广义估计方程处理重复测量数据的参数解释

摘要

Objective To study the principle and methods of generalized estimating equation, and investigate the specific significance of parameter estimation values in the application of different data types. Methods The effi-cacy of a medicine for improving pain in the gastral cavity of chronic superficial gastritis was taken as an example for the generalized estimating equation analysis on repeated measurement data. Meanwhile, the efficacy of the medicine for treating heartburn 3 week and 6 week of atrophic gastritis was taken as an example for the generalized estimating equation analysis on ranked data and specific parameter interpretation. Results ①The results of generalized estimat-ing equation showed that the test drug had better efficacy in improving pain in the gastral cavity of chronic superficial gastritis than that of the control drug. ②The results of generalized estimating equation for the efficacy of treating heartburn showed that there were differences in the treatment conditions between different treatment courses while the condition of treatment 6 weeks was improved. However, the generalized estimating equation showed that there were no statistically significant differences in the efficacy between the experimental group and the control group. Conclusion Generalized estimating equations is a powerful method to analyze the repeated measurement data. And there are dif-ferent means and perspectives for the interpretation of in practical applications. The interpretation of partial regression coefficients in quantitative and qualitative perspectives can provide more comprehensive information for the study of clinical medicine.%目的：研究广义估计方程的原理与方法，探讨参数估计值的具体意义及其在不同类型资料中的应用。方法以某药物治疗慢性浅表性胃炎主症胃脘疼痛的改善为例做对重复测量数据的广义估计方程分析；以该药物治疗萎缩性胃炎主症中胃灼热痛3周与6周疗效为例做对等级资料的广义估计方程分析，并做具体的参数解释。结果①利用广义估计方程分析结果表明，试验药治疗慢性浅表性胃炎对主症中胃脘疼痛症状的改善优于对照药。②对胃灼热痛疗效的广义估计方程表明，治疗疗程间疗效情况有差异，治疗6周情况改善。但广义估计方程表明试验组与对照组的疗效差异无统计学意义。结论广义估计方程是分析重复测量资料的强有力的手段。对于其参数的解释在实际应用中针对不同的资料有不同的解释方法与角度，从定量与定性角度对偏回归系数进行解释可以为临床药物的研究提供更全面的信息。