The idea that technology and AI can boost productivity and free up time for care in the NHS has become a major focus of health policy. Drawing on a recent Health Foundation survey conducted in partnership with eight professional bodies, this session will explore which technologies might help and what steps will be important to realise these gains.

Tim explores:

• Which technologies NHS clinicians say are making a difference right now, and which they think hold most potential for freeing up time over the next few years
• What challenges staff and organisations face in implementing and using technology effectively, and how they can be supported to realise the gains that technology offers
• The links between freeing up time and improving productivity, and what NHS clinicians tell us they might actually do with freed-up time

Expert: Tim Horton, Associate Director, The Health Foundation

Transcription:

 

Good afternoon or good evening, depending on where you're joining us from and welcome to today's webinar. My name is Ciarán Walsh. I'm clinical director at BMJ, and it's my pleasure to be your moderator for today's session.  The topic for today's webinar is how can technology help free up time for care in the NHS?

And we're delighted to be joined by our expert speaker, Tim Horton.  BMJ Future Health is a new community that we've created, which consists of a webinar series, podcasts, and a live event in November of this year.  We believe that innovation through digital health solutions are now essential to creating thriving health systems that are financially stable, that support the workforce, and that deliver better patient outcomes and healthier populations. 

At the end of this presentation, we'll have a few minutes of Q& A where you can submit your questions. Please do add your questions to the Q& A box.  Just a reminder that this session is being recorded and will be made available to watch afterwards. It's now my pleasure to introduce Tim Horton, Assistant Director of Insight and Analysis at the Health Foundation,  and Tim will be exploring what technologies clinicians say are making a difference right now and what they think has the most potential for the future. 

Also, what challenges staff and organizations face in implementing and using technology effectively. And lastly, the links between freeing up time and improving productivity. And the all important question, what NHS  clinicians might actually do with all the freed up time. Tim, your warm welcome to this BMJ spotlight session. 

Over to you, very much look forward to your talk.  

Thanks so much, Kieran. And it's great to be here. For this session today. Can I just check as I can see you, Kieran. Can you see my slides? Okay. Yeah, hear me. Okay. Fantastic. Thank you everyone for joining us. And it's a real pleasure to be here. As Kieran said to talk about how can technology help free time for care in the NHS. 

This is obviously a topic that's of big interest at the moment. There's huge policy interest in the potential of technology particularly in the UK thinking about how it could help potentially tackle some of the pressures the NHS finds itself under with escalating demand for care.  and workforce shortages as well.

I should therefore start with a little bit of a qualifier just to say that while we at the Health Foundation believe technology holds huge potential for improving care quality and productivity it's not a panacea. It's part of a broad solution to tackling pressures in healthcare. And it can't compensate for underfunding or workforce shortages.

So it's part of a mixed solution and I just want to be clear I'm talking about it in that context.  But now I want to tell you about work that the Health Foundation has been doing over the last year. Looking at technology and time for care in the NHS, if you don't know us, we're an independent charity and our mission is to improve health and health care in the UK.

And what I'm going to share with you is some work we've done over the last year, looking at this topic.  As I said, it's of huge interest to policymakers and debates about technology and how it can improve healthcare are very often dominated by policy voices, industry voices, technology commentators.

What we wanted to do in this bit of research is ask healthcare staff what they think about technology and help center their voices in this debate.  To do that. There's a quick view of the work I'm going to talk to you about today. It's three publications that we've put out over the last year  and to try to center healthcare workers voices in this debate.

As I've said, we collaborated with eight professional bodies.  through an online anonymous survey of UK clinical staff for which we've got 560 complete eligible responses. And that's supported by some qualitative interviews with experts on healthcare and technology.  We also commissioned a rapid evidence review.

of the impact of digital and telephone technologies on staff time in healthcare, really to help us understand the scope of the literature here. And this was all supplemented with our own in house research and analysis and stakeholder engagement.  Just before we crack on with the results, let me say that in this work, we were looking particularly to target our research on staff groups with workforce shortages.

Why was that? There's a strong sense that technology could help support workforce capacity and free up time for care.  And clearly, where there are workforce pressures, there are particular hopes that technology can be part of the solution.  For that reason, we targeted the following staff groups with our research.

Anesthetists, working with the Royal College of Anesthetists. GPs, working with the Royal College of General Practitioners. Intensive Care Doctors, working with the Faculty of Intensive Care Medicine. Nurses, working with the Royal College of Nursing. Obstetricians and Gynaecologists, working with the Royal College of Obstetricians and Gynaecologists.

Thank you  Physiotherapists, working with the Chartered Society of Physiotherapy.  Psychiatrists. by the Royal College of Psychiatrists, and finally, radiologists and oncologists by the Royal College of Radiologists.  So first of all, I want to show you some of the results thinking about which technologies might offer the biggest opportunities to save time in the NHS. 

