The heart of data science

Pizza + Data = Inspiration

Over the last year, data science and pizza have helped bridge the gap between the big data and medical communities. “I suppose we could have gone for ‘Data plus Beer’, but beer and hospitals are not traditionally seen as a great combination,” jokes Jeroen Maas, Challenge Lead for Health at the Amsterdam Economic Board, partner in the organisation of the evening.

Kicking off the new academic year, the 12th Medical Data plus Pizza meet-up at the Amsterdam UMC hospital on 17 September once again welcomed a full house – the group as a whole has now grown to around 1,200 members.

The rise of applying big data to a clinical setting is part of a larger trend. Amsterdam is quickly becoming a global hub for the development of these new technologies.

Earlier on this day as part of ‘Prinsjesdag’, King Willem-Alexander outlined the key policies that the Dutch government will focus on in the coming political year. The minister of finance, Wopke Hoekstra, also announced the setting up of a fund, worth as much as a billion euros, that seeks the further development of AI, robotics and big data.

“This fund will not be for fun things,” the minister said. “But rather for initiatives that will have a positive impact on Dutch society”. A few eyes lit up in the crowd.

Predicting outcomes for coma patients

The evening’s first presentation, ‘Deep Learning for Outcome Prediction in Postanoxic Coma’, was by Marleen Tjepkema. She is a technical physician of neurology at the University of Twente. Her team was acclaimed earlier this year for the release of a research paper that outlined how they trained an AI algorithm to read brain wave electroencephalograms (EEGs) and predict outcomes for cardiac arrest patients who fall into a coma.

With brain damage occurring within three to five minutes, half of such patients will have a poor outcome – ranging from death to a lifetime of full-time care. Hence, an accurate prediction of neurological outcome is essential to make a balanced decision on whether treatment is worthwhile.

One current method to predict outcome involves using trained specialists to analyse EEG traces for specific features. However, since this analysis should occur in the first 12 to 24 hours, these trained analysts are not always available. Hence: AI to the rescue.

Tjepkema’s team trained deep convolutional neural networks (CNNs) to analyse EEGs. The CNNs can then reproduce the abilities of human specialists. However, since this analysis occurs in a “black box”, the actual reasoning behind why the AI flags certain patients as potentially treatable remains unknown. Currently, the team is now involving neurophysiologists, computer scientists and mathematicians to make the results more transparent. This will allow a better understanding of the brainwave activities of comatose patients.

A smart partner for your heart

The numbers add up… In Europe, 85 million people live with cardiovascular disease. Forty per cent are over age 65 – and 32% of these live alone. In other words: there are 11 million elderly people with CVD who have no support – even though most heart attacks and strokes, which both require a speedy reaction to increase the chance of survival, happen outside of a hospital.

Hence there is a real need for a “smart partner for your heart”, according to Wavy co-founder Daryl Autar during his presentation ‘Wearables for Health’.

“Using smartwatches and AI, Wavy learns and monitors the personal heart health of vulnerable groups in real-time. With the help of seamlessly integrated smart home speakers, Wavy is even capable of intervening in case of emergencies. People can’t always be there to look after your loved ones, but Wavy can.”

The small Wavy team is moving fast. Just one year ago they came up with their idea during the TechCrunch Disrupt San Francisco Hackathon. It won the event’s top prize. Grants and support from big-name accelerators quickly followed. Next month, they will begin clinical trials with 100 patients – a remarkable achievement.

Autar mentions many nifty Wavy features, including: “We can use sentiment analysis to link recorded patient observations such as ‘I feel awful’ or ‘I just ate five pizzas’ to other data points. When taken together these might warrant an intervention.”

Indeed, pizza is not always good for you. But it certainly has helped to bring data scientists and doctors together in Amsterdam.

Please join our Amsterdam Medical Data Science group if you want to be on board!

About Amsterdam Medical Data Science

The Medical Data plus Pizza meeting aims to bridge the gap between health professionals and data scientists by bringing both together in an informal setting for presentations and pizza. At Amsterdam UMC data scientists, medical professionals and researchers discuss how AI (Artificial Intelligence) and medical data can be used to benefit patients and improve medical practices. Since it’s launch in August 2018 it has steadily grown in popularity with more and more people attending to hear the latest news and innovations in medical data and spark new collaborations.

For more reports from previous Medical Data plus Pizza Meet-ups, click here.

The Amsterdam Medical Data Science Group meetings are supported by The Right Data Right Now consortium, which includes Amsterdam UMC, OLVG, Vrije Universiteit, Pacmed, and the Amsterdam Economic Board.

Copy: Edenfrost by Steve Korver