Meta-analysis of test accuracy studies using imputation for partial reporting of multiple thresholds.
Meta-analysis of test accuracy studies using imputation for partial reporting of multiple thresholds.
复制标题
DOI：
10.1002/jrsm.1276
复制
发表时间：
2018-03
影响因子：
9.8
通讯作者：
Riley RD
中科院分区：
文献类型：
--
作者：
Ensor J;Deeks JJ;Martin EC;Riley RD
对于报告连续结果的测试，主要研究通常提供多个但通常不同阈值的测试性能。在每个阈值下进行Meta分析时，这会导致数据缺失。标准Meta分析（无插补[NI]）忽略此类缺失数据。最近提出了一种单插补（SI）方法来恢复缺失的阈值结果。在这里，我们提出了一种新的方法，使用离散组合（MIDC）进行多个填补的缺失阈值的结果。新的MIDC方法通过从位于2个已知边界阈值的结果之间的所有可能的离散组合的集合中随机选择来估算缺失的阈值结果。然后在每个阈值处合成插补和观察结果。这被重复多次，并且使用Rubin的规则将每个阈值处的多个合并结果组合起来以给出最终估计。我们通过仿真比较了NI、SI和MIDC方法。这两种插补方法在模拟中优于NI方法。SI和MIDC方法之间的差异通常很小，但后者在估计研究间方差方面明显更好，并且由于合并估计值的标准误略大，通常覆盖范围更广。考虑到阈值的选择性报告，插补方法还降低了汇总受试者工作特征曲线的偏倚。仿真结果表明，插补方法依赖于一个相等的阈值间距假设。给出了一个真实的例子。SI，特别是MIDC方法可用于检查试验准确性研究Meta分析中缺失阈值结果的影响。
For tests reporting continuous results, primary studies usually provide test performance at multiple but often different thresholds. This creates missing data when performing a meta‐analysis at each threshold. A standard meta‐analysis (no imputation [NI]) ignores such missing data. A single imputation (SI) approach was recently proposed to recover missing threshold results. Here, we propose a new method that performs multiple imputation of the missing threshold results using discrete combinations (MIDC). The new MIDC method imputes missing threshold results by randomly selecting from the set of all possible discrete combinations which lie between the results for 2 known bounding thresholds. Imputed and observed results are then synthesised at each threshold. This is repeated multiple times, and the multiple pooled results at each threshold are combined using Rubin's rules to give final estimates. We compared the NI, SI, and MIDC approaches via simulation. Both imputation methods outperform the NI method in simulations. There was generally little difference in the SI and MIDC methods, but the latter was noticeably better in terms of estimating the between‐study variances and generally gave better coverage, due to slightly larger standard errors of pooled estimates. Given selective reporting of thresholds, the imputation methods also reduced bias in the summary receiver operating characteristic curve. Simulations demonstrate the imputation methods rely on an equal threshold spacing assumption. A real example is presented. The SI and, in particular, MIDC methods can be used to examine the impact of missing threshold results in meta‐analysis of test accuracy studies.
登录
查看更多内容
流行病学和临床研究中缺少数据的多重归因：潜力和陷阱。
DOI：
10.1136/bmj.b2393
发表时间：
2009-06-29
期刊：
BMJ (Clinical research ed.)
影响因子：
--
作者：
Sterne JA;White IR;Carlin JB;Spratt M;Royston P;Kenward MG;Wood AM;Carpenter JR
通讯作者：
Carpenter JR
DOI：
10.1136/bmj.h910
发表时间：
2015-03-04
期刊：
BMJ (Clinical research ed.)
影响因子：
--
作者：
Roberts E;Ludman AJ;Dworzynski K;Al-Mohammad A;Cowie MR;McMurray JJ;Mant J;NICE Guideline Development Group for Acute Heart Failure
通讯作者：
NICE Guideline Development Group for Acute Heart Failure
影响因子：
1.7
作者：
Putter, H.;Fiocco, M.;Stijnen, T.
通讯作者：
Stijnen, T.
影响因子：
4
作者：
Steinhauser S;Schumacher M;Rücker G
通讯作者：
Rücker G
影响因子：
7.2
作者：
Aertgeerts, B;Buntinx, F;Kester, A
通讯作者：
Kester, A