So to explore this, we presented our survey respondents with a list of 25 common technologies or classes of technology, and we asked them which ones that they've used in their work. And then they were asked which ones of those were saving them the most time compared to non digital ways of working.  And we found the top three technologies that clinical staff we surveyed said were saving them the most time were, first of all, video conferencing to speak to colleagues.

Secondly electronic health records that was picked by 50 percent of respondents as being one of the top three technologies saving them time. And finally,  third place digital messaging tools for communicating with colleagues picked by 29%.  You probably can't read the footnotes at the bottom of this slide, but this shows you how the other technologies were ranked. 

And I think there's a few really interesting things that come through from these results. The first is the clear potential for technology to support administrative and operational work and into professional communication in the NHS.  This perhaps contrasts with a lot of policy and media focus you get on niche clinical technologies which of course are very important.

Often these technologies for supporting administrative work are of less news interest and less public interest. And what we've found in work we've done looking at the balance of investment in innovation programs in the public sector, very often there's a neglect of innovations to support administrative and operational work.

So this really speaks to the need to focus on the potential gains that these kinds of technologies  bring.  The second thing is it's interesting. How highly ranked tools for interprofessional communication were and this compares to, for example video conferencing to speak to patients that was chosen by only 9 percent of respondents.

Similarly, digital messaging tools to communicate with patients was picked by only 10 percent of respondents.  And again, I think this is really significant when we think back to discourse about service change in the NHS, it very often focuses on the patient facing aspects of it. An example would be during COVID, there was a lot of interest in the way consultations were moved online triage changed for primary care and so on and so forth. 

Actually, what these results are suggesting was that in the NHS in England, maybe it was the decision to implement Microsoft Teams across the NHS. A real revolution in ways of working. Maybe that was the most significant service change that happens during COVID.  It's also worthy of note that these kinds of technologies supporting interprofessional communication and also electronic health records are key aspects of how the NHS is trying to move to more joined up integrated care as well.

So it's really interesting that technologies that support that agenda are getting ranked so highly.  Next, we wanted to think about what kinds of technologies staff were saying might be most likely to generate time savings over the next few years. So we presented them with a whole list of technologies again, regardless of whether they'd used them in their work and asked them which technologies they thought were most likely to deliver time savings within the next five years. 

The results here were more spread out, but there were still some technologies that emerged as clear front runners. Top of the list picked by 31 percent of respondents were clinical documentation tools. Followed again by electronic health records and in third place picked by 23 percent software for the analysis of images and test results. 

So here it's really interesting when we've shifted respondents focus to the next few years.  What we found is that technologies like video conferencing to speak to colleagues or digital messaging tools to speak to colleagues, which came out top in terms of saving time right now. They fell down the list. 

And we've got a couple of new entrants here clinical documentation tools and software for the analysis of images and test results. It's really interesting though that electronic health records continue to be highly ranked when we're thinking about where gains come from over the next few years.

And it's suggesting that not only are electronic health records being perceived to help within the NHS right now, but staff clearly see them as a source of potential gains over the next few years. And this suggests that We should be keeping our interest in how we optimize the use of electronic health records and get more gains from them. 

There's a big focus, particularly in the NHS in England at the moment, in ensuring that providers have electronic healthcare records. What we can see when we look at the use of these kinds of technologies around the world is that when you've got them, that's, first base. in getting the gains of them can often be quite a long period of thinking.

How do you use them? How do you optimize the use of them to maximize the benefits from them? And this technology being ranked so highly here suggests that it's fertile territory for thinking about what the next phase of the strategy is. On electronic health records.  I'd also note in passing here.

It's interesting. These are technologies that whilst they might not be in every health care provider are currently in use in various locations within the NHS. And what this is suggesting is that Staff think gains over the next few years will probably come from better use of existing technologies.

Often debates about technology in healthcare like to focus on what's just being invented and cutting edge new technologies just emerging from the lab. Realistically, gains over the next few years are probably likely to come from the spread and optimization of existing technologies, like the ones we see on this slide. 

We also asked staff  what they thought the potential was for artificial intelligence to save them time in their work over the next few years. And as you can see from this slide, there's reasonable optimism about this. 57 percent of staff said that they thought a I was either very likely or somewhat likely to save them time in their work. 

And interestingly, a I could play a really important role in all of the three technologies we've just looked at that staff said might offer time savings over the next few years. Clinical documentation, for example, ambient voice technology, which uses both transcription, but also natural language processing analysis of images and test results. 

And of course, with electronic health records, layering in machine learning to that. It's interesting to note that AI might be highly relevant to the technology staff work we're flagging as potential sources of time savings in the next few years.  We also took the opportunity to ask staff what main barriers they were finding to using technologies effectively in their work. 

