Graph relates everything – Graph forms the foundation of modern data and analytics with capabilities to enhance and improve user collaboration, machine learning models and explainable AI. Although graph technologies are not new to data and analytics, there has been a shift in the thinking around them as organizations identify an increasing number of use cases. In fact, as many as 50% of Gartner client inquiries around the topic of AI involve a discussion around the use of graph technology.
Healthcare has a Data problem
Today, data rarely exists independently of the applications that use it – it exists in silos. Data is captured to serve the needs of an application and it is hard to access and share that data beyond that application. It is locked away and only those with the technical keys can get to it, internal IT or software vendor! The value of that data would increase IF it was openly available and securely shared as needed.
On average, a patient accumulates around 80Mb of new data every year and this is forecast to grow exponentially
But tomorrow’s Healthcare systems will need to support a lot more patient data. An example is genomic data from genome sequencing and real-world data from personal devices. Locking this data away with historic patient data makes no sense given the research value it promises. We need to make both historic and new patient data easily accessible by clinicians, administrators and researchers. And we need to make it available to AI to help avoid data overload, improve patient outcomes and operating efficiency.
An NHS acute hospital trust may have between 200 and 400 separate IT systems
As Artificial Intelligence (AI) in healthcare slowly becomes a reality graph technologies are the bridge from the world of application data silos to one where data is a shared and valued asset. One where decisions are consistently informed by timely and accurate data delivered at the point of need.
Graphs: From e-commerce and social media to hospital decision support
Graph databases are not new and commonly used by social media networks like Linkedin and e-commerce sites such as Amazon. They power the recommendation engines that suggest new contacts or products from millions of site members or items. And in milliseconds.
Over 80% of all healthcare data is in the form of text and until recently, very hard to access for research or clinical decision support
Hospital systems scope may be small compared to Linkedin or Amazon, but their data are much more complex with images and text common. And they are rapidly evolving as new medical diagnostics and monitoring are generating new data streams, often in real-time. And those systems face a tsunami of new data sources, including ‘omics’ and lifestyle data from our personal wearable devices.
Existing hospital patient record systems (or Electronic Health Records) and database technologies cannot meet this challenge. Graphs are ideal to capture and model this data in a way that both humans and machines can use to inform decisions. And as a common foundation for a new generation of explainable AI designed to augment clinicians and improve their performance.
Features of Graphs
Graphs are visual and intuitive. This means they can capture and visualise knowledge about a health system as it really is; as things (nodes) and with relationships between those nodes. They are visual so much easier to develop and evolve, unlike an SQL table in a database! And graphs can now be generated automatically from existing data sources using Machine Learning (ML). And graphs can quickly be extended with new data as needed, Covid-19 being a great example of such a need.
Graphs can help unleash a dormant NHS asset – its Data
Graphs, along with the cloud, modern data stacks and AI, are foundational technology pillars of a digital NHS. They can improve patient outcomes, increase productivity, and meet growing societal expectations of the NHS in the 21st century. Together these technologies can unleash the value of a much under-valued dormant NHS asset – its data.