FIRST-INF: Estimating the date of first infection by highly pathogenic avian influenza in a poultry flock

A tool to optimise contact tracing for avian influenza in poultry flocks

Background and objective

Epidemics of highly pathogenic avian influenza (HPAI) viruses are causing severe socio-economic damages in domestic poultry. Contact tracing is commonly used to fight HPAI, among other control measures. It aims at detecting early and controlling potentially infected farms that had contacts with outbreak farms. Because it is very resource-intensive, contact tracing is usually performed on a pre-defined time window, defined as 21 days for HPAI. However, pre-defined time windows may not be adequate for all farms, reducing the efficiency of contact tracing. Estimating the probable range of date of first infection in a farm would help tracing activities, by enabling official veterinarians to focus on a precise and focused list of relevant contacts.

Therefore, the Host-Pathogens Interactions unit (ENVT-INRAE) and Wageningen Bioveterinary Research (WUR) have developed the FIRST-INF web-based interactive application to help official veterinarians to optimise contact tracing for HPAI, by estimating the probable range of date of first infection in a poultry flock (within a poultry house). To do so, FIRST-INF uses a mechanistic model simulating the transmission of HPAI in a poultry flock. Model parameters, including the date of first infection, are estimated by fitting the model to observed daily mortality data.

Use of the application

The application is made of three different tabs. You are currently on the first tab (“Home”). You can start using the application by clicking on the “Enter your data” button at the bottom of this page, which will bring you to the second tab: “Data entry”.

On the Data entry tab, you are invited to enter the required data, i.e. the poultry species, the initial flock size, the date at which the flock was introduced in the poultry house, and the date of the last observation (last available daily mortality data). Based on the latter, a table is generated to enter the daily mortality data. The table spans between the date of the last observation and the date of flock introduction, with an upper limit of 5 weeks before the last observation. At this point, you need to enter the daily mortality data for as many days as possible (by default the value is “-1”, corresponding to missing values). After doing so, you can click on the “Confirm your data” button, which will bring you to the third and last tab: “Results”.

A “Start simulations” button is displayed at the top of the Results page. When clicked, model simulations are performed to estimate the time of first infection. Based on the estimates, recommendations for optimising contact tracing are formulated, except when the estimation is suspected to have failed, in which case a warning is displayed. A figure (see example below) is produced that compares the model results (50% of simulations in the dark blue area and 95% of simulations in the light blue area) with the observed data (black dots).

Warning

Once you click on the “Start simulations” button, you should see three blue bars moving. After a few minutes, a loading bar should appear on the bottom right corner of the screen, indicating that simulations are underway. After a few more minutes, the results and the figure should appear. If the screen becomes grey at any point, it means that something went wrong and you should then reload the app. In the eventuality where it still would not work, please contact us.

Contact

This App has been developed by Lisa Fourtune (ENVT-INRAE). The model and estimation algorithm have been developed by ENVT-INRAE (Sébastien Lambert, Lisa Fourtune, Julie Baca and Timothée Vergne) and WUR (Peter Hobbelen, Jose Gonzales, Armin Elbers). This work was performed in the framework of the “Chair for Avian Health and Biosecurity”, hosted by the National Veterinary College of Toulouse and funded by the Direction Générale de l’Alimentation, Ministère de l’Agriculture et de l’Alimentation, France. The Dutch contribution to this study was supported by The Netherlands’ Ministry of Agriculture, Nature and Food Quality (Veterinary Epidemiology and Risk Analysis project WOT-01-002-042). For more information and for technical support, please contact: timothee.vergne@envt.fr