Efficient contact tracing

AI can help navigate an unprecedented pandemic

During the Covid-19 pandemic, Systematic and the Danish Patient Safety Authorities developed a solution that contributed to streamlining contact tracing. The solution is now further developed with a combination of AI capabilities and machine learning technologies that help ensure better data quality in less time.

AI-driven solution

Initially, the Danish Patient Safety Authorities needed insights from their collected data to help navigate and prioritise their efforts. Due to the urgent nature of the situation, critical decisions needed to be taken based on insights, which made data quality a top concern. In collaboration the Danish Patient Safety Authorities, Systematic developed "Pandemic Control". The solution supports the prioritisation of the counselling of infected and the quality of the data collected.

The self-service solution is integrated with relevant systems and offers citizens the opportunity to enter data about their presumed location of infection transmission into a free text field. At first, localisation data was handled manually, but now Systematic has integrated a machine learning algorithm into the solution, which makes the algorithm able to handle data. The algorithm is named Clusters.

Faster qualification of data

AI handles time-consuming task

With the use of Natural Language Processing (NLP), which is based on artificial intelligence (AI), the Clusters algorithm can match the citizens’ entries with known locations. If for instance three citizens state that they have been infected at “Dokk1”, “Aarhus public libari” and “public library in Aarhus”, the algorithm is able to collect these entries under the same address, despite the fact that the text fields show different answers. 

In addition, Clusters also collects the answers that are not immediately identified and matched with a known location. Contact tracing staff can then review them and try to tie the answers to an address. With Clusters, staff can focus on qualifying data and advising relevant authorities, instead of spending hours on reviewing data.

The Clusters algorithm has made it possible for the Danish Patient Safety Authorities to re-prioritise resources and optimise a very time-consuming task. As more data is qualified, the algorithm is trained and learns from examples, which in time makes it better and better at matching and grouping locations. AI can therefore turn out to be a valuable tool in the fight against Corona and future epidemics or pandemics.