Bloom Bernard S, de Pouvourville Nathalie, Libert Simon
University of Pennsylvania, USA.
Int J Technol Assess Health Care. 2002 Winter;18(1):120-6.
The role of classical and Bayesian statistical approaches remains in dispute in health services research and policy. The goal of this study was to determine if results differ when both analytic techniques are used with the same data set.
We searched MEDLINE and related databases for English-language articles published January 1, 1978 through August 31, 1999. We combined Bayesian and classical statistics search terms and their variants with randomized control trials (RCTs) and meta-analyses.
Searches found 18 studies in 14 publications that met all review criteria--nine RCTs, eight meta-analyses, and one epidemiologic estimate. Statistical analyses using both methods agreed in five RCTs, four meta-analyses, and for the epidemiologic estimates. For four RCTs where results disagreed, classical analysis found the experimental intervention was efficacious compared with the control, and Bayesian reanalysis concluded the intervention was not proven efficacious. Classical meta-analyses of the four studies where results disagreed concluded the experimental intervention was not better than the control; Bayesian reanalysis concluded it was efficacious.
Classical and Bayesian methods in this review exhibited important divergence of results. Disagreement on many fundamental beliefs between classical and Bayesian statistics means continuing debate. One way to resolve this debate is for proponents of each technique to decide together the circumstances for use of each method and analytic framework. If the experts do not agree on the methodologic requirements, other decision makers likely will force their own views.
在卫生服务研究与政策中，经典统计方法和贝叶斯统计方法的作用仍存在争议。本研究的目的是确定当两种分析技术应用于同一数据集时结果是否不同。
我们检索了MEDLINE及相关数据库，以查找1978年1月1日至1999年8月31日发表的英文文章。我们将贝叶斯统计和经典统计的检索词及其变体与随机对照试验（RCT）和荟萃分析相结合。
检索发现14篇出版物中的18项研究符合所有综述标准——9项RCT、8项荟萃分析和1项流行病学评估。使用两种方法进行的统计分析在5项RCT、4项荟萃分析以及流行病学评估中结果一致。在4项结果不一致的RCT中，经典分析发现实验性干预与对照相比有效，而贝叶斯重新分析得出该干预未被证明有效。对4项结果不一致的研究进行的经典荟萃分析得出实验性干预并不优于对照；贝叶斯重新分析得出其有效。
本综述中的经典方法和贝叶斯方法在结果上表现出重要差异。经典统计学和贝叶斯统计学在许多基本信念上的分歧意味着争论仍将继续。解决这一争论的一种方法是每种技术的支持者共同决定每种方法和分析框架的使用情况。如果专家们在方法学要求上无法达成一致，其他决策者可能会强行推行他们自己的观点。
