More time
to be human.
It sounds counter-intuitive, but could it be that the rise of the machines will help deliver on the people priority in healthcare design?
POSSIBILITIES
Lately, AI has been high on the healthcare news agenda. There are reports of AI-powered tools predicting survival rates and recurrence for lung disease and heart attack, potential liver cancer cures being developed in just 30 days. Biotech company Insilico and University of Toronto researchers used protein structure predictor AlphaFold and drug discovery engine Pharma.AI to identify a previously undiscovered pathway for treating the disease’s most common iteration, hepatocellular carcinoma, and while clinical testing is required before use within hospitals, it’s undeniably ground-breaking stuff.
AMP-activated protein kinase (AMPK) fragment: a typical structure that could be used by AI predictors Image: ©Molekuul
The AI revolution could benefit the design of human-centric healthcare hubs that give us the power to facilitate more seamless, stress-free experiences.
There are hopes that AI will improve diagnostic processes, preventative measures and access to customised medicines and treatments – with cutting-edge technology helping to maximise efficient use of public funds.
What’s more, a new report from Europe’s largest sustainable venture capitalist investor, A/O PropTech, tells us that the rapid growth of startups harnessing AI for the built world has raised $12.3 billion in the last three years. We’re talking everything from computer vision and generative AI creating smart buildings and infrastructure, to construction robotics.
Indeed, early 2023 saw Dorset’s Royal Bournemouth Hospital – specialising in births, emergency and critical care, and child health – announced as the UK’s first NHS hospital to be built with AI technology, intended to provide the best value-for-money for taxpayers. It’s set to track healthcare standards, ensure design brief objectives are being met, improve efficiency and decrease costs by generating accurate, objective data and evidence-based real-time progress analysis via helmet-mounted 360-degree cameras on site.
Lessons around freeing up more time and funds to create spaces prioritising the human side of design can be taken into every sector. Yet healthcare is the area that could benefit most from the AI revolution, considering the potential impact of the chance to improve productivity, outcomes and satisfaction through design hubs that facilitate seamless, stress-free experiences. Connected, intelligent hospitals could cut queues and whittle down waiting lists, bettering system efficiency – and that’s the tip of the iceberg. Technology holds huge promise in reducing the burden on healthcare professionals.
In our latest piece of thought leadership, ‘Rehabilitating the Healthcare Machine’, we discuss the advantages of implementing patient monitoring and equipment tracking tech so that nurses spend less time hunting for what they need and more on patient contact. If tech can take care of the more mundane day-to-day essentials, such as ensuring fresh uniform is ready when staff arrive at work, they can get straight to clinical tasks and the all-important vocation of looking after people. This, in turn, could ease burnout and improve workforce retention, giving them more time to deliver the standards necessary and creating a ‘virtuous circle’ free of the moral injury that can occur when staff are physically unable to provide the level of care they’d like, due to strained systems. Optimising hospital performance frees up finances to invest in wellbeing, which leads to growth in satisfaction and staff retention, improved care and outcomes that feed back into optimised performance again.
When is a bed not a bed?
Case study: Kettering General. Source: NHS England
NHS AI lab Skunkworks helped Kettering General Hospital explore AI to improve bed management and generate options for moving patients in a way that would support the human team to make the best decisions, and achieve the goal of ‘right patient, right bed, right care, right time’.
Managed by a human team relying on individual expertise to deliver an air traffic control-esque system – deciding the best arrangement with a changing set of demands and patient numbers – bed scheduling is complex. No two patients are the same and staff are overwhelmed with options.
Patient admissions were likened to a game of Tetris, where the allocation of each bed can have a major knock-on effect to the smooth-running of admissions and patient welfare.
The team provided a proof-of-concept tool using historic data to create a virtual hospital predicting demand and making bed allocation suggestions to support better, faster decision-making. A Bayesian modelling approach was used to estimate how many patients with specific characteristics would present at the hospital over time, and compare different bed allocation approaches. Assigning beds based on the best bed available at the point of admission, the quick, less resource-intensive ‘greedy allocation’ method was implemented.
Showing current occupancy rates, forecasting bed demand and explaining bed suggestions, the user interface enabled staff to test the model on a range of patients with different attributes and constraints to validate its performance. They now hope to develop and operationalise the tool, building connections to patient data in real-time, refining the algorithm, and understanding how it can be integrated into real-world site management.
