Tick-borne diseases: spatio-temporal modeling of infection risk


Mining industry gives tools for accurate disease prediction.

Maps of Lyme disease infection risk for Poland. Credits: S. Moliński/Data Lions
Maps of Lyme disease infection risk for Poland. Credits: S. Moliński/Data Lions

Lyme disease is a serious problem in Central Europe. We observe na increasing trend in a number of cases. Actually 20 000 cases per year are registered in Poland.

Lyme borreliosis is a vector-borne disease which means that for successful transmission to the human it must be carried by other species. The main vector of Lyme disease in Europe is a hard-bodied tick Ixodes ricinus. Environmental conditions two decades back have a big impact on the population of this arthropod. Long winters with temperatures below -10 degrees were not favorable for ticks. Their larvae and eggs are frozen below this temperature threshold. But those winters are just history. Due to the climate change Polish winters become mild and temperatures below -10 degrees are very rare. Moreover, spring starts earlier. This is the recipe for a catastrophe: more ticks survive winters and they start their feeding activity very fast, in March. Risk of infection increases. That’s why Lyme disease is an example of climate-driven infection which will be a problem in the warmer Europe of tomorrow.

We are aware of this fact, that’s why we have created a set of Artificial Intelligence algorithms to predict Lyme disease risk at a local scale. We use remotly sensed data, meteorological measurements and sophisticated tools used by the mining industry, then we tune them to solve public health problems.

Fortunately, the problem of miners where they have only a few drills and must estimate where is the best place to run a full-scale excavation is similar to the problem of disease risk estimation, where we also have only a few data points and we must predict where the risk is greatest. Specific tuning of geostatistic algorithms done by our engineers provides us tool for accurate estimation of the risk.

Those tools may be easily transferred to the other diseases where environmental factor plays a crucial role (i.e. cancers). That’s why we won’t stop and we will utilize sophisticated machine learning tools for good of all of us.

If you want to find more just write to us: Contact.

(Project is realized for the European Space Agency).