Why Dr. Rob Brisk sees artificial intelligence as the ultimate medical recruitment tool



Twenty years ago, I had just hit upon the idea of applying to medical school. I was asking any clinician I could find about their experience of the profession. Would they do it all again if they had the chance? I spent several weeks shadowing my grandparents’ GP: a lovely, semi-retired doctor whose own medical journey began a few years after the birth of the NHS. I took a job as a ward clerk on a paediatric unit to get some first-hand experience.

Here’s what I learned:

The cost of the clinical life was substantial: physically, intellectually, emotionally. But the payoffs were huge. There was the sense of belonging to an enormous family. The knowledge that you were part of something much bigger than yourself. The untainted high of self-sacrifice in the service of strangers. I was sold.

Over the following years, few of my more romantic notions of medical life survived the brutal encounters with flatus-tube-shaped reality. I worried a lot about money. I saw too little of my family. I got so sleep-deprived that I regularly forgot my PIN code. But somehow it was always worth it. The feeling of curling up in bed at 6am, too tired to even take off your scrubs but knowing your patient would soon be waking up to another day because of your team’s great work – that was more valuable than any material reward. (And, perhaps, the sanity of my wife…)

Recently, though, the balance seems to have shifted. The nature of the job has changed. We spend less time focusing on the patient in front of us, more time worrying about bed shortages and ever-growing waiting lists. Service provision trumps training every time. Worst of all, constant demands for greater productivity from a workforce that already goes far beyond the call of duty have squandered our most valuable asset: goodwill. I’ve even worked in hospitals where there has been a crackdown on staff using NHS-provided teabags and instant coffee, never mind the staggering value of unpaid overtime they plough back into the system each year. The resentment on many wards is almost palpable.

The world has changed too. After decades of booming Western economies, the prospect of a secure, lifelong job holds less appeal than it once did. By making it easier to connect with family and friends, the internet has made physical distance less daunting. Young clinicians are readier to relocate to other health systems where they feel more valued. And for those who are really fed up, a wealth of online resources can help them retrain and jump the clinical ship altogether. More than a few times, I’ve heard senior doctors and nurses – usually frustrated that junior staff are unwilling to give their free time to plug yet more rota gaps – talk about the ‘snowflake generation’. That’s horribly lazy thinking. Human nature hasn’t changed a jot in the last few decades. Junior clinicians are just facing a different cost / benefit analysis. The sooner we accept that, the sooner we stop moaning and start to redress the balance.

That’s where AI comes in.

It’s going to take many years to grow the workforce enough to meaningfully spread the load. It may take longer still to restore morale. In the meantime, though, AI applications could take the sting out of some of the worst parts of the job. Like filling in a mountain of request forms after each round and clinic. Or trawling through endless case notes to write the discharge letter at the end of a long admission. Or being called to review a patient in the middle of the night because some arbitrary threshold has been reached on a one-size-fits all risk score, even though the patient is perfectly stable and the nurse isn’t worried.

We’re still waiting to see which frontline application of clinical AI is going to be the snowball that triggers the avalanche. Computer vision technologies developed by other industries, plus a strong focus on single-step problem solving, have certainly given radiology a head-start. Pathology, dermatology and ophthalmology are up there too. But the earliest examples of any new technology are always the most expensive, and these fields don’t necessarily offer the most compelling business cases. Cut down the reporting time on CT scans by a few hours and you’ll get a mention at the next quality improvement meeting. Cut the resignation rate of ED staff by a quarter, and you’ll have a transformative impact on the entire organisation.

So how do we get there? It won’t be easy. Interacting with PACS servers is tough enough. Getting AI into the hands of staff in the ED, clinics, wards and theatres is another level entirely. It requires smooth integration with large numbers of disparate clinical systems. Just overcoming the technical and data governance issues is really hard. But even once you have access, much of the useful data is unstructured: couched within the machine-boggling complexity of natural language.

Yet there is hope. Here in the UK, for example, a team of HDR UK-funded researchers are developing a framework to integrate unstructured data from a range of clinical systems and run intelligent analytics. Their solution is called CogStack and it’s already live within the NHS. (You can hear directly from this team and many other leading clinical AI innovators at the upcoming GTC conference – free registration.)

In the meantime, what can us mere mortals do? Stay engaged. Keep opening up the conversation within your organisation. Be the person who reassures their colleagues that AI isn’t coming for their jobs.

In fact, it might be the best recruitment tool since they took away the teabags.


Dr. Rob Brisk is Developer Relations & Alliance Manager, EMEA Healthcare, NVIDIA and Specialty Doctor, Dept of Cardiology, Southern HSC Trust