Table 2.
True prevalence results (mean and quantile of the 95% credible interval (Q) of the toy example with 24% and 5% apparent prevalence, 500 sample size, 120/25 positive tests with consistent assumptions on sensitivity (80%) and specificity (90%). Speybroek et al. 2012 Model 1 assumes fixed values for Se and Sp, while Model 2 assumes Se∼Unif(0.7,0.9) and Sp∼Unif(0.85,0.95). For the |tPRiors| example, Se∼Beta(195.76,48.94) and Sp∼Beta(135.86,15.1), while for Setup A: πt∼Beta(1,1) and Setup B: πt∼Beta(1.3,1.3)
| Speybroek et al. 2012 | tPRiors | |||||||
|---|---|---|---|---|---|---|---|---|
| Model 1 | Model 2 | Setup A. | Setup B. | |||||
| Measure | 24% | 5% | 24% | 5% | 24% | 5% | 24% | 5% |
| Mean | 0.2015 | 0.0048 | 0.2008 | 0.0096 | 0.2004 | 0.0103 | 0.2026 | 0.0127 |
| 2.5% Q | 0.1493 | 0.0001 | 0.1133 | 0.0003 | 0.1266 | 0.0003 | 0.1203 | 0.0008 |
| 97.5% Q | 0.2577 | 0.0173 | 0.2927 | 0.0336 | 0.2705 | 0.0327 | 0.2764 | 0.0367 |