Interview, Secondary Care

Interview Series: Dr Tom Lawton, Head of Clinical Artificial Intelligence at Bradford Teaching Hospitals

In our latest interview series, Angela Sharda speaks to Dr Tom Lawton about his new role, his aspirations for taking Artificial Intelligence (AI) innovations from the labs to wards, what he has in the pipeline for the next 12 months and whether AI can help solve the NHS crisis.

Dr Lawton is a Consultant in Critical Care and Anaesthesia and has taken up a new role of Head of Clinical Artificial Intelligence (AI) at Bradford Teaching Hospitals NHS Foundation Trust, believed to be the first role of its kind in the NHS.

What will you be focusing on initially in your new role? 

The first thing is to gain a handle on the information that we have. We have data from around the region and in primary care and that gives us an understanding of what we may be able to achieve with AI.

Secondly, we are starting to look at the computing power side of things. The computing set up that we have had so far hasn’t been as high-performing as we need so we are looking at things like the use of the cloud.

What are the main problems that AI will help fix in your trust? 

It’s really important to remember that AI is not here to try and replace clinical judgement and nor is it a replacement for doctors and nurses.

Health professionals that work within hospitals are presented with more and more information about each patient – which is all very useful. However, you do end up with a sensory overload of information but we have made attempts to distil things down with AI.

What advice would you give to other organisations who are looking to do something similar? 

AI is a massive area and most importantly you need to get a good handle of what information you have and which information you can make available to generate systems. You then need to start to try and recruit expertise – hopefully there will be more available in the future.

We are working with local universities and the command centre dashboards. There is a great deal of expertise out there but you just need to approach your links to try and bring them into the NHS.

Can you talk us through any examples of large data sets initiatives that are currently ongoing? 

Google have done some amazing stuff with predicting flu outbreaks a few years ago.

It’s difficult because we are all very concerned about relationships with big companies and health tech but if Google were able to give that data to the NHS that would have been very helpful to predicting what may have happened.

We don’t yet have reliable ways to tie those things together but in the future I think this will become a bigger thing.

How will you ensure that there is better translation from the labs to the wards? 

The really nice thing about the Bradford Institute of Health Research is that you have links with universities but it’s on the hospital site and therefore, you have immediate access to clinicians.

I am a clinician so it means that you have got that sense check right from the start.

You see some technologies where someone has done something amazingly clever but you take one look as a doctor and think that is never going to work in reality.

The idea of clinical involvement right from the start is a great thing – we have nurses involved, doctors, analyst and allied health professionals which means that hopefully we can develop technology that is really useful for the patient.

What is in the pipeline for the next 12 months? 

We are looking at some smaller projects to demonstrate the power of what we can do. We have been looking at predicting test use in A&E and cutting patient waiting times from the point of which they arrive in A&E to the point of treatment and seeing whether it is possible to reduce that time.

We are also looking at a way in which we help save hospital beds by looking at patients who may stay in hospital after they have been declared that they are clinically fit for discharge.

Another project that we are looking at is how long patients will stay in hospital as part of their illness and discharge.

More medium term, we are looking at how the Government wants to use AI in radiology.

The possible challenges that will lie ahead with this could be around the safety and assurance aspect of this and getting people to trust you with data. Although, we are starting from a good place because we are the NHS and we are trusted with the data already.

Another challenge is that the role of the computer and AI would only ever be to advise rather than to dictate.

We also need to be realistic about technology and some of the outlandish claims that are made about what AI can do and it may well do some of those things in the future but it can make life difficult for people that are trying to make more realistic claims.

We need to make sure that things are tempered with realism.

Is AI the answer? 

AI is not the solution to solve all of the NHS’s problems but it is part of the solution.

By directing resource to where they can be used in the most efficient manner then AI has a role to play and by picking up some of the more simpler tasks.

You can save people’s time by presenting some of the information that AI has condensed down, it can help make things more efficient and it can assist in solving some of the problems but obviously not all of them.

Evidently AI can’t create physical hospital beds but it can tell you when you are likely to need beds and when there is to be a likely spike predicted in demand which will allow the NHS workforce to be more efficient with the resources that they do have.