And we presented them with a range of common barriers. We find that clinicians in the UK site the most highly ranked barrier was a lack of IT support and expertise followed by lack of funding to implement new technologies. And in third place, poor internet connectivity.  To me, this really highlights an important point. 

If an obvious point, the gains from technology don't come from just having the technology, but from implementing it and using it effectively. And that means we've got to worry about the barriers that staff say they're facing in doing that.  We won't get the productivity gains from technology unless it's being effectively implemented and used.

So these findings suggest a range of fronts where we need to act if we're going to support providers in the UK, but in any country to effectively use technology. It requires, as I say, not just having the kit itself, but work to embed it effectively in a local context. Having the right skills, the right ways of working and the right roles and processes surrounding that technology,  it's a challenge we often see at the health foundation, where we fund teams on the ground to implement innovations, including technologies.

So let me just say a little bit more on how the implementation of technology can be supported.  In short, there's a range of things that. People adopting a technology might need to support them to do it effectively, in addition to the technology itself, which could be training or evaluation.

It could be skills in change management or design or quality improvement as well as support like peer learning to understand how peer organizations and teams are doing this.  None of these are things that cost huge amounts of resources, but often investment in them can reap significant dividends in terms of getting better benefits out of service innovations in healthcare, including technology. 

Seeing as I've mentioned implementation, one important aspect of that is whether or not you're involving users in discussions about not just the design of technologies, but their application to healthcare processes in the first place. And several of our experts interviewees raised concerns about really the lack of clinician involvement, typically in technology development and procurement where decisions might not be sufficiently driven by staff demand. 

So a couple of interesting quotes here. One from Dawn Dowding, Professor of Clinical Decision Making,  who said, quite often what you end up with is technology companies coming to you with a solution looking for a problem to solve.  And also a nice quote from Joseph Alderman anaesthetic and intensive care registrar and doctoral researcher said we need to move towards a future where technology is seen as something that is developed with clinicians and patients rather than something that is developed for them and then applied to them. 

The risk if you don't involve the end user in the design of the intervention and the user about the use of the technology is that you're getting get tools being implemented that might not be suited to the needs in the context, or you might end up getting them in effectively implemented and used as a result. 

Out of this first bit of research, we compiled some recommendations that we thought were relevant for policy makers and service leaders. Again thinking about the UK, but relevant in any healthcare system.  First of all, there's probably a need to give greater priority to technologies that can help with administrative and operational tasks. 

Second, we should recognize that many of the most immediate gains will come from the optimization and spread of existing technologies, and it might be worth having specific strategies to think about the optimization of those technologies, for example, electronic health records.  A key part of helping organizations, teams and health systems do this well is supporting rigorous real world testing and evaluation of emerging technologies. 

We also need action to tackle the barriers staff face in the effective implementation and day to day use of technology, such as improving underlying infrastructure providing IT support. Maybe offering embedded implementation support in procurement programs or service transformation programs. 

And finally there'll be benefits from increasing staff involvement in demand signaling for new technologies and in technology development and deployment to ensure that the tools that are being developed and the ways in which they're being applied to healthcare processes fit the local context and match the local needs. That was the first bit of our research. We then wanted to move on to look at a related and a highly important issue. How would clinicians use time freed up by technology?  We were keen to explore with this bit of work, how time freed up is repurposed in practice.  And to be honest, this involved challenging some very common assumptions, particularly policy assumptions about the way freed up time might translate into increased productivity. 

What's the evidence base on freed up time? Just to give you the context much discussion on technology is really focused on the first part of the implementation challenge I was describing to you. Focus on ensuring that technology is successfully implemented and so frees up time.  But of course, another critical causal link in the chain that has to work, if time saving technologies are going to deliver productivity benefits, is to ensure that freed up time is then used effectively.

And this is something that's often assumed, but by no means guaranteed.  So we wanted to explore the evidence base for this through the evidence review we commissioned.  And what it found was only a tiny proportion, it estimated less than 1 percent of studies on the impact of technology on staff time.

Actually consider how freed up time is used.  To estimate this, it looked at around 35, 000 papers looking at the impact of technologies on staff time in healthcare. And sure enough, only a tiny proportion considered how freed up time was actually used. So this is a significant gap in the literature.

Occasionally you find a study that will seek to observe how time is used or ask staff themselves to report on it. But the vast majority of studies don't provide data to support claims about the use of freed up time. Really giving the impression that it's assumed that it can automatically be repurposed for whatever the desired activity is. 

So we wanted to explore this a bit further, and given that one method that's sometimes used in the literature, for exploring the use of freed up time is self reporting by clinicians. We thought we'd expand this methodology. And what we did was we presented our clinicians with some scenarios where a technology might free up a certain amount of time in their day job.

And we asked them how they would be likely to use that time.  So in the first scenario we gave them, we asked clinicians how they might use one hour, a freed up time, and we presented them with a range of activities and asked them to select one activity on which they would be most likely to use that time. 

