A large disparity in the seroprevalence results betweenthe two HI tests was demonstrated in the field sera.Overall, significantly (P<0.001) more HI antibodypositiveducks were detected with HI-HRBC than HICRBC,but no sera that were positive in the HI-CRBCwere negative in the HI-HRBC test. This indicates thatthe type of HI test used for assessing field sera has to be carefully chosen as the test performance will influencethe reported measures of disease frequency, such asprevalence or incidence, for the population evaluated. Incontrast a discrepancy between the proportions of positivesidentified with the HI tests was not evident in theexperimental study. Compared to the experimentalstudy, the H5N1-exposure status of ducks in the fieldstudy was unknown. Therefore, it is possible that someof the birds that tested negative in the HI-CRBC testwere in the early response of H5N1 infection, which asthe experimental study showed, is more likely to bedetected with the HI-HRBC test than HI-CRBC.The interpretation of the diagnostic test results shouldbe conducted cautiously, particularly if it is performed toassess an agreement between tests using samples fromdifferent populations. The seroprevalence of ducks inthe field was relatively low (8% or 2% using HI-HRBCand HI-CRBC, respectively) in contrast to the overallseroprevalence of 57-60% obtained in the experimentalstudy. Factors related to the stage and severity of diseaseand the immune status vary within a population (andcould be different in low- and high-prevalence populations)[43]. Therefore the distribution of individualssampled at different stages in the infection process willinfluence the prevalence estimated. On the other hand,it has been highlighted that the statistical relationshipbetween kappa and prevalence (and bias) is multifacetedand complex [33,44,45]; two tests will have a higherkappa value if the prevalence is moderate (e.g. 50%) ratherthan very low or very high [33]. Therefore the unadjustedkappa statistic was higher for the experimental study compared to the field study. In fact, it has beenrecommended by some statisticians to concentrate onpopulations with prevalence near 50% when comparing
การแปล กรุณารอสักครู่..
