Resource Library

Highlights and Key Takeaways from the Digital Health Networks Exchange Event on AI, Data, and Analytics

Written by CereCore | Apr 26, 2024 11:27:58 AM

AI is the buzzword of our times, occupying the spotlight of almost every digital health event we attend, and with good reason. AI represents a leap forward in technology, with potentially profound impacts on healthcare. With this, two CereCore International colleagues, Stephen Ashmead and James Walsh, attended the Digital Health Networks Exchange event Unlocking the transformative power of AI, data, and analytics in digital health in Birmingham at the Altera offices.

Here are our 5 key takeaways.

Diagnostics leads the way

Diagnostics, covering a range of data-rich, self-contained sciences, is ahead of the game when it comes to incorporating AI. But simple improvements in data integration can also lead to dramatic changes. Professor Kevin Moore, Chief Medical Officer at Salutare, presented a 'digital blood form', a single phlebotomy interface where orders can be tracked through every step of the process, improving patient experience and eradicating pinch points such as the need for labelling samples more than once.

Data, data, data…

It's not just about having access to key patient data but also about the quality of the data itself. Artificial intelligence can only be as good as the information it is provided with. It was clear throughout the day that we need to work harder – and smarter – at this crucial step.

That's just as true when looking at data about AI products themselves. Dr Clare McGenity, reporting on a meta-analysis of AI digital pathology tools, found that 99% of studies on AI products had a high or uncertain risk of bias or applicability issues. The ENSURE-AI project is developing a framework for addressing these gaps.

AI will never be perfect.

The real world is full of ambiguity, bias and complexity. While well-designed AI tools can help reduce health inequity, they will never be perfect. But what happens when AI gets it wrong? We have processes for dealing with clinical errors, but we're only starting to get to grips with the potential for AI errors. And what standard should we hold AI to? Are we setting an impossible bar, much higher than what we demand of human clinicians? Having clear equity and governance frameworks can help navigate these challenges – information governance should be seen as a friend, not a foe, in deploying AI products.

Education is key

When rolling out technology, we need to ensure staff at all levels have the knowledge they need to use these new solutions. Beatrix Fletcher, Programme Manager at Guy's & St Thomas' Trust, recognises that clinicians are the most prominent advocates for patient-focused care, and we should be providing structured opportunities for them to learn about AI.

But we also need patients to be on-side. By focusing on patient outcomes, we can look past the hype as well as distrust surrounding AI. An honest discussion of what AI can and can't do can engage patients in AI and data analytics.

Start somewhere

The message that resounded throughout the day was to be courageous and start somewhere. It was recognised that not every organisation is prepared to incorporate AI in clinical decision-making, but there are many fewer daunting uses for AI, such as supporting back-office functions, that can significantly improve patient care.

If you're a healthcare organisation looking to begin or are already on your AI and data analytics journey, CereCore International can help support you by providing data and technology expertise, including EPR and software deployment and optimisation, maturity assessments, and data infrastructure.