倾向性评分方法及其在伽玛刀治疗垂体腺瘤疗效评价中的应用

目的]本课题采用传统多因素分析和基于倾向性评分的3种方法比较伽玛刀分次和单次治疗垂体腺瘤患者的疗效,证实倾向性评分可以用来平衡两个组的协变量,从而降低偏倚。方法]采用t检验和简单线性回归方法逐个分析每一因素的作用,利用多因素回归平衡混杂因素来探讨处理方法的效果;倾向性评分方法均衡数据后采用1︰1配对、分层及分层后回归调整比较处理方法的效果。结果]①多因素分析平衡了其他因素后处理方法仍然为术后疗效的显著影响因素。②经配对、分层后,伽玛刀分次治疗组垂体肿瘤的体积差小于对照组,差异均具有统计学意义,采用分层后回归调整处理方法后,分次伽玛刀治疗的体积增大风险是单次伽玛刀治疗的22.1%,分次伽玛刀治疗的效果要优于单次伽玛刀。③两方法结果一致,平衡了其他因素后处理方法不同术后疗效不同,分次伽玛刀治疗的效果要优于单次伽玛刀。结论]倾向评分法能够有效地均衡各对比组间特征变量的分布和构成,并在组间均衡的基础上评价干预措施或危险因素与结果变量间的联系或作用。

PROPENSITY SCORE METHODS AND ITS APPLICATION IN GAMMA KNIFE TREATMENT OF PITUITARY ADENOMA EVALUATION

Objective]The traditional multi-factor analysis and three methods based on propensity score was used to compare the efficacy of Gamma Knife treatment of multiple and single in patients with pituitary adenoma.We confirmed that the propensity score can be used to balance the covariates in both groups to reduce the bias.Methods]t test and simple linear regression was taken to analyze the role of each factor one by one,a multi-factor analysis of weakening confounding factors was used to investigate the effect...