As you can see, the most highly ranked choice was direct clinical activity or patient care, which was chosen by 27 percent of respondents. Followed by reduction of my overtime chosen by 17 percent of respondents and in third place was quality or service improvement activity.  So it's interesting. Very often I've seen an assumption that 100 percent of freed up time will get repurposed like for patient care.

What we're finding here is that it is the most highly chosen option, but it was only picked by 27 percent of respondents.  And in fact, when we looked at the results from primary care, we found that the most highly ranked choice there was reduction of my overtime.  I'll also just mention with these results we've got here interestingly, delivering education and training ranked higher than receiving education and training, but probably reflected the fact that within our survey sample,  There was quite a high proportion of people at a later stage in their careers who might be more involved in delivering training than receiving it. 

Amongst those at an earlier stage in our careers, receiving education ranked more highly.  So that was a glimpse at how staff might use one hour of freed up time.  We also asked how they might use three hours of freed up time. And here we allowed them to pick three different activities.  You can see that when we did that the pattern changed.

Now quality or service improvement activity comes top, chosen by 48 percent, followed in second place by direct clinical activity and patient care, selected by 46 percent of staff.  We, for that 46 percent that  said they would use some of that freed up time on direct clinical activity or patient care. We also asked them how much of the three hours freed up would you spend on direct clinical activity or patient care? 

And to cut a long story short they said that 59 percent of that three hours around 108 minutes would go on direct clinical activity or patient care.  And when we look at 46 percent of respondents saying they would spend 59 percent of their three hours on that, that once again equates to 27 percent of total freed up time.

being used on direct clinical activity or patient care. It's always nice in research when two different methods converge on the same result. And both of these scenarios created a result of 27 percent of freed up time being used for direct clinical activity or patient care. Something we should also note is that spending more time on patient care could imply treating more patients.

and increasing throughput. It could also imply longer consultations, spending longer with patients, with an assumption that would improve quality. And many of the experts we spoke to highlighted the potential to use freed up time to improve the quality of consultations. Here's a quote from Asif Bashlani, a consultant psychiatrist.

He said, maybe if we had more time to reflect about patient care, more resources in terms of both time and capacity, we could have more directed beneficial treatments for people who are disadvantaged or marginalized. The ideal scenario, we'd have a mixture of focusing on population health, determinants of health, better personalized care.

I think potentially even early intervention. So really interesting hearing some of the ways that people at the cutting edge of this are thinking about how that freed up time could be used in this case to improve the quality of care on offer. But I wanted to highlight that we shouldn't assume that devoting it to patient care will just be about increasing volumes of care activity.

It could be about improving  quality as well.  What all of this is suggesting is that planning for the reuse of freed up time is essential when you are using technology to try and free up time. I've got on this slide a copy of a slide from Eric Topol's review for the NHS in England, which he did in 2019,  which highlighted some of the technologies that could make a difference in freeing up time and some of the key considerations for the NHS to get this right. 

And in some of those case studies, there were assumptions made about the use of freed up time in this case, suggesting image interpretation could give you back the equivalent of 500 radiologists time.  But again, we shouldn't just assume that there will be a like for like translation of freed up time to patient care.

What you can see on the right hand side of this slide is all of those different calls on the marginal hour that might get freed up.  Certainly we don't want to suggest that the 27 percent figure that came out of the hypothetical scenario in our survey is some kind of standard rule or let alone an upper limit on the reuse of freed up time for patient care.

But what it does suggest is we can't assume that freed up time will automatically get repurposed for patient care.  In reality this is going to be amenable to how you manage change in practice, and it suggests that planning conversations about how freed up time should get reused will be essential to deriving the gains from it. 

So again, we produce some recommendations for this aspect of our work. First of all, a call for more realistic estimates and modeling about how freed up time gets used. Like I say, often see policy assumptions, there'll be 100 percent shift of freed up time into patient care. What from our survey, that won't necessarily happen. 

Second, We need effective implementation and change management processes really to ensure that time freed up is used purposefully. There's lots of examples in the literature of technology being successfully implemented, but not delivering productivity benefits precisely because there wasn't effective planning done for how that freed up time should be used. 

We need more evidence on how any freed up time by technology is used in practice, try to tackle that evidence gap that I highlighted.  We also need a broad view of how we can use freed up time to improve NHS productivity. So using freed up time to increase care volumes is only one way in which freed up time can improve productivity.

And in fact, all of those other activities on which time might get spent  also have benefits for productivity. So training, for example, higher skill staff are associated with higher productivity. Quality improvements can lead to more streamlined services and improve productivity. Taking breaks will lead to less fatigued staff, less risk of burnout better health and higher recruitment, retention, and productivity as a result.

