Natalia Ciria

SVEPM 2024 Poster

Download pdf

Modular risk analysis to assess complex risk pathways: An application to farm biosecurity

Natalia Ciria1, Giovanna Ciaravino1, Alberto Allepuz1
1Universitat Autònoma de Barcelona, Spain

1. Background

Quantitative risk analysis, using Monte Carlo simulations, is a method to numerically determine the probability of a risk event occurring taking into account uncertainty. It considers all possible values that each variable could take and the probability of their occurrence.

robability that a herd is disease free h_free = (1 - h_prev )/(1 - h_prev · plan_sensi) Probability that a herd is infected h_inf = 1 - h_free Probability that an animal in a herd is infected a_inf = h_inf · w_prev Probability that an infected animal is infectious a_infectious = a_inf · infectious_inf

For example: To calculate the probability of buying an infectious animal from a farm, we use surveillance estimates that indicate the herd prevalence of BVD in the farm region is between 0.64 and 0.57, with a mean of 0.55.

2. Problem

Detailed risk quantification in complex systems, such as the different pathways by which pathogens enter a farm, involves numerous steps and parameters that make the model unwieldy and difficult to analyse, explain and visualise.

3. Methods

Modular risk analysis helps to analyse complex systems by looking at parts of a system separately and then combining the results for an overall assessment. Our modular risk analysis approach consists of mathematical expressions grouped into modules.

Expressions: Set of mathematical functions. Core expressions calculate probabilities, link expressions connect core expressions with different inputs and between modules.

Modules: Set of expressions and input data that constitute a self-contained part of the system.

Core expressions: Origin, Test, Direct Contact, Indirect Contact, Zone Effect, Time and Totals. Link expressions: Farm, Unknown farm, Transport, Quarantine, Fattening, Visit, Neighbour and Wildlife.
Modules: Farm, Unknown farm, Transport, Quarantine, Fattening, Visits, Neighbour farms, Wildlife

4. Implementation in R

Based on the mc2d package, we are developing a new package to automate the creation of stochastic variables, simplify the building of modular models and visualise the model and its results.

R script: transport<-eval_module( model_expression = c(transport_link_expression, dir_contact_expression, indir_contact_expression), prev_mcmodule = c(farm, unk_farm) data = purchase_data)

5. Next steps

We are testing this model on real cattle farms to assess the probability of pathogen introduction (IBR, BVD and TB) and the impact of biosecurity measures, and will extend it to other species and pathogens.
We plan to publish the model’s framework as an R package that could be applied to other types of risk analysis beyond biosecurity.

Contact: natalia.ciria@uab.cat

BIOSECURE
UAB (Universitat Autònoma de Barcelona)

This research project was funded by BIOSECURE Horizon Europe project (www.biosecure.eu)