The Big News Is Big Data
Might the next Nobel Prize in Physiology or Medicine go to a coder, rather than a researcher or doctor?
Mark Hillen |
In May, I attended the annual meetings of both ARVO and the Royal College of Ophthalmologists. And what was the big news at both? Big Data.
As ever, many of the best conversations I’ve had on these topics have been in hotel bars next to the congress venues... So what did I learn?
Electronic medical records (EMRs) are the future. Well-designed ones with lots of data are powerful – and will soon offer real-time information on the safety and efficacy of interventions in your institute and beyond. Of course, most of you dislike filling in the forms, clicking drop-down menus and radio buttons, spending more time typing than talking to patients. Fortunately, all of you people happily using Siri, Cortana and Google Now are making speech-to-text tech pretty awesome. Within 5–10 years, your EMRs could be filled in by the computer listening to the conversation with your patients, with any diagnostic scans being added in automatically. A quick check by the doc and it’s done.
The automated algorithmic analysis of retinal image work is now well-known, and it is going to represent some amazingly helpful decision support and assist with the accurate triage of patients – separating the ‘worried well’ from those truly needing attention from ophthalmologists. Pearse Keane worries about the High Street: optometrists are all now adopting OCT. The amount of image data needing quality analysis will soon explode. It has to be dealt with somehow – and I think we have the answer.
And AI approaches can do even more. I watched Cambridge University’s Peter Thomas present his group’s work on automated eye tracking, pupil and face analysis tech. It was so smart that it could map every muscle visible in the face and track each one’s every movement, so patients undergoing a short eye test need only be filmed to diagnose and assess any number of gaze, pupillary or facial nerve disorders.
What’s most interesting to me is this: a future Nobel prize in Physiology or Medicine could go to an artificial neural network researcher, whose contribution was entirely in silico – someone who has never seen the inside of a medical school, let alone interacted with a patient. And might that person and their team have done more than Lister, Fleming or even Hippocrates to transform medicine?