基于经验似然修正的艾滋病药物疗效对比分析研究

ACTG 175数据是比较研究艾滋病药物疗效的重要临床试验数据.为了更稳健有效的对不同治疗方案进行比较，该文通过观察第20周时的CD4＋ T细胞数目的中位数差异，提出基于经验似然修正的中位数回归方法来分析研究不同药物治疗方案对艾滋病的疗效，并利用Bootstrap抽样方法分析各组方案的疗效差异.分析结果表明zidovudine和didanosine联合用药效果优于zidovudine单独用药的效果，zidovudine和didanosine联合用药的抗艾滋效果显著最优，zidovudine单独用药的抗艾滋效果要显著劣于zido-vudine、zalcitabine联合用药和didanosine单独用药，zidovudine和zalcitabine联合用药和didanosine单独用药的抗艾滋效果没有显著差别.

Comparative analysis of AIDS drug efficacy based on empirical likelihood correction

ACTG 175 data is an important clinical trial data for comparative study on the efficacy of AIDS drugs. In order to compare the different treatment schemes more robustly and effectively， this paper proposed a median regression method based on empirical likelihood correction to analyze the efficacy of different drug treatment schemes on AIDS by observing the median difference in the number of CD4＋ T cells at week 20. And Bootstrap sampling method was used to analyze the difference in efficacy of different treatment schemes. The results show that combination of zidovudine and didanosine is the best for anti-HIV， and using zidovudine alone is worse than using combination of zidovudine and zalcitabine and using didanosine alone. However， there is no significant difference between using combination of zidovudine and zalcitabine and using didanosine alone.