Fireside Chat with Duncan McRobbie

Hi Duncan, can you tell me a bit about your work and how your interest in digital health developed? 

Yes of course! My interest in technology in healthcare really came about organically,  as well as from the realities of working in pharmacy in a hospital. Medicine is ubiquitous in pharmacy – identifying, tracking, recording and the necessity of ensuring prescriptions are legible and safe.

Early in my career, I worked with systems that were largely inventory-based with some clinical components – they worked very well for what they designed for, but were not integrated with the other digital systems across the trust. Being part of the process of deploying clinical systems in the hospital made me interested in how we could do things better and how to better utilise digital tools. I was interested in seeing how people did their roles without digital support and how we could use digital tools to enhance care. 

Over time, my role has increasingly involved supporting teams through technology, change and looking at how digital can make medicines use safer and more efficient.

It’s interesting to see how your role has developed. With that in mind, what do you think are the biggest transformational digital shifts for pharmacy and medication?

For me, I think it’s integration. The fact that everything in the hospital can be on one system, and that we can see everything about a patient, is transformative. Paper-based and unconnected systems were prone to error and were immensely time consuming. So, the integration of clinical data and interoperability are real benefits in terms of being able to see what other people are doing to support patients across the hospital. 

Definitely, digital integration has been significant in recent years! Could we also delve into dosing and the work you are doing on this within clinical decision support?

Dosing is an area where we are putting in a lot of effort at the moment, and I’m quite excited about what is going on. The main aim is to ensure patient safety by alerting clinicians when doses are outside the expected licensed ranges. Dosing can be particularly complex in paediatrics, where unlicensed doses often need to be used. 

We are working on this collaboratively with FDB to develop clinically informed dose guidance that will provide consensus support for unlicensed medications especially in children.  I would love this collaboration with FDB to go to a national level, rather than having each NHS trust work out its dose warnings in their own systems. We are also working with stakeholders on a project which will help us achieve this. Using the expertise of specialist paediatric pharmacists across the NHS will create more consistency and ensure that dose-range checks genuinely reflect real patient needs. Good clinical decision support should not be about telling people what to prescribe. but about giving safe parameters that support good prescribing decisions.

Taking a bit of a sidestep, genomics and pharmacogenomics are a hot topic at the moment, can you share how you see their future role in changing decision support? 

Pharmacogenomics is an exploding field and is becoming increasingly relevant to how we work with medicines. I started chairing an internal genomics group with a lot of really engaged experts and it is clear that the amount of information available is expanding quickly, but without the tools to interpret it effectively, this can be somewhat overwhelming. 

I think that the biggest opportunities at the moment lie in areas such as cancer care, where drugs can be expensive, complex and often toxic. If genomic information helps us to select a drug that is more likely to work and less likely to cause serious side effects, that benefits both patients and the NHS. Although genomics has long been talked about as transformative, implementation is still challenging because metabolism, genetics and clinical response are incredibly complex. In the UK, we have just had the first guidance on genomics testing to aid patient selection when prescribing Clopidogrel, and I am aware of quite a big list of medicines coming. 

There is potential that this is going to change the world we work in! Over time, I expect more genomic data to be directly incorporated into clinical decision support so that prescribers have clearer, evidence-based guidance at the point of care. 

Absolutely. Healthcare is changing rapidly with the exploration of genomics. Not only that, but the increasing implementation of AI is going to dramatically alter the future of the NHS. In your opinion, how is AI and machine learning shaping the future of healthcare technology? 

Technology in healthcare has been exciting for last 30 years – including the sort of work that FDB has been doing in the background for ages – which means we don’t have to go and check every drug-drug interaction manually. Now, AI and large language models (LLMs) are starting to influence the way we work in healthcare in many ways, but it’s really important that we use them thoughtfully. There is useful work emerging around ambient voice technology that can listen to clinical discussions and structure the information. However, the quality of the data feeding these models is critical. AI trained on unfiltered internet sources is far less reliable than language models trained on clinically validated content. Explainability is also essential because clinicians must be able to understand why a model has made a clinical recommendation. 

With that in mind, where do you think AI will go next and what’s in store in the imminent future? 

Some of the earliest practical gains may come from operational areas such as inventory management. AI may be able to analyse years of stock patterns and patient data and recommend more efficient ordering processes that reduce cost and waste. In the longer term, AI may support remote medicines reviews and more personalised dosing models.

Thanks so much Duncan, it's really great to get your take on all these exciting areas – I appreciate you sharing your knowledge and expertise!