Department of Statistics, Florida State University, Tallahassee, Florida, USA.
Statistical Research and Innovation, Global Biometrics and Data Management, Pfizer Inc, New York, New York, USA.
BMJ Open. 2022 May 9;12(5):e055336. doi: 10.1136/bmjopen-2021-055336.
Several methods are commonly used for meta-analyses of diagnostic studies, such as the bivariate linear mixed model (LMM). It estimates the overall sensitivity, specificity, their correlation, diagnostic OR (DOR) and the area under the curve (AUC) of the summary receiver operating characteristic (ROC) estimates. Nevertheless, the bivariate LMM makes potentially unrealistic assumptions (ie, normality of within-study estimates), which could be avoided by the bivariate generalised linear mixed model (GLMM). This article aims at investigating the real-world performance of the bivariate LMM and GLMM using meta-analyses of diagnostic studies from the Cochrane Library.
We compared the bivariate LMM and GLMM using the relative differences in the overall sensitivity and specificity, their 95% CI widths, between-study variances, and the correlation between the (logit) sensitivity and specificity. We also explored their relationships with the number of studies, number of subjects, overall sensitivity and overall specificity.
Among the extracted 1379 meta-analyses, point estimates of overall sensitivities and specificities by the bivariate LMM and GLMM were generally similar, but their CI widths could be noticeably different. The bivariate GLMM generally produced narrower CIs than the bivariate LMM when meta-analyses contained 2-5 studies. For meta-analyses with <100 subjects or the overall sensitivities or specificities close to 0% or 100%, the bivariate LMM could produce substantially different AUCs, DORs and DOR CI widths from the bivariate GLMM.
The variation of estimates calls into question the appropriateness of the normality assumption within individual studies required by the bivariate LMM. In cases of notable differences presented in these methods' results, the bivariate GLMM may be preferred.
诊断研究的荟萃分析通常采用多种方法，如双变量线性混合效应模型（LMM）。它可以估计汇总受试者工作特征（ROC）曲线估计的总敏感度、特异度、它们之间的相关性、诊断比值比（DOR）和曲线下面积（AUC）。然而，双变量 LMM 做出了潜在不切实际的假设（即，研究内估计的正态性），这可以通过双变量广义线性混合效应模型（GLMM）来避免。本文旨在通过对 Cochrane 图书馆中的诊断研究荟萃分析来研究双变量 LMM 和 GLMM 的实际性能。
我们使用汇总敏感度和特异度的相对差异、95%置信区间（CI）宽度、研究间方差以及（对数）敏感度和特异度之间的相关性来比较双变量 LMM 和 GLMM。我们还探讨了它们与研究数量、研究对象数量、总体敏感度和总体特异性之间的关系。
在提取的 1379 项荟萃分析中，双变量 LMM 和 GLMM 得出的总体敏感度和特异度的点估计值通常相似，但 CI 宽度可能明显不同。当荟萃分析包含 2-5 项研究时，双变量 GLMM 通常产生比双变量 LMM 更窄的 CI。对于包含 <100 个研究对象或总体敏感度或特异度接近 0%或 100%的荟萃分析，双变量 LMM 可能会从双变量 GLMM 产生明显不同的 AUC、DOR 和 DOR CI 宽度。
这些估计值的变化表明，双变量 LMM 中所需的各研究内正态性假设可能不适用。在这些方法的结果出现显著差异的情况下，可能更倾向于使用双变量 GLMM。
