Preventing Disease

Riddles with Disease

At the Zoo

What do you do if you are a young epidemiologist and your boss tells you to solve a complicated disease problem?  It will take 50 years to solve it, but you only have 24 months. Say “hello” to computer simulation technology.

Complicated Disease Problem

In 1991, our veterinary pathologists identified an intestinal infection in a few of the hoofed mammals at the Safari Park. This bacterial infection called Johne’s (pronounced YO-nees) disease is common and well studied in dairy cattle, but less is known about effectively managing this disease in exotic species.

Because the bacteria can spread from animal-to-animal, we need to maintain a level of disease surveillance and management at all times to ensure that the healthy animals stay healthy.  This involves testing hundreds of animals to detect only one or two cases per year, a costly endeavor that has never been able to fully eradicate the disease. The complicated problem: how do we improve surveillance?

Finding a Better Way to Manage Disease

Support from the Ellen Browning Scripps Foundation allowed us to construct a disease simulation model to identify which scenarios could reduce (or eradicate!) this infection in our animals.

The advantage of a simulation model is that we can experiment with different management and surveillance techniques in a virtual world (in silico) and make an informed decision on improving our testing strategy in real life. 

The model not only simulates disease transmission and intervention strategies, but can also be programmed to track cost, total number of tests, and how well we are identifying infected animals. 

Computer simulation also allows us to test extreme conditions.  For example, we can test each animal 100 times a year or choose not to test at all. While both strategies are impractical in the real world, the beauty of computer simulation is that it doesn’t matter…we can explore, explore, explore! 

Check Out the Cool Model

Click here to watch a video of our running model.

On the left-hand side of the screen you will see a group of 500 virtual animals that are infected (red and orange) or healthy (gray).

As time progresses, the infected animals can spread disease to the healthy animals based on the rules of the underlying model, which were established using scientific literature, our own data, and the occasional expert opinion. 

As more time passes, the graphical output on the right shows the number and percent of animals that become infected (“latent”) and those that are already infected and can spread it to others (“infectious”). 

You will see animals appear and disappear, because, like in reality, our virtual animals have babies (appear) and grow old (disappear). Watch the counts of “incidence” (the number of newly acquired infections) and “proportion detected” (ranges from 0 to 1 and is an indicator of how well the current testing plan is doing – 1 means the testing strategy found all of the infected animals). Notice the change in cost of testing over time.

What the Model Tells Us

By changing the input parameters we can modify the predicted outcome and identify the best testing strategy. While we still have many other options to test, so far the model suggests that testing a fraction of the population will allow us to maintain our current level of surveillance while greatly reducing our testing costs. Such computer simulation technology is helping us do a better job caring for our animals and tackle seemingly impossible problems.
You can read more about our use of disease modeling to improve animal health in the Fall 2010 Conservation Update.

Carmel Witte, Scientist, San Diego Zoo Global