So we need a broad and nuanced approach. about how freed up time will go to improving productivity. It won't just be through increasing care volumes.  And we also think there's a chance here really to make an offer to the health and care workforce to ensure that some of the time freed up by technology could be used in ways that improve job quality and make work more rewarding, perhaps as part of an offer where time is also agreed to increase care volumes as well. 

In the final part of my presentation now, I want to come on and look specifically at some issues  to do with automation and artificial intelligence and how they might affect work in healthcare.  The material that I've shown you so far is really looking at technology in general. And we touched on artificial intelligence within it, but artificial intelligence recently has assumed a new role. 

a lot of prominence in thinking about reform of the labor market, particularly last year, with the leap forward in generative AI, large language models, for example, chat GPT barred and tools like this. So I just want to finish by looking at how automation and AI will affect work in healthcare.

And to start with thinking about this, I want to just try to share a framework for thinking about different modes of automation in healthcare. So technology will impact on work in different ways, depending on how it's used. And this quadrant here shows four potential different modes of automation.

Let me just talk you through this figure. On the right hand side, we've got instances where a machine might replace a human worker. So in the bottom substituting for a worker. So for example, a robot to transport samples from a ward to a lab. And in the top right, Where a machine might supersede what people can do and replace them as a result.

For example. A pill counting machine.  On the left hand side of this figure we've got by contrast examples where machines will be used to assist human workers rather than replace them. Whether that's supporting them, for example, a software for dictating notes or to strengthen and enhance their performance, for example, an augmented reality headset to support surgery. 

So the right hand side is where machines replace humans. The left hand side is where machines assist humans.  Then if you look at the top half, that's really focused on instances where machines are being used because their performance exceeds human performance like our pill counting machine or our augmented reality headset.

And in the bottom half of this figure, you've got instances where machines are really being used because they can match human performance or they're good enough rather than exceed it. So that gives us these four use cases substituting, superseding, strengthening, and supporting.  We find technologies used automation and AI technologies used for supporting and substituting, particularly to with the motivation of freeing up staff to focus on other tasks,  whereas we tend to find technologies used for strengthening or superseding in order to improve performance, though they can also free up.

Staff as well.  Another thing I should say is strictly this is not a way of categorizing technologies, but of categorizing uses of technology. So we could take the same kind of tool like a large language model here. and look at different uses of it in this quadrant. For example, large language, using large language models to generate clinical letters instead of human administrative staff would be an example of substitution.

Whereas using large language models to rapidly translate letters into different languages would be an example. far supersede what human staff could do. Or if we look at instances of large language models assisting human task performance they could be used to screen for clinical trials and support that work, or they could use to really strengthen human task performance, for example, by synthesizing large volumes of clinical information and using that analysis to inform decision making.

So this is really a framework for understanding different uses of technology.  I should also say that where you might situate a technology on here might change as it evolves and becomes more powerful. So you could imagine a simple clinical decision support system really supporting a clinician with simple prompts and reminders.

But then as it's made more powerful, for example, by creating greater data linkage, Or maybe layering in analytics or even machine learning it would get to a point where it's performing work that the human laborer cannot do and therefore strengthening their performance.  You could imagine the role of it changing and actually moving to a system of automated decision making for certain kinds of questions.

That's not something that's on the table at the moment. And it would pose some serious issues around risk and accountability and ethics, but you can imagine the way a technology interacts with work evolving as the technology itself becomes more powerful and gains more functionality.  Now,  when chatGPT 4 came onto the scene last year, it seemed to spawn a bout of soul searching about what this would mean for the labour market, whether AI was gonna take people's jobs And what it might mean for the future labor markets. 

What we find when we look at labor market modeling is that automation and AI are believed to be less likely to lead to widespread job displacements. in healthcare than in other sectors of the economy. Certainly they'll have a big impact, but they are estimated to have a much less intense effect on job displacement than in other economic sectors like manufacturing or transport. 

Why is this? I would suggest there's several reasons why healthcare is different. First of all, many healthcare tasks require traits or competences that technologies currently struggle to replicate, such as critical thinking skills, negotiating skills various kinds of creativity.  So tasks requiring traits that are at the moment quite hard to automate. 

Secondly, a lot of healthcare is seen as intrinsically human. We place great value on the relational interpersonal dimensions of healthcare. And in some ways that can limit and constrain the scope of the use of automation and AI. We might regard there as being certain tasks which can't be delegated to a machine because there are issues of compassion, respect, dignity or other aspects of care quality like person centeredness at stake. 

A third reason why healthcare might be different to some other sectors is that there are actually few roles in healthcare that consist wholly of automatable tasks. So technology might remove certain tasks from people's roles, but not others. And that will enable those workers to expand their time on non automatable tasks or indeed to evolve or grow their role in other ways. 

So I suppose the good news from a healthcare perspective is that at present, it's not modeled that automation and AI will lead to widespread job losses. Nevertheless, they will have a transformative impact on much work in healthcare. And I wanted to finish by just thinking through this process a little more.

