3. Results
Descriptive statistics of the predicted number of infected birds, probability a market is infected, and prevalence at the market for the three scenarios are shown in Table 2 and Fig. 2. Also shown are the test statistics and p-values for the Kruskal–Wallis tests for significant differences in predictions for each scenario. The mean probability a LBM is infected was high (0.78) and this was similar between scenarios. The estimated median prevalence of H5N1 at a LBMwas 0.005 (0–0.025, 95%uncertainty interval), and decreased when Inspection 1 and 2 were adopted. Prevalence was 1.2 and 2.4 times lower than the median when Inspection 1 and 2 were adopted respectively. Consequently, LBM were less likely to be infected
and less contaminated when Inspection 2 was implemented.