Diagnosis of tuberculosis in groups of badgers: an exploration of the impact of trapping efficiency, infection prevalence and the use of multiple tests

Buzdugan, S N and Chambers, M A and Delahay, R J and Drewe, J A (2016) Diagnosis of tuberculosis in groups of badgers: an exploration of the impact of trapping efficiency, infection prevalence and the use of multiple tests. EPIDEMIOLOGY AND INFECTION, 144 (8). pp. 1717-1727.

[img]
Preview
Text
10168.pdf - Accepted Version
Available under License Creative Commons Attribution Non-commercial.

Download (1MB) | Preview

Abstract

Accurate detection of infection with Mycobacterium bovis in live badgers would enable targeted tuberculosis control. Practical challenges in sampling wild badger populations mean that diagnosis of infection at the group (rather than the individual) level is attractive. We modelled data spanning 7 years containing over 2000 sampling events from a population of wild badgers in southwest England to quantify the ability to correctly identify the infection status of badgers at the group level. We explored the effects of variations in: (1) trapping efficiency; (2) prevalence of M. bovis; (3) using three diagnostic tests singly and in combination with one another; and (4) the number of badgers required to test positive in order to classify groups as infected. No single test was able to reliably identify infected badger groups if <90% of the animals were sampled (given an infection prevalence of 20% and group size of 15 badgers). However, the parallel use of two tests enabled an infected group to be correctly identified when only 50% of the animals were tested and a threshold of two positive badgers was used. Levels of trapping efficiency observed in previous field studies appear to be sufficient to usefully employ a combination of two existing diagnostic tests, or others of similar or greater accuracy, to identify infected badger groups without the need to capture all individuals. To improve on this, we suggest that any new diagnostic test for badgers would ideally need to be >80% sensitive, at least 94% specific, and able to be performed rapidly in the field.