And I was going to do that by taking a case study to talk through. Before I do that, let me show you a survey result,  where we asked staff in the NHS what they thought the impact of automation and AI would be. This was a survey from 2021. We asked staff to choose between two different statements, one suggesting that the main impact of automation would be to improve the quality of work,  and one suggesting that the main impact would be to threaten jobs and professional status. 

And it was finely balanced, but on balance, NHS, more NHS staff, 45 percent said that the main impact of automation in healthcare would be to improve the quality of work by supporting them and enhancing their capabilities. Only 36 percent said that the main impact would be to threaten jobs and professional status. 

I should qualify that though, by saying that amongst different professional groups, the results were different. For example, medical and dental staff were much more positive about the potential future impact of AI than, say, healthcare assistance.  And that really flags that the impact of different technologies will be distributed unevenly through the NHS workforce and the healthcare workforce in general.

And it's going to be really important to think about that differential impact in thinking how we're supporting different roles and staff to adapt.  Let's look at a case study to finish and I've picked here the role of a GP receptionist. This is a receptionist in a primary care setting. And we wanted to think about how the role of the receptionist might evolve.

with the increasing use of automation and AI.  So on this slide, we've used that framework to  analyze how a variety of tasks that make up the GPU receptionist's role might be affected. And you could see that All modes of automation are relevant here. For some of the receptionists, historical tasks might get substituted out like checking in patients now replaced by touchscreen check in or scanning letters or sorting posts. 

On the top right you might also get the use of technology which far supersedes what a receptionist could do, and therefore is used not only to replace the receptionist on the, on those tasks, but potentially to do it faster or more accurately as well. For example, translating the patient communications into other languages, sending out reminders,  and letters and so on. 

There's also a range of ways that technology will support the receptionist role, whether it's supporting work like processing prescription renewals or transcribing letters or triaging appointment requests or strengthening what they do, for example, planning staff rotas or helping analyze and identify failed patient contacts.

So there's lots of tasks that make up the receptionist role that will be impacted by technology in different ways.  However, it's also clear that there's lots of other tasks in the GP receptionist role, perhaps with a more significant human dimension, where even though technology might play a role, are much less likely to be automated.

For example, attending a front desk and welcoming people, handling queries gathering information from colleagues and supporting communication with staff in other healthcare settings.  So to summarize, the GP receptionist role has a range of tasks that might be replaced by technology, a range of tasks that might be supported by technology, and a range of tasks where technology might play a much smaller role. 

So what does this all add up to? It's clear that This role will evolve and looking at the kind of tasks on this yellow post it on the front of this slide. There's this clear scope to grow the receptionist role to be adding value on more of these tasks, doing what humans do best. You could see the role, for example, becoming one that grows much more towards supporting patients. 

care navigation, maybe identifying and helping vulnerable patients, maybe supporting communication within a GP practice or between different settings and play an important role in the integration of care. So there are many different ways this role could evolve. It's critical that.  The receptionists and staff themselves that are most directly affected are those that are playing a role in thinking about how this process will happen,  how jobs will change, what we want those jobs to change into. 

But there's an example of how you can think about different modes of automation to really envisage how a role might evolve in the future.  What we think is really important is looking across the health and care workforce, is that  We develop a shared vision of the way in which technology is going to impact on roles and how they will evolve as a result. So we need a shared vision for how professions and occupations will evolve. It's critical that healthcare staff, of course, play a central role in developing that, with all staff groups supported to take part. So actually, many of the clinicians we work with on our survey are lucky to have such powerful and effective professional bodies representing them.

GP receptionists don't. have those kinds of representative bodies. So it's also going to be important for employers, trade unions, and others to play a role here. It's critical that strategies for workforce education, training, and skills ensure all staff can capitalize on technologies and the opportunities that they present. 

And also professional development might be needed in other skills as roles develop. For example, let's say that receptionist role evolves more towards care navigation. That might require skills in coaching and shared decision making, for example. So it's really important that these offers are available with support prioritised for those who have to make larger adjustments. 

So look, that was a whistle stop tour through these three pieces of work. I hope you found it interesting, and I'm really keen to hear any questions and discuss it further. 

Thank you.  Thank you very much indeed, Tim. That was a fantastic  way through very important issues.  In technology and healthcare.

Yeah, we will now want to move on to our Q& A. If I can ask the audience members to please add questions they might have to the Q& A and while we're waiting for them to do that, why don't I kick off Tim with a question.  You've asked what?  clinicians might want to do with their read up time or would do with their read up time.

I wonder, has anybody asked patients and the public,  What they think should happen  with the freed up time of healthcare professionals? 

It's a really great question. And it's clearly important that happens. Patients and care users are key stakeholders that need to be involved in these discussions about how roles might evolve.

