Estimation of sensitivity and flock-sensitivity of pooled faecal culture for Mycobacterium avium subsp. paratuberculosis in sheep.

Pooled faecal culture (PFC) is a widely used test in ovine Johne's disease (OJD) control programmes in Australia but information about its characteristics is limited. We conducted this study to estimate sensitivity and flock-sensitivity of PFC in sheep with different OJD histopathological lesions in simulated flocks with a range of infection prevalence levels. Initially, a known quantity of faeces from sheep with pauci- or multibacillary lesions was pooled with uncontaminated faeces from confirmed non-infected sheep and cultured using PFC technique. PFC sensitivity, calculated as a proportion of the pools of a particular size that tested positive, was determined to be 90% in sheep with the multibacillary form of the disease but varied with pool size in sheep with the paucibacillary form of OJD. Subsequently, probabilistic models were developed to estimate overall pool-sensitivity achieved in a flock (Sek) and flock-sensitivity of PFC (FSe) in various simulated scenarios. In flocks with a given ratio of multi- to paucibacillary sheep and with low to moderate infection prevalence level, Sek decreased with increase in pool size, but increased with pool size in flocks with >or=10% prevalence. FSe, in contrast, increased with pool size in all the tested scenarios. Both Sek and FSe increased with infection prevalence, ratio of multi- to paucibacillary sheep and the number of pools sourced from flocks. Sensitivity analyses indicated that the estimates and trends were robust to moderate changes in input parameters. The results suggest that the current testing of seven pools of size 50 is adequate for most scenarios, however, for very low prevalence flocks, a gain in FSe can be made by increasing the number of pools tested and a higher Sek can be achieved by reducing pool sizes.