Government and healthcare providers around the world are well aware of the potential of patient data to revolutionise treatment. Indeed, a report on the recent National Health Service (NHS) Health and Care Data Day highlights the £10 billion commercial value placed on the NHS’s patient records — a figure that reflects the huge potential that data holds to transform the health of current and future generations.

And while artificial Intelligence (AI) is transforming the speed and accuracy at which this data can be turned into actionable insights, the ability to share and process data is still being held back by a fragmented IT landscape and the demands of market regulators. In parallel, public concerns over how the data is used are coming to the fore. A December 2019 article highlights that 74 per cent of the UK public is concerned about how their personal data is shared, and only 28 per cent are confident that current laws and regulations will protect them.

The healthcare sector has seen spiralling volumes of data in recent years. Medical imaging equipment has become much more powerful, quicker to operate and much more widely available. At the same time, this has often created multiple data silos. To enable clinicians to make use of a single data set, the several sources of data need to be integrated, and many healthcare providers have chosen to do this with an electronic health record (EHR) software suite. For example, the Royal Papworth Hospital in Cambridgeshire, United Kingdom, has built its entire digitisation strategy around its EHR platform, which for completeness was also integrated with a different EHR used in a neighbouring campus. While this ensures easy access to authorised users, the integration process can be difficult and lengthy.

In parallel, there has been an explosion in wellness data recorded by citizens’ personal health-monitoring devices, whether worn for fitness tracking or to manage a medical condition. Most of these personal devices are already network-connected, but to date there has been little integration between medical systems and consumer wellness trackers. However, Fitbits have already helped to predict which post-operative cancer patients are likely to require re-admission, and as health trackers evolve, there is no doubt they will become valuable diagnostic aides.

Automation has proved essential for making efficient use of these huge volumes of healthcare data. Big data analytics technologies were originally developed by search engine providers trying to make sense of huge volumes of internet data. Today, big data analytics solutions are helping make sense of huge volumes of healthcare data — or, as Professor John Danesh of the British Heart Foundation (BHF), puts it, “We can now measure haystacks of biology in large numbers of people, which we hope will help us find the needles responsible for disease.”

At the same time, citizens are understandably sensitive about how and with whom their personal medical records are shared. The medical sector actually benefits from a high level of trust relative to other sectors. A recent survey by YouGov and the Open Data Institute found that the NHS is the only organisation that most UK citizens trust with their personal data. But since healthcare providers turn to partners for analytics and machine learning (ML) platforms, this only highlights the need for effective de-identification.

This is the process which ensures that identifiable patient data is accessible only to authorised users. Beyond reassuring patients, this process is also a regulatory requirement for sharing data in many geographies — in the United States as part of the HIPAA Privacy Rule, and for GDPR compliance when relating to EU citizens. However, as the IEEE has shown, this is a nontrivial exercise for which healthcare professionals will often seek specialist support.

Alongside big data approaches, developments in ML and AI are providing increasing support to clinicians. Much of the current focus is on using AI for image analysis, supporting increased use of digital imaging capabilities. As the Imperial College of London (ICL)’s Professor Daniel Rueckert said,  “AI has tremendous potential in radiology and medical imaging.” However, the possibilities of AI in healthcare go beyond image analysis, valuable as this is. Oxford University researchers have recently used AI  to develop a biomarker that helps predict heart-attack victims, and there is huge interest in developing further predictive capabilities.

In summary, while citizens continue to have legitimate concerns about medical data privacy, the potential for big data and AI in healthcare is undeniable and unstoppable — as witnessed by the recent decision to invest £250 million in the UK’s NHS for AI alone.

Interested in learning more about data-driven care? Read how University Hospitals of Morecambe Bay NHS Foundation Trust is leveraging patient data to deliver better health outcomes.