. 2011 Oct 26;6(10):e26935. doi: 10.1371/journal.pone.0026935

Table 7. Sensitivity analysis: descriptive statistics for 1,000 simulated outbreaks under different model assumptions and datasets.

	Model 1 – dataset A			Model 3 – dataset A			Model 2 – dataset B
Parameter	median	2.5^th percentile	97.5^th percentile	median	2.5^th percentile	97.5^th percentile	median	2.5^th percentile	97.5^th percentile
Duration (days)	79	72	90	86	42	400	135	100	400
T_peak (days)	43	39	51	47	17	235	41	18	175
I_peak (%)	52.4	51.3	53.4	46.7	1.4	87.2	11.3	1.2	16.2
R_final (%)	99.6	99.5	99.7	99.3	38.4	100	99.5	49.7	100

Model 1 assumed that all 1,000 simulated outbreaks had the same R ₀, κ_L, ρ_L, κ_I and ρ_I, all medians from the posterior distributions. Model 2 assumed all simulations with the same κ_L, ρ_L, κ_I and ρ_I, but with different R ₀. In model 3 all simulations had different κ_L, ρ_L, κ_I, ρ_I, and R ₀. Dataset A includes the results of PCR assay, whereas dataset B includes the results of virus isolation. Duration, duration of the epidemic in days; T_peak, time of the epidemic peak (days after infection); I_peak, peak number of infective birds; R_final, seroprevalence at the end of the outbreak.