Confidence interval for the AUC of SROC curve and some related methods using bootstrap for meta-analysis of diagnostic accuracy studies

04/09/2020 ∙ by Hisashi Noma, et al. ∙ 0

Background: The area under the curve (AUC) of summary receiver operating characteristic (SROC) curve has been a primary statistical outcome for meta-analysis of diagnostic test accuracy studies (DTA). However, its confidence interval has not been reported in most of DTA meta-analyses, because no certain methods and statistical packages to compute it are provided. Methods: We provided a bootstrap algorithm for computing the confidence interval of AUC, and developed an easy-to-handle R package dmetatools (https://github.com/nomahi/dmetatools). Also, using the bootstrap framework, we can conduct a bootstrap test for assessing significance of the difference of AUCs for multiple diagnostic tests. In addition, we provide an influence diagnostic method based on the AUC by leave-one-study-out analyses. The bootstrap framework enables quantitative evaluations for statistical uncertainty of the influential statistic. Results: Using the R package dmetatools, we can calculate these statistical outcomes by simple commands. We also present illustrative examples using two DTA met-analyses for diagnostic tests of cervical cancer and asthma. Conclusions: The bootstrap methods can be easily implemented by simple codes using dmetatools. These statistical outcomes would be recommended to involve in the statistical outcomes of DTA meta-analyses if the authors report the AUC of SROC curve. These various quantitative evidence certainly supports the interpretations and precise evaluations of statistical evidence for DTA meta-analyses.

READ FULL TEXT
POST COMMENT

Comments

There are no comments yet.

Authors

page 1

page 2

page 3

page 4

This week in AI

Get the week's most popular data science and artificial intelligence research sent straight to your inbox every Saturday.