Preparation for a potential outbreak of bluetongue virus in Ireland: surveillance design to estimate local prevalence after an initial case detection

TL;DR

This paper outlines a surveillance plan to estimate bluetongue virus prevalence in Ireland after an initial case is detected, using simulations and web tools to support outbreak response.

Contribution

The study introduces a two-stage surveillance design and interactive web tools for simulating bluetongue virus outbreak scenarios in Ireland.

Findings

Sampling 10–11 cattle per herd was sufficient for estimating within-herd prevalence.

Simulated between-herd prevalence estimates varied based on the number of sampled herds and assumed prevalence levels.

Web applications were developed to simulate scenarios and visualize outcomes for different diagnostic test parameters.

Abstract

Bluetongue virus serotype 3 emerged in northern Europe in 2023 and 2024. As of September 2025, Ireland is bluetongue free. However, to inform control decisions in the event of a possible incursion, a surveillance plan to detect cases and estimate prevalence is required. We created an active surveillance plan for 20 km radius temporary control zones (TCZs) after initial case detection. Potential TCZs (n = 1,062) covering Ireland were generated, and surveillance sample sizes were estimated based on cattle data in each TCZ. A two-stage (herd and animal level) design accounted for within-herd clustering. We simulated implementation of the surveillance plan in each TCZ to understand surveillance performance in the Irish cattle population. Within herd prevalence of 30% was used for all simulations, and we assumed the use of RT-PCR with perfect specificity and 99% sensitivity for diagnosis. Based on 100 simulations per TCZ, mean apparent animal level prevalence within tested infected herds ranged from 25 to 32%. With simulated between-herd prevalence of 5%, when between 39 and 61 herds per TCZ were sampled, apparent between-herd prevalence ranged between 3% and 6%. With simulated between-herd prevalence of 30%, and 10–11 herds per TCZ sampled, apparent between-herd prevalence ranged from 20 to 36%. Within herds, sampling 10–11 cattle was sufficient for prevalence estimation. We integrated Shiny and ArcGIS web applications to allow users to simulate scenarios under different settings. These include different test sensitivity and specificity, and different within- and between- herd prevalence contexts. This interface presents infected herds sampled, and true- and false- positives and negatives in a variety of conditions. Evidence from this scenario analysis can be integrated into a multi-pronged early warning and potential follow-up surveillance programme to facilitate decision making in the event of an incursion of BTV into Ireland. The URL for the web application associated with this paper is: https://www.arcgis.com/apps/dashboards/16722dde78d240f4a96303173bc6da2c.

The online version contains supplementary material available at 10.1186/s13620-025-00315-1.

Figures