雷公藤多苷导致闭经是影响育龄女性用药的重要因素，高准确度的闭经预测模型可以对闭经发生进行早期预警、预后和转归评估，为临床用药提供有力指导。传统研究往往重视闭经发生相关的个体因素，忽略其中的群体层面特征影响以及群体因素与个体因素的相互作用，因而目前尚缺乏高准确性的风险预测模型，且无法评估闭经的预后与转归。本研究拟在采用Meta分析方法全面收集并量化该类闭经发生相关危险因素的基础上，联合巢式病例对照研究，全面筛查类风湿关节炎育龄女性服用雷公藤多苷后发生闭经危险因素，应用Multilevel Model，综合全部相关因素，从患者个人水平及群体水平特征进行考虑，明确影响闭经发生的各层次关键因素。应用前瞻性随访验证模型的预测价值，对具有重要预测意义的危险因素进行评估，建立群体层面聚集性下能够准确评估个体闭经发生、预后和转归的模型，以期弥补单水平分析下传统指标预测能力的不足，指导临床更精确地选择用药人

Amenorrhea is an important adverse outcome among women of childbearing age with Tripterygium Glycosides therapy. Prediction model with high accuracy will help clinical decision. Many previous studies included with individuals’ common characteristics and failed with association between individual and cluster characteristics. So it is in high demand for prediction methods to predict amenorrhea among women of childbearing age. We will systematically review studies with unadjusted OR or RR values to address major high risk factors associated to amenorrhea and applied Meta analysis to measure the association. Use multilevel models to estimate the association between amenorrhea and the characteristics of.higher level (cities, hospitals), the characteristics of lower level (women), based on the data collected from nested case-control study. To verify the prediction models, we used it to analyze cohort study data. We applied ROC approach to define the best cut point of the model with its sensitivity and specificity. Finally, through this study, our established model is an appropriate forecasting method to yield more accurate results than the original methods. It is highly successful in predicting an individual's risk of amenorrhea among women of childbearing age with Tripterygium Glycosides therapy.