Sweden’s new AI hub enters the stage

The Swedish AI sector is starting to pick up speed, and the strategic hub AI Sweden is in full swing with its work. Now, the first sharp implementations in healthcare are also here.

AI SWEDEN IS based in Lindholmen Science Park in Gothenburg. A name change (from AI Innovation of Sweden) took place in the autumn of 2020, which at the same time meant a significant broadening of the work field. With funding from Vinnova, AI Sweden now has over 100 partners.

Henrik Ahlén has a long background as a consultant in digital care and e-health, and he has often expressed frustration that the healthcare sector has not progressed faster in digitalization issues. As newly appointed AI Change Agent Healthcare at AI Sweden, he gets the opportunity to work with healthcare issues at the absolute forefront of tech development.

At the moment, he is busy with the project Information-Driven Care and Health, where he surveys the leading players in Sweden, in order to spread good examples further. Partners in the project include Swedish regions, hospitals, universities, municipalities, as well as large and small companies. Among the regions is Region Halland, which contributes with exciting AI applications from its operations, says Henrik Ahlén.

– Halland is at the forefront of AI. They have implemented the technology in clinical care and have succeeded in resolving certain legal issues. Among other things, they have developed algorithms that measure a patient’s care costs throughout the system, so that a basis for decisions can be obtained and help with optimizing processes. It is very useful for many caregivers.

To predict how long a patient lives after their emergency visit, mortality predictions have also been developed at hospital emergency departments. Together with AI Sweden, Halland’s AI model for this will be packaged together with The Karolinska University Hospital to see how the findings can be scaled up nationally.

Decentralized AI unties regulatory knots

One challenge for artificial intelligence in healthcare is that large amounts of patient data are needed. This data is often collected at a single hospital or region and can usually not be shared with other regions for legal reasons.

One solution to the problem with disseminated data and patient integrity is called Edge AI. AI Sweden has built a testbed in Gothenburg called Edge Lab, where it is possible to work with the technology.

– Machine learning usually requires you to upload large amounts of data to a common server where the algorithms can work. With Edge AI, you can instead send out AI algorithms to the local data sources to run calculations and analyses on-site.

Then the insights from the locally trained AI models – not the data – can be transferred to the central database without risking any patient data. This also heavily reduces the amount of data that is shared, which reduces costs. This procedure is still at an early stage so it needs to be tested. But it has great potential, and not just in the healthcare sector, says Henrik Ahlén.

Are you more hopeful about the future of health care now that you see what is happening behind the scenes? – Yes, because now I am in an environment where we focus on the opportunities and want to contribute to the distribution of good ideas. Of course, we also have lawyers on site who keep track of the legal obstacles that arise. AI is not a technology in the first place, but a transformation of business development. Therefore, the issue must be raised in the management teams, and we run specialized podcasts and training on the topic “AI for CxOs”.

Henrik Ahlén points out that healthcare processes are largely about logistics, which makes it excellent for optimization with the help of machine learning and AI.

– Everything that can be streamlined, automated, or quality enhanced benefits from AI. It’s a lot about raising the utilization rate of the availableresources, and the healthcare sector has a lot to learn from other industries. The ten largest companies on the US stock market are heavily involved in AI, regardless of whether it is Amazon in e-commerce or mobility companies that develop self-driving vehicles.

“Everything that can be streamlined, automated, or improved in quality benefits from AI. It’s a lot about raising the utilization rate of the available resources, and the healthcare sector has a lot to learn from other industries.” - Henrik Ahlén

Henrik Ahlén

By allowing the algorithms to analyze patient data locally, tough regulations can be evaded.