连续性变量自动分级数据库设计在最小随机化分组系统中的应用*

目的:使最小随机化分组系统支持对连续性变量因素的自动分级,并检验其均衡能力。方法:以Po-cock和Simon最小随机化法进行分组,使用EAV模式的数据库结构存储数据。通过一个实际的临床试验检验本方法的可行性和分组系统对连续性变量分组因素的均衡能力。结果:该方法简化了数据录入人员操作,临床试验共入组患者215例,通过SAS对3个连续性变量在两个治疗组间的分布进行均衡性检验,P值分别为0.9615,0.7461和0.6843。结论:通过本研究的数据库设计方法,最小随机化分组系统可以简便地对治疗组间的连续性变量因素进行均衡分组,同时更好地确保了分组结果的不可预测性。

Database design for automatic classification of continuous prognostic factors in minimization random allocation system

Objective: To design and demonstrate a continuous prognostic-factor's automatic classification in minimization random allocation system.Methods: This function was constructed with EAV database schema design and Pocock and Simon's minimization method.A real clinical trial with 3 continuous prognostic factors was randomized through this system to test its facility and applicability.Results: There were 215 patients enrolled in the clinical trial.The result of random allocation was analyzed by SAS and 3 continuous prognostic factors were balanced.The P values of them were 0.961 5,0.746 1 and 0.684 3.Conclusion: Clinical trials with continuous prognostic factors can be easily balanced with the method in this article,and at the same time,the predictability of next assignment can be decreased.