Over a number of years at the Health Foundation, we've surveyed the general public to understand more about their views, their preferences. and their hopes and fears around this agenda. Actually, something here and we often find coming through is a belief that technology can help in healthcare, particularly by freeing up staff.

And we found the UK general public are by large up for more use of many types of technology in healthcare.  But when they're asked about the risks they rank as highest is the risk that healthcare might become more impersonal with less human contact.  So I would say if I could summarize one message from what we've heard from patients in the public, it's making sure  we're capitalizing on technologies and embedding them in care settings in ways that don't compromise the human dimensions of healthcare and those interpersonal and relational factors that we all hold dear.

But yes, it's critically important. And I should say, we were very keen to speak to staff.  In the research I've presented to you, mainly because we felt that the debate was dominated by policymakers and industry voices, and we wanted to bring staff voices in. But of course, they're one of a number of voices that are fundamental to determine the future direction of healthcare. 

And there's no more important voice than patients and the public.  

Thanks, Tim. Thank you. So please do add your questions to the Q& A. We've still got some time and lots of time to get through lots of questions, but I've got another one, Tim, while we're waiting. Nearly everybody when you talk to technology companies and healthcare or anyone else, they all say they involve clinicians.

None of them will say, oh, we went away ourselves and we created this out of our own imagination. They all say they involve clinicians. How deep do you think that involvement genuinely is?  

It's a really good question. And  I certainly didn't mean to imply that technology developers, including an industry are somehow seeking to exclude users from the development process that would be mad, but I guess your question implicit in your question is a sense of levels of depth in how you do that.

You could engage clinicians to identify an opportunity, for example and then really look to use technologies that you've already developed or which are close to technologies you've already developed to try and solutionize in that space.  Much deeper would be to involve users further upstream. In the design of technologies even when you've got to that point though and you've got a validated technology, there's a second stage of the process.

How is that applied to healthcare and used? What's, if you want the use of the technology? And there, there needs to be much more work with clinicians and staff around  how you actually apply it, including in their local context as well. And some of that's about implementation, but some of it does care and reach upstream to issues like, say, the user interface of a technology and the customizability of it.

Is it going to be able to fit into different kinds of workflows? For example, does it have rigidities that are going to cause problems? So I don't know. I can't answer.  In general, how deep it goes, but what we hear from staff is that very often they're being presented with technology as a solution that doesn't quite fit the gap. 

Yeah, no, I love the quote. I'm going to steal it. If that's, if I have your permission, Tim, it's great. By all means.  We've got a question. Why do you think EHRs are viewed positively by NHS doctors when they have been cited as a major cause of burnout so often in the US?  

It's a really good question. And I should say that just because staff in our survey ranked them as one of the technologies with the greatest potential for saving time, doesn't mean we didn't also hear frustrations from staff about the difficulties of capitalizing NHS. 

As you saw the lack of support to use them effectively. So it's possible to both believe that there's great potential of them, but think there's a lot more to do.  I think in the US there's been a very specific issue, which is the way that they have been used. And in many respects, this is a positive thing to open up communication.

between a physician or clinician and a patient for email, for messaging and so on and so forth. And that's been done in a way I know that has really expanded the clinician workload in the U S and has happened in I understood it a relatively surprising and unplanned way. So they have been a significant source of stress. 

It goes back to this point of good implementation and change management. which requires planning, requires thinking about how benefits will materialize and how they'll be used. Otherwise, you will get situations like that. And of course, the US is now trying to think about how it moves forward to still get these benefits, but in a way that doesn't drive burnout and fatigue.

Got it. Thank you. We have another question. Thank you, Samar Badmuni.  On implementation adoption of technology, what on average is the proportion of investment in change management versus procurement of technology in healthcare?  Tim. 

Wow that is a good question. I, a lot is spent on the procurement of technology without a doubt.

We have, I guess that to simplify slightly, there's a lot of  people who think that technology is about having the kit, and when you've got that, voila, you've got the benefits.  As a result, we find that  The management of change and the related aspects of implementation. For example trialing designing interventions, trialing them, evaluating them, bringing stakeholders together to discuss it.

That,  that area is relatively underinvested in for example, when the Health Foundation's funding change projects, we've, we find quite small amounts of results go a very long way to facilitating successful change. But I would say it's underinvested in. That's probably not because someone has, sat down and thought, we're not doing that.

But actually  we find that the complex labor involved in adopting these innovations is often underestimated. So it's a sense of really helping to spread the view that getting the gains from technology is more than about just having the new kit. We would say, don't just fund the, don't just fund the tech, fund the 

change.

Thanks, Tim. Thank you.  Now we've got another question while we're waiting. So you mentioned AI, so there has to be a question about AI.  A lot of people say AI will bring efficiency,  but also, because it will automate certain things, but also they'll say, Of course, there'll still be human oversight.  So which is it going to be?

