Commentary on “On the robustness of latent class models for diagnostic testing with no gold-standard” by Schofield et al.