药物经济学评价中Meta分析结果应用方法的比较与选择

Comparison and selection of application methods of meta-analysis results in economic evaluations

目的比较Meta分析结果在药物经济学评价中应用的四种常见方法,为今后开展类似药物经济学评价提供参考和建议。方法对奥美拉唑和埃索美拉唑治疗消化性溃疡的Meta分析结果采用四种方法计算干预措施的有效率,并代入决策树模型进行成本-效果分析,对药物经济学评价结果进行对比分析。结果方法一以风险差为增量效果,增量成本-效果比(incremental cost-effectiveness ratio,ICER)为2 420元,概率敏感性分析中成本-效果可接受曲线的等概率点约为2 600元;方法二以高质量文献中研究组的有效率为基准,根据风险比计算对照组有效率,ICER为2 016元,概率敏感性分析中成本-效果可接受曲线的等概率点约为2 000元;方法三以高质量文献中对照组有效率为基准,根据风险比计算研究组有效率,ICER为2 420元,概率敏感性分析中成本-效果可接受曲线的等概率点约为2 200元;方法四使用文献权重加权计算有效率,ICER为2 420元,概率敏感性分析中成本-效果可接受曲线的等概率点约为2 400元。结论四种方法的成本-效果分析结果差异较小,敏感性分析结果表明4种评价结果均较为稳健。但应用过程中,方法一缺少两组具体有效率,在单因素敏感性分析时,低估了增量效果的变化范围;方法二和方法三在敏感性分析中可能会出现有效率大于1的情况,需要通过相关假设进行限定。综合比较,通过方法四计算效果值的缺陷较少,更推荐在药物经济学评价中使用。

Objectives To compare the common application methods of meta-analysis results used in economic evaluations so as to provide reference and suggestions for similar economic evaluations in future. Methods Four methods were used to calculate the effectiveness deriving from meta-analysis of omeprazole and esomeprazole in the treatment of peptic ulcer, then substituted into the decision tree model to perform cost-effectiveness analysis. Results Method I used the risk difference as the incremental effectiveness. The ICER was ￥ 2 420, and the equal probability point of the cost-effectiveness acceptability curve （CEAC） in the probability sensitivity analysis was approximately ￥ 2 600. Method 2 used the effective rate of the study group in high-quality literatures as the benchmark, calculated the effective rate of the control group according to the RR. The ICER was ￥ 2 016, and the equal probability point of the CEAC was approximately ￥ 2 000. Method 3 was based on the effective rate of the control group in high-quality literatures to calculate the effective rate of the study group according to RR. The ICER was ￥2 420 and the equal probability point of the CEAC was approximately ￥ 2 200; Method 4 used literature weights to calculate the effectiveness, the ICER is ￥ 2 420, and the equal probability point of the CEAC was about ￥ 2 400. Conclusions The results of the four methods share little difference, and the sensitivity analysis results show that the base case analysis results are more robust. However, in the application process, method 1 lacks specific effectiveness of the two groups and underestimate the variation range of the effectiveness difference when one-way sensitivity analysis was performed. Relevant assumptions are further required to limit the possibility of effectiveness calculated greater than 1 in sensitivity analysis among method 2 and 3. Comprehensively, method 4 can be recommended in the economic evaluations for fewer defects of calculating effectiveness.