I am currently in the process of running a diagnostic accuracy meta-analysis of sensitivity and specificity values. I am using the R program meta4diag, which uses a Bayesian model. As part of this model, Guo and Riebler (2015; citation below) suggest setting the penalized complexity (PC) priors to µ = 3 and α = 0.05. This would create a contrast of P (σ > 3) = 0.05. Guo and Riebler note that this corresponds to believing “that the sensitivities or specificities lie in the interval [0.5, 0.95] with probability 0.95” (p. 5). Now, my question is how did they determine that the PC prior parameter P (σ > 3) = 0.05 would correspond to this interval? I tried looking at their citations but could not determine how they estimated this confidence interval. I would like to know how to do it so I can estimate the confidence intervals for other PC priors. Any help would be greatly appreciated.

Guo, J., & Riebler, A. (2015). meta4diag: Bayesian bivariate meta-analysis of diagnostic test studies for routine practice. arXiv preprint arXiv:1512.06220.

  • $\begingroup$ You should contact the authors. $\endgroup$ – Xi'an Jan 3 '19 at 16:39
  • $\begingroup$ You could try the mailing list stat.ethz.ch/mailman/listinfo/r-sig-meta-analysis/ although I do not know whether anyone there uses Bayesian models for diagnostic test meta-analysis $\endgroup$ – mdewey Jan 3 '19 at 16:50

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.