Researchers at Purdue University's School of Veterinary Medicine, one of 32 veterinary colleges in North America, are using a national pet database to monitor how diseases in dogs and cats could serve as early warning indicators of bioterrorist activities and other health concerns for humans.

Utilising the database, temporal and spatial analyses are performed with solutions from SAS, leaders in business and analytical intelligence, to identify unexpected clusters of disease. Unusual findings are further explored to determine causality.

The goal is to create an integrated surveillance system that allows veterinarians and government officials to log on and see which diseases are occurring where and whether disease rates are increasing or decreasing.

According to researchers, dogs and cats are highly sensitive to four of the six Class A biological agents which can be a threat to national security: anthrax, botulism, plague and tularemia.

With the system, researchers can monitor certain Class A diseases in pets, and because the incidence of such diseases is so rare in animals, they could quickly identify an increase in occurrence and take steps to reduce or prevent harm to humans. Researchers plan to use the system to measure zoonoses (diseases that are transmissible between animals and humans) and other health outcomes, like adverse drug-related reactions or adverse vaccine reactions, as well as to monitor chronic diseases and other epidemics in pets.

The avian influenza epidemic that has spread to large gaming pets like leopards and tigers is an example of how the surveillance system could be used to help monitor and detect potential disease outbreaks and health threats in humans.

Data to populate the system is being gathered in collaboration with Banfield Pet Hospitals, the largest provider of veterinary care in the US.

Banfield's medical records are uploaded weekly to a central data warehouse. The research team uses various SAS tools, to load, merge and extract the data they need for specific analyses. SAS Enterprise Miner is used to model the risk factors associated with pets diagnosed with diseases like cancer, as well as the risk factors for flea and tick exposure. Disease frequencies are tracked using SAS Enterprise Guide.

"In the past, we used SAS to analyse very small data sets - 50 or 100 cases - but we're now looking at 10 million cases, which is totally different," says Dr. Nita Glickman, research coordinator and computer analyst at Purdue.

"We really didn't consider working with anyone other than SAS because there is nothing else around with the capability of doing what we needed to do. I think we've established strong with SAS, and they've been extremely helpful in getting our project moving."