Ian A. Gardner

Epidemiologic issues in the validation of veterinary diagnostic tests

In this review, we critically discuss the objectives, methods and limitations of different approaches for the validation of diagnostic tests. We show (based on published data and our own experiences) that estimates for the diagnostic sensitivity and specificity may vary among populations and/or subpopulations of animals, conditional on the distribution of influential covariates. Additional variability in those parameter estimates may be attributable to the sampling strategy. The uncertainty about diagnostic parameters is of concern for the decision-maker in the context of clinical diagnosis or quantitative risk assessment as well as for the epidemiologist who uses test data for prevalence estimation or risk-factor studies. Examples for the calculation of diagnostic parameters are presented together with bias-avoidance strategies. We suggest guidelines for an epidemiologic approach to test validation of veterinary diagnostic tests.

Coastal freshwater runoff is a risk factor for Toxoplasma gondii infection of southern sea otters (Enhydra lutris nereis)

The association among anthropogenic environmental disturbance, pathogen pollution and the emergence of infectious diseases in wildlife has been postulated, but not always well supported by epidemiologic data. Specific evidence of coastal contamination of the marine ecosystem with the zoonotic protozoan parasite, Toxoplasma gondii, and extensive infection of southern sea otters (Enhydra lutris nereis) along the California coast was documented by this study. To investigate the extent of exposure and factors contributing to the apparent emergence of T. gondii in southern sea otters, we compiled environmental, demographic and serological data from 223 live and dead sea otters examined between 1997 and 2001. The T. gondii seroprevalence was 42% (49/116) for live otters, and 62% (66/107) for dead otters. Demographic and environmental data were examined for associations with T. gondii seropositivity, with the ultimate goal of identifying spatial clusters and demographic and environmental risk factors for T. gondii infection. Spatial analysis revealed clusters of T. gondii-seropositive sea otters at two locations along the coast, and one site with lower than expected T. gondii seroprevalence. Risk factors that were positively associated with T. gondii seropositivity in logistic regression analysis included male gender, older age and otters sampled from the Morro Bay region of California. Most importantly, otters sampled near areas of maximal freshwater runoff were approximately three times more likely to be seropositive to T. gondii than otters sampled in areas of low flow. No association was found between seropositivity to T. gondii and human population density or exposure to sewage. This study provides evidence implicating land-based surface runoff as a source of T. gondii infection for marine mammals, specifically sea otters, and provides a convincing illustration of pathogen pollution in the marine ecosystem.

Conditional dependence between tests affects the diagnosis and surveillance of animal diseases

Dependence between the sensitivities or specificities of pairs of tests affects the sensitivity and specificity of tests when used in combination. Compared with values expected if tests are conditionally independent, a positive dependence in test sensitivity reduces the sensitivity of parallel test interpretation and a positive dependence in test specificity reduces the specificity of serial interpretation. We calculate conditional covariances as a measure of dependence between binary tests and show their relationship to kappa (a chance-corrected measure of test agreement). We use published data for toxoplasmosis and brucellosis in swine, and Johnes disease in cattle to illustrate calculation methods and to indicate the likely magnitude of the dependence between serologic tests used for diagnosis and surveillance of animal diseases.

Application of diagnostic tests in veterinary epidemiologic studies

Important applications of diagnostic tests in non-clinical areas of veterinary medicine include surveillance, monitoring or screening for disease, prevalence estimation, and risk-factor studies. In this review, we address the processing and analysis of diagnostic data with special emphasis on serological data from enzyme-linked immunosorbent assays (ELISAs). We describe methods that are used to adjust for misclassification (imperfect sensitivity and specificity) in the different applications of diagnostic tests. These methods often have limitations attributable to the uncertainty of sensitivity and specificity estimates. We review descriptive and analytical statistical methods that allow adequate presentation of non-clinical diagnostic data.

Herd-level interpretation of test results for epidemiologic studies of animal diseases

Correct classification of the true status of herds is an important component of epidemiologic studies and animal disease-control programs. We review theoretical aspects of herd-level testing through consideration of test performance (herd-level sensitivity, specificity and predictive values), the factors affecting these estimates, and available software for calculations. We present new aspects and considerations concerning the effect of precision and bias in estimation of individual-test performance on herd-test performance and suggest methods (pooled testing, targeted sampling of subpopulations with higher prevalence, and use of combinations of tests) to improve herd-level sensitivity when the expected within-herd prevalence is low.

Correlation-adjusted estimation of sensitivity and specificity of two diagnostic tests

Models for multiple-test screening data generally require the assumption that the tests are independent conditional on disease state. This assumption may be unreasonable, especially when the biological basis of the tests is the same. We propose a model that allows for correlation between two diagnostic test results. Since models that incorporate test correlation involve more parameters than can be estimated with the available data, posterior inferences will depend more heavily on prior distributions, even with large sample sizes. If we have reasonably accurate information about one of the two screening tests (perhaps the standard currently used test) or the prevalences of the populations tested, accurate inferences about all the parameters, including the test correlation, are possible. We present a model for evaluating dependent diagnostic tests and analyse real and simulated data sets. Our analysis shows that, when the tests are correlated, a model that assumes conditional independence can perform very poorly. We recommend that, if the tests are only moderately accurate and measure the same biological responses, researchers use the dependence model for their analyses. Copyright 2003 Royal Statistical Society.

