance their applica-tion in areas like syndrome differentiation diagnosis,prescription recommendations,TCM knowledge graph construction,and TCM edu-cation.

The development of large language models in the field of Traditional Chinese Medicine(TCM)using artificial intelli-gence has significantly contributed to both the innovation and preservation of TCM.This paper outlines the current research status and process of large language models in TCM,focusing on three key tasks:collecting TCM data,fine-tuning models with specific instruc-tions,and using different methods to evaluate model performance.It also highlights cutting-edge techniques,such as prompt engineer-ing,retrieval-augmented generation,and reinforcement learning from human feedback,which have enhanced the models'adaptability in various TCM applications.The challenges faced by TCM language models,such as data privacy,ethical biases,model interpret-ability,technical difficulties,and evaluation standards,are also analyzed,indicating areas for further improvement.Looking ahead,the combination of these models with advanced AI techniques like deep learning,and the integration of multimodal information such as TCM diagnostic data and herbal images,can open new possibilities for large language models in TCM.This will enhance their applica-tion in areas like syndrome differentiation diagnosis,prescription recommendations,TCM knowledge graph construction,and TCM edu-cation.

Objective: To observe the short-term changes of Body Mass Index (BMI) and Waist Circumference (WC) in junior high school students in Harbin, and to explore its association with the changes of high-sensitivity C-reactive protein (hs-CRP) in the same period. Furthermore, to analyze the value of hs-CRP to predict the BMI and WC changes in children and adolescents. Methods: Convenient sampling method was used,and the students in grades 6 and 7 in junior high school in Harbin were selected by stratified clicstering. Physical examination and serum hs-CRP were performed for two consecutive years. According to the results of the two surveys, the subjects were divided into control group (normal → normal), case group (abnormal → abnormal), improved group (abnormal → normal) and progression group (normal → abnormal). The multinomial logistic regression was used to analyze the predictive value of hs-CRP changes for BMI and WC changes. Results: The rate of overweight and obesity among samples from junior high school was 30.6%. The increment of BMI in female was more obvious than that in male(U=17 358.0，P<0.05). Both increments in BMI and WC were positively correlated with increments in hs-CRP(P<0.05). The risk of occurrence of "BMI always abnormal" and "WC always abnormal" increased sequentially in hs-CRP "normal → abnormal", "abnormal → normal" and "abnormal → abnormal", were 3.45 times, 5.98 times and 38.84 times of "BMI is always normal", respectively; and were 3.45 times, 4.26 times and 28.50 times of "WC is always normal", respectively. The risk of "BMI improvement" was 7.70 times more than that of BMI "always normal" when hs-CRP "abnormal → normal". Conclusion The prevalence of overweight and obesity in junior high school students in Harbin is high. The BMI increases faster in female from junior high school. The trends of change of BMI and WC are consistent with the trend of change of hs-CRP. The “hs-CRP change” has a good predictive value for BMI and WC development.

Objective To explore the application conditions of 22 S rule (one of the Westgard rules) in single concentration quali‐ty control .Methods The application scope of 22 S rule was used into the same batch detection of single‐concentration quality control materials .Results The 22 S rule improved the detection rate of system errors .Conclusion 22S rule should be applied to two detec‐tions within a batch when using one quality control material ,as to improve the detection of system error more efficiency .

Objective To evaluate the clinical efficacy and safety of flexible fiberoptic bronchoscopy (FFB) and bronchoalveolar lavage (BAL) in treatment of neonatal atelectasis.Methods Eligible patients, who were diagnosed as neonatal pulmonary atelectasis and admitted consecutively to the neonatal intensive care unit of the Children's Hospital of Hebei Province from January 2013 to January 2015, were included in the study.They were randomly assigned to FFB group (n=30) and control group (n=28).Newborns in the FFB group received BAL under FFB, while those controls received tracheal irrigation after intubation.The duration of lung recruitment, oxygen exposure and antibiotic administration, hospital stay, culture results of respiratory secretions, prognosis and total expenses during hospitalization were compared between the two groups.Complications of FFB were also recorded.Chi-square test and t-test were performed for statistical analysis.Results (1) In the FFB group, atelectasis occurred in the upper fight lobe (n=26), upper lobes of both sides (n=1), lower right lobe (n=2) and lower left lobe (n=1), while in the control group, atelectasis occurred in the upper right lobe (n=26), lower left lobe (n=1) and middle right lobe (n=2) (x2=0.094, P ＞ 0.05).(2) The positive rate of bacteria culture results showed no difference between bronchoalveolar lavage fluid in FFB group and tracheal secretions in the controls [43％(13/30) vs 32％(9/28), x2=0.770, P ＞ 0.05].(3) The duration of lung recruitment, antibiotic administration and hospital stay of the FFB group were all shorter than those of the control group [(4.7±3.4) vs (7.4±6.6) d, (14.0±4.5) vs (20.3±10.9) d, (15.1±4.7) vs (21.8±12.3) d, t=-5.718, 8.604 and 7.733, all P ＜ 0.05].(4) Among babies in the FFB group, nine experienced fever and returned to normal after physical cooling;three showed more shadow in chest X-ray with aggravated dyspnea during a short period, and relieved 12 hours later;two had minimal hemorrhage from tracheal mucous membrane;one showed crying hoarse.Serious complications, such as pneumothorax, massive bleeding or cardiac arrest, did not occurred.No death or refuse of treatment was reported.Conclusion FFB and BAL is much more effective than tracheal irrigation after intubation in treatment of neonatal atelectasis without any severe complications.