Threshold Modeling
The threshold model avoids choosing a single threshold per study or per test. The latent transformed test result is assumed to follow a logistic distribution. The model uses a parametric logistic relationship between threshold and the prob…
1 sources - 4 claims
The threshold model avoids choosing a single threshold per study or per test. The latent transformed test result is assumed to follow a logistic distribution. The model uses a parametric logistic relationship between threshold and the probability of testing positive within each disease class. Reparameterization around the most commonly reported threshold makes the model compatible with standard bivariate DTA meta-analysis at that threshold.