Table 3.
Probability posterior estimates for the prevalence of enteropathogenic Yersinia in serum at pig and herd level. Sensitivity and specificity for the diagnostic test are also shown. Estimates with different set of priors are also presented, as part of sensitivity analyses.
| Sample | Parameter | Posterior estimates, median (95% probability interval) | ||
|---|---|---|---|---|
| Original model | Set of priors 1 | Set of priors 2 | ||
| Serum | Sensitivitya | 0.795 (0.736–0.848) | 0.802 (0.744–0.855) | 0.919 (0.833–0.990) |
| Specificitya | 0.969 (0.853–0.999) | 0.961 (0.826–0.998) | 0.978 (0.894–0.999) | |
| Herd prevalence slaughter-age pigs | 0.776 (0.522–0.937) | 0.782 (0.528–0.943) | 0.777 (0.522–0.936) | |
| Herd prevalence sows | 0.868 (0.625–0.981) | 0.868 (0.628–0.979) | 0.869 (0.634–0.978) | |
| Pig prevalence slaughter-age pigs | 0.675 (0.632–0.709) | 0.645 (0.572–0.709) | 0.654 (0.613–0.689) | |
| Pig prevalence sows | 0.740 (0.573–0.824) | 0.749 (0.586–0.829) | 0.720 (0.578–0.803) | |
aEstimates calculated considering all pigs, that is, slaughtered-age pigs and sows.
Herd prevalence and pig prevalence are based on the estimates of tau and Tp, respectively.
The following are other priors for sensitivity analysis.
Set of priors 1: using different prior distributions for pig prevalence.
Set of priors 2: using noninformative prior distributions beta(1, 1) for pig prevalence and for sensitivity beta.