Polish scientists have created a platform based on artificial intelligence that analyzes the need for blood transfusions. The hybrid system uses neural networks and thus allows you to accurately determine how much blood a given hospital should order. Algorithms improve the management of transfusions by hospitals and help reduce the scale of blood waste. The algorithm is not only informed about historical data, but thanks to the integration of electronic health records the hospital is also able to adjust its forecasts in real time.
Blood has a very limited amount of time it can be held. In Poland, it is 42 days. Until now, there have been no advanced statistical tools available to predict how blood will be distributed and how it should be ordered each day. This is provided by artificial intelligence, machine learning, i.e. statistical models that allow you to predict the future, find correlations in historical data about how blood was used in a given blood bank, and thus predict how to best approach today’s inventory – explains Mateusz Pawełczuk, co-founder of AIDA Diagnostics.
The AIDA Blood: Hospitals system uses genetic algorithms to achieve a high level of forecast accuracy. It constantly collects new data from the hospital in which it is installed, reaching its full potential after a month of use. The algorithm is not only informed about historical data, but thanks to the integration of electronic health records the hospital is also able to adjust its forecasts in real time. This allows teams responsible for blood storage and management to make decisions based on data gathered in one place.