AI can be of great benefit to a stretched healthcare system, ranging from tailoring treatment to individual patients, to making more secure diagnoses, and by quickly training staff in new competencies. AI and digitalization have increased rapidly during covid-19 and many new working methods and habits may continue even after the pandemic. This is the opinion of the international AI expert Göran Lindsjö: “When we look back at the pandemic, we will realise that it has given a real boost to both AI and digitalization”.
During the spring of 2020, the coronavirus pandemic has meant that much of what we previously knew has of course changed. What strategies have been successful to better cope with the crisis and what is important to think about when society is restarted when the crisis is over? Future by Lund has begun an interview series within our network that looks for Future by Lund’s new normal. We aim to draw attention to the expertise that is close to us so that we can jointly prepare for the future. Do you know someone that has knowledge and perspectives that may be of interest for us to share? Please send suggestions for people to interview to email@example.com or firstname.lastname@example.org
Göran Lindsjö is an international AI expert with a background at the Swedish Ministries of Finance and Enterprise and Innovation. He has worked with AI for 35 years, having ran several companies in the field, and now resides in the USA. This interview was conducted by Future by Lund’s business advisor Lars Mattiasson.
The pandemic has placed a great strain on healthcare. What are the possibilities and dangers of using AI to make improvements to healthcare?
- There are huge possibilities to use AI in healthcare and, as I see it, these are greater than in any other industry. The benefit is not least about better diagnoses. It has been found that humans and machines work best together rather than individually. But also, that AI can be used to develop treatment methods that are more personalized, thus offering more specific treatment methods based on a particular patient.
- Another important benefit of using AI is to train staff new skills faster. By using AI there are as many opportunities as possible. In parallel, we need to be vigilant about the disadvantages that AI can bring to the healthcare industry. One danger could be that the technology can use “prejudicial” data that can come about when, for example, the technology has only been trained to investigate heart attacks on older men. Another form of danger relates to integrity, an issue which has existed prior to AI, but with a higher risk due to AI because it is dependent on large data sources. Of course, we can try to anonymize the data, but it is difficult to know when it is successful and if the data being used is not traceable after all.
- I think nevertheless that the danger for Sweden right now is bigger, so we have not sufficiently exploited the opportunities with AI. In the Örebro region they have taken a good initiative and want to use AI for many different tasks. This can by extension lead to shorter waiting queues, better diagnoses, and lower taxes. They have then looked at what has been done around the world. Otherwise, we have a tendency in Swedish healthcare and AI to completely want to devise this ourselves. Some regions want to start from the beginning and work using their own most advanced staff. This means that Sweden participates and contributes in different ways. However, a report from the National Board of Health and Welfare released this autumn states that we find it difficult to develop this further in Sweden. We start many projects within AI, but something makes it slow to begin using the results properly.
The AI projects that have been initiated during the crisis – do you think they will remain afterwards?
- I think many of the initiatives that are being utilized now in relation to both digitalization and AI within healthcare and education will remain, while elsewhere it will almost fall back to where things were before. But when we look back at the pandemic after it is over, we will see that it has given a real boost both to AI and digitalization. The combination of healthcare and AI is complicated, and thus it is not as easy to evaluate any increase, compared to the number of digital meetings that has radically changed working life.
- The healthcare system has discovered that, for example, John Hopkins University has been a forerunner in AI and are the source most often used by Sweden to get reasonably accurate data on what is happening in Sweden and in its vicinity. In this way we can piggy-back on the big AI-users.
Do you see any interesting projects that combine healthcare and AI in Sweden?
-There are many sharp projects with an international connection. One such project is in Linköping where they are far ahead in image processing, working together with Stanford for example. Region Halland is also quite progressive in several ways and have collaborated with the best. It is not that activity is missing in some way, but there is even more that is ready that we could use right now.
In Lund, the innovation system is collaborating across the border during the crisis. Do you notice a similar trend to break down verticals even in the USA?
-Those who have come quite far have tried to do this, but it is difficult to break down verticals everywhere. It becomes very software orientated, so you often apply it in a specific context. It is difficult even within a large healthcare organization to draw on common lessons, which is a shame because there is much to share and develop within platforms, tools, methodology and knowledge. In fact, lessons can be reused from, for example, pattern recognition in the banking, insurance, and manufacturing industries. Compared to that it should then be even easier to reuse skills within the same healthcare organization. It is often the case that people are thinking of the final application and are used to thinking that one clinic is totally different from another clinic and that both differ even more from the finance department at the hospital. Whereas there can be so much in common where one could benefit from the other. Those who have come the furthest, for example the Cleveland Clinic and Mayo Clinic, have succeeded in developing a horizontality within the different parts of the organization. During the crisis these are also the healthcare organizations that have been given more and more responsibility even outside their usual catchment area.
It is easy to see how the crisis has changed our lives on a personal level – both in terms of travel and meetings. Is there anything special you have reflected on about this?
- There have been so many changes in our personal lives that we have experienced over the past two months and of course it has affected our values. One thing I have reflected about is that we have had physical distancing, but perhaps we should not talk so much about social distancing – the crisis may even mean that we have contact with more people because we did not Zoom previously but now we are already used to it and have additional time to spend at home. This means that we have now added more people that we have conversations with.
Translation: Ben Dohrmann