Using Hidden Markov Models for Animal Disease Surveillance
Author Information
Author(s): Colin Robertson, Kate Sawford, Walimunige S. N. Gunawardana, Trisalyn A. Nelson, Farouk Nathoo, Craig Stephen
Primary Institution: Wilfrid Laurier University
Hypothesis
Can hidden Markov models effectively analyze veterinary data for early detection of zoonotic diseases?
Conclusion
The study demonstrates that hidden Markov modeling is a useful approach for analyzing veterinary surveillance data to detect unusual disease patterns.
Supporting Evidence
- The study analyzed 5758 submissions reporting animal health issues.
- Hidden Markov models were used to account for variability in disease reporting.
- The model identified significant differences in disease patterns across regions.
Takeaway
This study shows that by using special math models, we can better understand animal health data to spot diseases early, which can help keep both animals and people safe.
Methodology
Data were collected from field veterinarians in Sri Lanka using mobile phone surveys, and analyzed using hidden Markov models to identify patterns in disease prevalence.
Potential Biases
Underreporting of severe diseases by farmers could skew data towards common diseases.
Limitations
Potential selection bias due to reliance on farmers to report cases and exclusion of data from private veterinary clinics.
Participant Demographics
Field veterinary surgeons from four administrative districts in Sri Lanka.
Digital Object Identifier (DOI)
Want to read the original?
Access the complete publication on the publisher's website