Europe is looking to revolutionize healthcare with a massive project that would make millions of patient records available to artificial intelligence algorithms.
The multi-billion-dollar European Health Data and Evidence Network project will standardize 100 million patient records and permit A.I. algorithms to tap them, with the goal of improving all aspects of the medical system. Applications open Sep. 1 for grants to participate in the project, which formally launched in November.
The system will boost the sharing and analysis of health data with its network of compatible data sources. They will comply with Europe’s General Data Protection Regulation and other European privacy laws. The network recently announced its first open data partner call.
“Every day, patients are suffering or dying because we are not able to tap into the transformational potential of real-world data,” according to EHDEN, a five-year public-private partnership under the $3.7 billion Innovative Medicines Initiative of the European Commission.
Participants range from universities such as Oxford, pharmaceutical companies including Sanofi and Pfizer to tech firms like EMC.
The project is designed to overcome European privacy sensitivities around medical data, and to close the gap between Europe and the rest of the world in the fast-growing digital health sector.
“In Europe we have brilliant minds in our academic universities but have created a lot of knowledge that has gone nowhere,” Pierre Meulien, executive director of the Innovative Medicines initiative, told a conference in Brussels last year. “We have to find new models where innovation can be created and translated into use.”
A report by Silicon Valley Bank showed investments in digital health in Europe in 2017 and 2018 were less than 10% of similar investments in the U.S. The number of investments and the average size of early-stage rounds in Europe fell between 2017 and 2018.
A.I. needs large, diverse sets of data to train algorithms and cut inconsistencies and errors, says Nigel Hughes, a coordinator of the EHDEN project and a scientific director at Janssen Clinical Innovation in Belgium.
EHDEN will build a “federated network for open science research in Europe over the coming five years” by mapping datasets from hospitals, private health networks and others, says Hughes.
EHDEN will convert detailed patient data from hospitals, primary healthcare providers and other bodies using the Observational Medical Outcomes Partnership Common Data Model, a data tool that enables information to be anonymized and harmonized so it can be shared and analyzed by remote researchers.
The original data remains with its owner but through the tool it will provide the basic information required to train AI algorithms, allowing companies, academics and other researchers to run queries on records from multiple sources and platforms to develop digital health solutions.
“Hospital data units and academic research teams are becoming increasingly aware of the scientific potential of the data they hold,” Vasa Curcin, senior lecturer in health informatics at King’s College, London University, said.
“Until now, the sharing of patient data has been held back by privacy concerns that threatened to leave Europe behind as the U.S. and China raced to develop digital health platforms based on artificial intelligence-fueled algorithms,” he said.
The project may help European competitiveness at a time when it risks falling behind, experts said. The continent’s regulatory bodies have focused on privacy, possibly at the expense of competitiveness, they said.
“If Europe doesn’t act at speed to change the course of its regulation and policy from protecting to driving competitiveness we will miss the innovation opportunity that AI brings and will not be able to create global European champions in AI,” says Loubna Bouarfa, CEO of Okra Technologies in Cambridge, England.
The EU’s caution causes “a problem in terms of competitiveness” when healthcare AI entrepreneurs choose where to locate their companies, says Daniel Castro, vice president of the Information Technology and Innovation Foundation in Washington.
“The EU is naïve to think that it can become the global leader in AI by focusing on ethics. In a race, the fastest car wins, not the most ethical,” Castro says. Opportunities to improve healthcare will “remain out of reach if the data is locked away. Data protection rules like GDPR can make it difficult to collect and aggregate, even when it is being used for purposes like medical research.”
