Kieran Walsh: Your new interdisciplinary team member—the robot

Interdisciplinary team learning and working have been in vogue for several years now. Both are based on the fact that healthcare is a team activity. For example, we cannot function as doctors in any specialty without nurses, therapists, pharmacists and a range of other healthcare professionals. But until now, there hasn’t been much mention of another interdisciplinary team member—the robot. There is a good deal of discourse about robots in healthcare—but much of this is dominated by how robots will do our jobs and make us redundant. And not enough on how we might work with robots.

Daugherty and Wilson have recently written a book on exactly this subject. [1] In an early example, they cite the case of online maps. The first online maps were reproductions of paper-based maps. They were handy though, as you could access them for free on your computer and then on your phone. The next development was being able to enter your destination onto the online map and getting directions. Also, really useful. But then the real revolution came when mobile apps started taking advantages of real time data such as traffic jams and driver speed to give you just in time directions that would take you by the fastest route at that particular time.

It is easy to think of equivalents in medicine and healthcare. Patient flow is a perennial problem in hospitals and healthcare more generally. Patients often get stuck at various points of their journey—such as waiting for an investigation or for an appropriate bed. We create care pathways to help healthcare professionals and patients alike navigate their way through the system. But a bit like paper-based maps, the pathways are static. Does it have to be like this? Could online pathways be based on algorithms that learn from real time data on usage of the system and suggest to patients the shortest way through. So an appropriate investigation might be an MRI or a CT, but the dynamic clinical decision support pathway might suggest one or the other depending on actual data on waiting times. A robot might do this but no-one will lose their jobs. The healthcare professionals will still be needed to make judgements and advise and empathise. They will also need to continually develop new algorithms as the evidence base for treating various conditions change.

Will working with machine learning robots make our jobs harder or easier? This is more difficult to say. It is tempting to give the robots all the easy and boring jobs that we know that they can do. But a problem with this approach is that it leaves us all the difficult jobs. I spoke to a clinician a little while ago and she told me that twenty years ago when she qualified as a GP she would see a few “easy” patients in every session. These were patients who only needed to have their blood pressure checked. They probably didn’t need to see the doctor but there was no one else to see them. But nowadays she doesn’t see any patients like this—they all go to the practice nurse. She only sees complex patients with lots of things wrong with them. And this is cognitively and emotionally exhausting.

We don’t want robots to make this even worse. It would be nice to think that we could give them some of the more difficult jobs. This might be good for patients and for ourselves.

Kieran Walsh is clinical director of BMJ Learning and BMJ Best Practice. He is responsible for the editorial quality of both products. He has worked in the past as a hospital doctor—specialising in care of the elderly medicine and neurology. 

Competing interests: Kieran Walsh works for BMJ which produces which produces the clinical decision support resource BMJ Best Practice.


  1. Daugherty PR, Wilson HJ. Human+ Machine: Reimagining Work in the Age of AI. Harvard Business Press; 2018 Mar 20.