Table 3 Parameter estimates of the best-fit models for H5N1 and H5N6.

From: Optimising the detectability of H5N1 and H5N6 highly pathogenic avian influenza viruses in Vietnamese live-bird markets

 

H5N1

H5N6

Point estimate

95% CrI

Point estimate

95% CrI

Se duck sample

0.17

0.06–0.86

0.66

0.27–0.98

Sp duck sample

0.99

0.99–1.00

0.98

0.98–0.99

Se solid waste

0.19

0.05–0.59

0.49

0.27–0.81

Sp solid waste

1

0.99–1.00

1

0.99–1.00

Se liquid waste

0.21

0.06–0.53

0.38

0.20–0.60

Sp liquid waste

0.99

0.97–1.00

0.99

0.98–1.00

Se drinking water

0.25

0.04–0.63

0.23

0.08–0.46

Sp drinking water

1

0.99–1.00

0.99

0.98–1.00

Se faeces

0.22

0.04–0.64

0.14

0.03–0.35

Sp faeces

1

0.98–1.00

0.99

0.98–1.00

True prevalence

0.05

0.01–0.13

0.08

0.04–0.13

Apparent prevalence

0.13

0.09–0.18

0.20

0.15–0.25

CovN

—

—

0.003

0.000–0.015

DIC

58.5

61.2

  1. Estimated sensitivity (Se) and specificity (Sp) refer to the individual (pooled-by-five) samples of the different sample types; the prevalence refers to the proportion of contaminated live-bird markets (LBMs); 95%CrI is the 95% credible interval of the posterior distribution of the parameters, as estimated by the best-fit models; CovN is the estimated covariance between liquid waste and drinking water samples in non-contaminated LBMs; DIC stands for deviance information criterion. Note that the two selected models were those associated with the smallest DIC.
Back to article page