Molecular and bioassay-based detection of Toxoplasma gondii oocyst uptake by mussels (Mytilus galloprovincialis)

Toxoplasma gondii is associated with morbidity and mortality in a variety of marine mammals, including fatal meningoencephalitis in the southern sea otter (Enhydra lutris nereis). The source(s) of T. gondii infection and routes of transmission in the marine environment are unknown. We hypothesise that filter-feeding marine bivalve shellfish serve as paratenic hosts by assimilation and concentration of infective T. gondii oocysts and their subsequent predation by southern sea otters is a source of infection for these animals. We developed a TaqMan PCR assay for detection of T. gondii ssrRNA and evaluated its usefulness for the detection of T. gondii in experimentally exposed mussels (Mytilus galloprovincialis) under laboratory conditions. Toxoplasma gondii-specific ssrRNA was detected in mussels as long as 21 days post-exposure to T. gondii oocysts. Parasite ssrRNA was most often detected in digestive gland homogenate (31 of 35, i.e. 89%) compared with haemolymph or gill homogenates. Parasite infectivity was confirmed using a mouse bioassay. Infections were detected in mice inoculated with any one of the mussel sample preparations (haemolymph, gill, or digestive gland), but only digestive gland samples remained bioassay-positive for at least 3 days post-exposure. For each time point, the total proportion of mice inoculated with each of the different tissues from T. gondii-exposed mussels was similar to the proportion of exposed mussels from the same treatment groups that were positive via TaqMan PCR. The TaqMan PCR assay described here is now being tested in field sampling of free-living invertebrate prey species from high-risk coastal locations where T. gondii infections are prevalent in southern sea otters.

Comparison of methods for estimation of individual-level prevalence based on pooled samples.

We review frequentist and Bayesian approaches for estimating animal-level disease prevalence using pooled samples obtained by simple random sampling. We determine the preferred approach for different prevalence scenarios and with varying knowledge about sensitivity and specificity values. When sensitivity and specificity are perfect or known, we can choose between the large-sample theory estimates and the one-to-one relationship exact estimates. When sensitivity and specificity are unknown, we must use large-sample theory estimates or Bayesian methodology (which gives exact estimates). However, when the large-sample theory produces a negative lower confidence limit, we must use one of the exact methods. We compare estimates from each approach using culture results from pools of 20 eggs from three flocks on a California ranch that were producing eggs that were contaminated with Salmonella enteritidis phage type 4.

Salmonella enteritidis, phage type 4 infection in a commercial layer flock in southern California: bacteriologic and epidemiologic findings.

Salmonella enteritidis, phage type 4 (SE PT4), was isolated from five of six 27-wk-old layer chickens submitted for necropsy from a flock of 43,000. Bacteriologic and epidemiologic investigations on the ranch revealed that five of the eight flocks (n = 176,000) were infected. The prevalence of SE PT4 in randomly selected healthy birds ranged from 1.7% (in caged birds) to 50% (in free-range birds) and prevalence in culled birds (kept on dirt floor houses) ranged from 14% to 42%. The estimated overall prevalence of group D Salmonella in eggs contaminated with group D Salmonella was 2.28 per 10,000. The estimated prevalence of group D Salmonella in eggs from caged birds in three infected houses ranged from 1.5 to 4.1 per 10,000, whereas in two houses of free-range birds, prevalence was 14.9 to 19.1 per 10,000. Three of the eight flocks on the ranch remained negative for Salmonella between May 1994 and December 1995 or until removed from the ranch. Salmonella enteritidis PT4 was also isolated from 12.5% (6 of 48) of mice; 57% (four of seven) of cats; and two of two skunks tested. Environmental drag swabs and well water samples yielded multiple serotypes of Salmonella (23/180 and 5/14, respectively) but not S. enteritidis.

Range expansion, population sizes, and management of wild pigs in California

The introduction and spread of nonnative organisms to new regious can disrupt ecosystems by causing declines or extinctions of native species. Widely mtroduced wild pigs (Sus scrofa ) have contributed to declines and extinctions of numerons species on oceanic islands and can have pronounced negative ecological effects on mainland areas when population densities are high. Although range expansion by introduced wild pigs has ceased in many regions of the United States, it has increased significantly since the 1950s in California. Our analyses of data from annual hunter survevs and mapped locations of hunter-killed wild pigs shows that the distribution of wild pigs increased from around 10 coastal connties in the early 1960s to parts of 49 of Californias 58 counties by 1996. An index to density based on locations of hunter-killed wild pigs plotted in a Geographic Information System (GIS) indicated that within the 79,550 km 2 (25%) of the total land area of the state now occupied by wild pigs populations are most abundant in the central and north-coast regions. By stratifying each county into 1 of 3 relative abundance classes and assigning density values based on research at multiple sites, we estimated tbere were around 133.106 (range = 106,485-159,727) wild pigs in California in 1996. The recent increase in the range of wild pigs in California was facilitated by a combination of multiple hunting-related introductions, deliberate releases of domestic pigs, and, possibly, increased forage availability associated with agricultural development. Natural range expansion also has occured, however, and the dynamics of the spread of wild pigs in California appear typical of invasions observed among other organisms. Forage and water availability are important factors influencing the dist ribution and abundance of wild pigs in California. and predation may be important to an unknown extent. Up to 40% of Californias wild pigs are removed from the population annually hich may control populations in some areas but not others. Wild pigs may be cansing ecological damage in some coastal regions where population densities are very high, and detailed studies are needed in those areas to help mitigate potential problems with this increasingly widespread mammal in California.