Department of Medicine, University of Calgary, Calgary, AB, Canada.
Department of Community Health Sciences, University of Calgary, Calgary, AB, Canada.
Patient. 2022 Nov;15(6):629-639. doi: 10.1007/s40271-022-00587-7. Epub 2022 Jul 13.
Systematic reviews of discrete-choice experiments (DCEs) are being increasingly conducted. The objective of this scoping review was to identify and describe the methodologies that have been used to summarize results across DCEs.
We searched the electronic databases MEDLINE and EMBASE from inception to March 18, 2021, to identify English-language systematic reviews of patient preferences that included at least two DCEs and extracted data on attribute importance. The methods used to summarize results across DCEs were classified into narrative, semi-quantitative, and quantitative (meta-analytic) approaches and compared. Approaches to characterize the extent of preference heterogeneity were also described.
From 7362 unique records, we identified 54 eligible reviews from 2010 to Mar 2021, across a broad range of health conditions. Most (83%) used a narrative approach to summarize findings of DCEs, often citing differences in studies as the reason for not formally pooling findings. Semi-quantitative approaches included summarizing the frequency of the most important attributes, the frequency of attribute statistical significance, or tabulated comparisons of attribute importance for each pair of attributes. One review conducted a meta-analysis using the maximum acceptable risk. While reviews often commented on the heterogeneity of patient preferences, few (6%) addressed this systematically across studies.
While not commonly used, several semi-quantitative and one quantitative approach for synthesizing results of DCEs were identified, which may be useful for generating summary estimates across DCEs when appropriate. Further work is needed to assess the validity and usefulness of these approaches.
越来越多的离散选择实验（DCE）系统评价正在进行。本范围综述的目的是确定和描述用于汇总 DCE 结果的方法。
我们搜索了电子数据库 MEDLINE 和 EMBASE，从建库到 2021 年 3 月 18 日，以确定至少包含两项 DCE 并提取属性重要性数据的患者偏好的英语系统评价。用于汇总 DCE 结果的方法分为叙述性、半定量和定量（荟萃分析）方法，并进行了比较。还描述了用于描述偏好异质性程度的方法。
从 7362 条独特记录中，我们确定了 2010 年至 2021 年 3 月期间的 54 项符合条件的综述，涵盖了广泛的健康状况。大多数（83%）采用叙述性方法总结 DCE 的研究结果，通常引用研究之间的差异作为不正式汇总研究结果的原因。半定量方法包括总结最重要属性的出现频率、属性统计显著性的出现频率，或为每个属性对列出属性重要性的比较。一项综述使用最大可接受风险进行了荟萃分析。虽然评论经常评论患者偏好的异质性，但很少有（6%）系统地在研究之间解决这个问题。
虽然不常用，但确定了几种用于综合 DCE 结果的半定量和一种定量方法，当适当的时候，这些方法可能有助于在 DCE 之间生成汇总估计值。需要进一步的工作来评估这些方法的有效性和实用性。
