Moorfields and DeepMind Health research partnership latest update

In 2016 Moorfields Eye Hospital NHS Foundation Trust and DeepMind Health started a partnership to explore how cutting edge technologies such as machine learning and artificial intelligence (AI) can help medical research into eye diseases, including age-related macular degeneration and sight loss as a result of diabetes.

The partnership has brought together leading NHS eye health professionals with some of the UK’s top technologists at DeepMind. Together, they have used thousands of historic de-personalised eye scans to successfully train computer systems to identify signs of eye disease and recommend how patients should be referred for care. 

In August 2018 we announced ground-breaking results of the first stage of the partnership. The results showed that the AI system could match world-leading experts in diagnosing a range of eye conditions, recommending the correct referral decision for over 50 eye diseases with 94% accuracy.

The next step is to build on these initial results to continue with the overall goal of exploring whether AI technology could achieve real patient benefit through the detection, diagnosis and prediction of sight threatening conditions. 

As part of the next step, the data used to train the AI systems will be moved to Google’s cloud computing infrastructure, with initial access in the UK and USA, but it may one day include cloud facilities globally. This is one of the most powerful and secure computer systems in the world. 

Traditionally, computers process data locally, using their own systems. The cloud though is a network of connected computer systems that work together to process data. Recent guidance published by NHS Digital enables data to be stored on the cloud and the NHS is using cloud systems more and more because they are often faster and more secure than traditional fixed systems. You can read more about this guidance on the NHS Digital website.

 Why is the data being moved? 

This research requires a great deal of processing power. Cloud based systems can offer this as well as higher levels of safety and reliability than fixed systems. This will allow the research to be carried out to a higher standard and at a faster pace than what is possible with current systems available in the UK. 

All the data being moved over to the cloud is completely depersonalised and encrypted. That means that we cannot identify Moorfields patients from the data. Furthermore, the only people who will be able to access the data are the researchers and there are strict legal and regulatory approvals in place overseeing this.

Last updated: 5th November 2018