Doctor data and the future of AI in healthcare

Jack Watts, head of Business Development AI EMEA, NetApp, writes about the growing influence of AI, and what’s needed for its success.

In today’s digitally connected world, data is everything. This is especially true within the medical field, which arguably demonstrates the best examples of how data is changing the world for the better. Though the industry is still at the forefront in its ability to use that data with emerging technologies, like artificial intelligence, the room for growth cannot be understated. According to recent reports, the industry is expected to reach $6.6 billion by 2021 and almost double to $13 billion by 2025.

Due to the sensitivity around personal health records, there are strict controls on how companies and organisations can use this data. While personally identifiable data is in no case allowed to be shared for marketing or insurance purposes, it’s often used day to day in conjunction with artificial intelligence to assist with diagnostics, clinical support, preventative medicine and new drug discovery. In an attempt to ensure that patient data is only used for research purposed, the UK Health Data Research Alliance was recently created. Working to unite expertise and act as a “clearing house” for best practices in the guidance of the UK’s health data, the alliance ultimately enables faster, more efficient access to research in these fields.

Another route some organisations have been moving towards is the idea of federated learning, the process of training a high-quality centralised model with data distributed across various servers. Because the centralised server aggregates the data, each local data set is kept secure within the organisational infrastructure. This allows institutes like hospitals and universities to develop sophisticated models without having to directly share sensitive clinical data – ultimately facilitating innovative research methods between the two.

AI, data, silo complexities

Research around data and AI within the medical field has significantly progressed in recent years, so much so that there have been multiple use cases – many of which have already started to majorly influence the wider industry. AI’s been able to reduce diagnosis errors in breast cancer patients by 85% and has helped 61% of heart patients avoid invasive angiograms – reducing treatment costs by more than a quarter (26%). And in 2019, the UK government announced that £250 million in public funds would be rerouted for the National Health Service (NHS) to set up an artificial intelligence lab that will work to expand the use of AI technologies within the service.

The success of the artificial intelligence is invaluable when coupled with the right data sets. It can help identify patterns, enable accurate autonomous systems, and develop predictive insights. Unfortunately, in order to come to fruition, these projects must have the flexibility to move without the limitations of where data exists. But as more than three-quarters of organisations still claim to have internal data silos – it’s a huge hurdle to consider.

Having traditional and separate infrastructures – or silos – results in bottle necks which can substantially extend the time needed to complete each cycle. This complexity not only reduces the productivity of the pipeline, but lowers general business value as well. To alleviate this, organisations must adopt a true Data Fabric strategy, one that considers the journey towards AI implementation. A Data Fabric strategy weaves the entire deep learning pipeline together, enabling more iterations to be completed in the same amount of time – leading to software that can work more effectively and efficiently.

To some, the importance of this type of data pipeline may not seem like a priority, but it’s especially essential when dealing with a consistently growing volume of data. A comprehensive pipeline can help organisations capture, prepare, access, move and protect large volumes of data from multiple sources, which is particular useful in the medical industry as patient data is considered especially complex. In the end, a Data Fabric strategy can ensure that the organisation is future proofed, enabling them to support any combination of on-premise and cloud, as well as offering complete flexibility.

Though major progress has been made with data driven AI in the medical industry, there’s still a way to go. Not all organisations have the advanced infrastructure to support the increasing amount of data and its complexities. To support this technological evolution, organisations within the sector must start to put a focus breaking down silos and moving models from prototype to production. If they can, they not only have the opportunity for efficiency and flexibility, they have the chance to save lives.



About


'Doctor data and the future of AI in healthcare' has no comments

Be the first to comment this post!

Would you like to share your thoughts?

Your email address will not be published.

This site uses Akismet to reduce spam. Learn how your comment data is processed.

© 2019 Rapid Life Sciences Ltd, a Rapid News Group Company. All Rights Reserved.

Privacy policy

Terms and conditions