Apparent Prevalence of Fish Pathogens in Asymptomatic Salmonid Populations and Its Effect on Misclassifying Population Infection Status

Abstract

A mail survey of Canadian fish diagnosticians was conducted to determine how frequently samples of asymptomatic fish yielded an estimated apparent prevalence (ap) of greater than zero but less than 2% or 5% for infectious pancreatic necrosis virus (IPNV), infectious hematopoietic necrosis virus (IHNV), Aeromonas salmonicida, or Ceratomyxa shasta. The relationship of host-specific factors to the frequency of samples with low ap (≤5%) was also investigated. Fourteen percent of IPNV-positive samples and 26% of A. salmonicida-positive samples had an ap of less than 5%, with Atlantic salmon and rainbow trout, respectively, having the highest frequency of samples with low ap. Forty-seven percent of IHNV-positive samples and 35% of C. shasta-positive samples had ap values of less than 5%; no factors were significantly associated with the ap of either pathogen. The results are used to discuss the probability of misclassifying the infection status of a population of fish. This probability is influenced by the sample size and the apparent prevalence, which is a function of the true prevalence and the sensitivity and specificity of the diagnostic test. Formulae for estimating the probability of correct classification of a population's infection status are reviewed.