Introduction of DTAmetasa (1.0.0)

DTAmetasa is an Rshiny application aiming to assist users conducting meta-analysis of diagnostic test accuracy.

(For more detailed information about meta-analysis of diagnostic test accuracy, please refer to the Cochrane Screening and Diagnostic Tests Methods Group.)

DTAmetasa provides the following functionalities:

  • Summary descriptive statistics of multiple diagnostic studies with data of true positive, false positive, true negative, false negative. Note that, only data of sensitivities or specificities extracted from the published diagnostic studies CANNOT be analyzed.

  • Meta-analysis of diagnostic test accuracy with results of the difference estimates of summary operating receiver characteristics (SROC) curves and the area under the curves (SAUC).

  • Detection of publication bias in meta-analysis using funnel plot and the trim-and-fill method.

  • Adjustment of publication bias by the likelihood-based sensitivity analysis method. Note that this method is specific to the estimates by the bivariate normal model.

Requirements on the format of data:

  • We suggest users prepare their data in CSV format or text format

  • The column names of data MUST BE TP, FN, FP, TN, indicating true positive, false negative, false positive, and true negative, respectively. The order of the columns does not matter. We present an example data in the first panel for reference.


How to cite

Mizutani, S., Zhou, Y., Tian, Y.-S., Takagi, T., Ohkubo, T., & Hattori, S. (2023). DTAmetasa : An R shiny application for meta‐analysis of diagnostic test accuracy and sensitivity analysis of publication bias. Research Synthesis Methods, 14(6), 916–925. (https://doi.org/10.1002/jrsm.1666)

Main R packages

All the results in this application can be reproduced by using the following R packages


Find information of helps

  • The help button in the top of each panel will provide some helps and notations

  • The panel of Wiki gives the detailed description of models


Release History

Date (yyyy-mm-dd) Version Details
2024-08-01 1.0.0 We added the introduction page
2023-06-30 0.9.1 Beta We updated the followings:

- Change the color of SROC plot in Meta-analysis to distinguish the results from different models; added the legends

- Added the Feedback button

- Added the Reporting issue button

- Change the names of Help tab into Wiki tab

- Added Release history in Wiki

- Added button for taking the survey
2023-02-01 0.9.0 Beta First release

Summary of Prognostic Studies

Meta-analysis of prognosis studies is used to summaries prognostic accuracy at different time points using multiple prognosis studies

1. Functionalities

  • To upload data files, preview data set, and check the correctness of data input
  • The model for meta-analysis is bivariate normal model (Hattori & Zhou. 2016)
  • To produce time-dependent summary ROC (SROC) curves and summary AUC (SAUC)
  • To detect the publication bias in the results
  • To do sensitivity analysis for the publication bias

2. The format of your data

  • Data of log hazard ratio and standard errors or confidence intervals
  • Data of survival probabilities at different time points
  • Data of median survival probabilities

Case Example

We used the meta-analysis for Ki67. (de Azambuja et al. 2007)

Please follow the Steps, and Outputs will give the analytical results.

Download Manual

Data Preview

Bivariate Normal Model for Meta-Analysis

Meta-analysis


Summary ROC (SROC) Plot


Forest Plot for lnHR
Forest Plot for Sensitivity
Forest Plot for Specificity

Likelihood-based Sensitivity Analysis for Publication Bias

Estimates of sensitivity analysis