By Michael Reakes, CEO PKS
Nothing seems to inspire a headline in health quite like Artificial Intelligence (AI). Ever since IBM Watson moved from the Jeopardy studios to hospitals and research centres, the industry and world have been abuzz with excited optimism about the revolutionary impact AI would have on diagnosis, treatment and prevention of a myriad of diseases.
One of the flagships of this brave new intelligent world was MD Anderson, the cancer centre within the University of Texas who announced in 2013 that they would be using IBM Watson to help eradicate cancer.
This was huge. Finally we were living in the future where technology could help solve some of the most pervasive and devastating diseases on the planet; doing what thousands of doctors, researchers and medical professionals have spent decades working to solve. AI was not only living up to the hype, it was exceeding it.
Fast forward three years and there is sadly different headlines that appear. In a recent Forbes article it appears the IBM and MD Anderson partnership is falling apart, with the illustrious Watson project having been placed on hold since 2016, costing the University millions of dollars and failing to meet its goals.
So what does this mean?
When your flagship site fails to deliver, you have to imagine this will raise questions around the value of AI and whether it is ready to move from think-tank to the real world. The reality is that AI and machine learning can and do have a direct and obvious impact on the day-to-day treatment and diagnosis of patients.
Whilst we don’t know precisely what’s gone on at MD Anderson, we shouldn’t lose faith in what AI and machine learning can deliver. What we can take away from this experience is there are no free passes – in an already overburdened health system, the days are numbered for any technology (including AI or machine learning) that can’t demonstrate material value to current patients in a timely manner.
What we’ve repeatedly seen as critical success factors, and what resonates with our clients, is when our CDS technology is used alongside the client’s existing clinical experts, processes and systems to support, amplify and scale their immense and inherent value. Whilst always admirable in their intent, data-based discovery projects that don’t prioritise existing people, process or systems and take a long time before any value arrives at the patient bedside, ultimately attract the wrong type of attention.
So we should in no way lose heart about AI, but instead we should recognise that there are different types and applications of machine learning and AI, and in all cases the focus should be on enhancing the patient outcome as quickly as possible.