Can you have efficiency? And will the human oversight  stop you from getting as much efficiency as you should? Tim, what is the solution?  

Yes it's a good question. In many cases particularly where there are issues of clinical risk involved, where AI is being used as a medical device, that oversight is essential.

And it means that  Yes, it's not just a case of freeing up staff.  The staff that are freed up might be the ones that are involved in the oversight. Alternatively, new roles might need to be created to oversee these technologies. You're right, we shouldn't think about this in a simplistic way. And there's going to be work involved in making sure the use of artificial intelligence is safe and effective.

We already know that there are certain kinds of risks involved there. For example, if you don't have enough capacity on it, if staff are tired and burnt out, there's a risk you might trust an automated system too much and not question the results enough. So it's really important that we have proper capacity as well as the Regulatory and governance frameworks and assurance around the use of these technologies.

And yes, that will limit the efficiencies involved. It won't eliminate them. There could still be efficiencies, but it's a really important consideration.  

Fantastic. Thank you. I got another question,  Tim. Sorry. So what if AI gives you a really quick answer, but you can't work out how it came up with that answer. 

And then you spend ages trying to work out.  How it came up with that answer.  Tell me how you cut that Gordian knot. 

It's really difficult, isn't it? When you've got  machine learning and adaptive algorithms and a black box it can be really difficult to know simply how recommendations have been arrived at. In some cases, that will be a problem in terms of, assessing the accuracy of a tool and with the evidence standards and assurance around it.  In other cases, it might pose a procedural problem. For example, Kieran, what if a patient wants to know why that decision has been recommended? And we regard it as important to be able to explain that to them.

There there's a procedural issue we've got then in how to do that. So this is a big issue with machine learning that is coming down the track into healthcare. I don't want to generalize. My instinct is there will be some decisions where provided we're confident about the accuracy of the outputs of an AI system.

We're just happy to know what they are.  In other cases obviously matters of life and death, but maybe also where fairness is involved, like why are you telling me I'm at the back of the queue, rather than in the middle of the queue, where  a patient might want to know why. It's going to be a big issue, and we're only just getting to grips with it. 

Okay, great. Thank you. One last question. Unless other people have more questions, I'll take the opportunity. My question is still about AI. Did anybody in any of the surveys mention anything about the environmental effects of AI? The amount of energy used by machine learning algorithms and big data and what effects that might have?

On the environment question.  

Yeah. So as survey probably wasn't set up to elicit those wider thoughts, but what you've touched on is a massive issue, and it's something that's come up in work. The Health Foundation's engaged on in thinking about sustainability in health care, including how the NHS meets its net zero commitments.

The  training of AI models is requiring vast amounts of energy. You hear on the grapevine that large AI developers. are setting up their own power plants indeed to provide the energy to develop and train AI systems. So it's a huge issue. And part of a responsible AI agenda is thinking not just how we avoid negative health outcomes but how we ensure the development of AI is going to  mean positive.

benefits for healthcare, society, and the planet as well. So it's a huge issue, and it's something that needs to be discussed as this agenda evolves.  

Thanks, Tim. We've got one last question. I know we're in danger of running over, but then this is it. No more.  Thank you, Samar Bhatmuni on AI.  How well do you think healthcare is addressing the potential for AI to exacerbate?

Healthcare inequalities.  

That's a risk. It's a risk for many reasons. If more patient facing technologies in healthcare might exclude some groups who could have trouble using those. Alternatively, if AI systems are trained wrong on unrepresentative data they might be biased against certain populations.

We've already seen algorithm that failed to recommend treatment for Black, African, or Caribbean patients. So these are big risks. On the other side of the ledger though, AI could be a huge driver of reducing inequalities as well. When we think how transformative technology can be for people with mobility problems, for people with sensory impairments, for people with language barriers it could be transformative.

So we absolutely need to worry about the risk of exacerbating inequalities, but also I would like to see much more focus on how the use of tech and AI can be a driver to reduce inequalities too. 

Thanks. Thanks, Tim. That, that's fantastic. And of course, what our audience doesn't know is that we had terrible tech problems before you joined us and somehow you managed very calmly and coolly to get it all sorted out and then deliver a fantastic presentation.

So thank you very much again. very much. Tim Horton for, thank you for having me, having my terrible que answering the terrible questions that were thrown at you. Just a few last words from me. On behalf of BMJ Future Health, we are excited to announce that registration for our event on the 19th and 20th of November is now open.

Please scan the QR code on the screen. Visit our website, BMJ Future Health for Future information. Our next webinar in this series will be on Tuesday, 25th of June, one to 2:00 PM where we will be joined by Rajesh Agarwal, Michael McDonald. And Russell Gruen, who will talk about how to approach digital transformation across global markets.

And finally, I'd like to to thank all of you for participating. We really hope you find it helpful. Thank you once again.