Diagnostic Test Accuracy Meta-Analysis using Template Model Builder

Fits the hierarchical summary receiver operating characteristic (HSROC) model of Rutter and Gatsonis (2001) , the bivariate binomial-normal model of Reitsma et al. (2005) , the threshold-based bivariate time-to-event model of Hoyer et al. (2018) , and the latent class extensions of Liu et al. (2015) for diagnostic studies with an imperfect reference standard. Provides subgroup analyses, HSROC meta-regression, likelihood-ratio tests, summary ROC plots, and coupled forest plots.

dtametaTMB

Diagnostic Test Accuracy Meta-Analysis using Template Model Builder

dtametaTMB provides a unified framework for frequentist meta-analysis of diagnostic test accuracy (DTA) studies in R. It implements conventional gold-standard models, latent class extensions for imperfect reference standards, subgroup analyses, and multiple-threshold models within a consistent interface.

Key Features

Models

• Reitsma bivariate random-effects model
• Rutter and Gatsonis (HSROC) model
• Hoyer multiple-threshold model
• Latent class Reitsma and HSROC models for studies with an imperfect reference standard

Extensions

• Subgroup analyses
• HSROC meta-regression
• Parameter constraints for sparse-data settings
• Likelihood-ratio tests for nested model comparisons

Output

• Summary ROC (SROC/HSROC) plots
• Coupled forest plots
• RevMan-compatible exports

Implementation

• Frequentist estimation using exact binomial and multinomial likelihoods
• Template Model Builder (TMB) backend
• Unified interface across model families
• Validation against published examples and Cochrane Handbook analyses