“I regularly hear that a bed is a bed and I know it’s not... But when you have those front-door pressures, you can’t get ambulances offloaded and you have beds in the wrong place, this is the time you need the real support, real-time data, an automatic risk assessment that is generated for each patient.”
Bed management staff Kettering General Hospital
Royal Liverpool University Hospital. Photo: Architecture NBBJ / HKS
“The growing pressure on clinicians and health systems to provide timely, efficient and effective care is here to stay,” says Dr Melanie Flory, neuroscientist, psychologist and associate director of research at the Royal College of Art’s Helen Hamlyn Centre for Design. Her research enquiry sits at the intersection of design, systems thinking, and cognitive neuroscience, seeking new insights into the interplay between emotion and cognition.
“The theoretical conflux of AI and precision medicine promises insight generation, swift decision-making, personalised diagnosis and prognostication, and better survival rates, among a plethora of other advantages.
“The opportunity that human-centric AI design and AI-assisted care affords is that of enhancing rather than diminishing the humanity of patient and clinician.”
Smart hospital systems – with improved operational efficiencies, automated workflows and connectivity between systems, devices, assets, data, and people – could allow integration with speciality clinics, primary care providers, and other healthcare venues to ensure patients receive the right care at the right time and place, anywhere in the community. Instead of every care service being warehoused in a one-size-fits-all building, more individual specialist units could be established to suit specific patient populations – each then hooked up to one big digital infrastructure network that monitors supply and demand in these smaller hubs.
“The ability for AI to efficiently and effectively manage the huge amounts of data in healthcare will not only help free up valuable time and resources but also directly improve health outcomes,” agrees Sam Shooter, Director, Hoare Lea. He believes the social value and impact of AI, meaningfully implemented in healthcare design, could be vast. Of course, it must be responsibly executed, with control mechanisms in place to maintain ethical practice. While there are known biases in algorithm writing and AI design can be seen to pose the biggest risk to the creation of inclusive design, if we remain acutely aware of these, AI-powered tools also have the potential to provide the answers and design for the widest range of people.
The opportunity that human-centric AI design and AI-assisted care affords is that of enhancing rather than diminishing the humanity of patient and clinician.
Dr Melanie Flory Neuroscientist and Psychologist
Eradicating bias: a continuous journey
“[Inclusive design] is particularly important in healthcare because typically excluded people are likely to make up a much higher proportion of the total user population within the healthcare space when compared with other industries,” says Becky Ferraro, Inclusive Design Lead at Deloitte. “It is becoming increasingly documented that there are patterns of exclusion that occur within healthcare and this can be the difference between life and death. For example, pulse oximetry devices used for warning of low blood oxygenation in covid-19 and other diseases may be missing three times as many cases of hypoxaemia in black patients as in white patients.* (*From a study report published in the New England Journal of Medicine.)
We know that bias is the enemy to an inclusive AI. But it can be challenging to get this right. As with all approaches to embedding inclusion into design processes, eradicating bias is a continuous journey rather than a one-time act.”
If there is one thing to keep in mind, it’s that AI will reflect the people who build it, so we need to ensure diversity in its creation, with a diverse team that’s always paired with a collaborative and open design process.
“This is an all-too-common pitfall, where people choose to design behind closed doors through fear of rejection, failure or loss of ownership,” says Becky. “In reality, these fears have a much smaller negative impact than reputational damage or harm to others because of an AI that showed bias and wasn’t inclusive.”
The other lens to collaboration in the AI development process is the idea of trust, given that the people that we need to input into algorithms and design process are perhaps also the people who trust us the least, who may be the least willing to be involved.
“Through the stories of AI going wrong and the fear of the unknown, many people may be hesitant to get involved,” Becky adds. “I think part of this comes down to thinking of AI as something already built and out of the box; something that is less controlled. However, if we work on flipping the narrative to talk about the limited nature of AI at the start, and the ability to train and continuously shape its output, it will help people understand the value their input can have on creating a more trustworthy intelligence.”
This is an all-too-common pitfall, where people choose to design behind closed doors through fear of rejection, failure or loss of ownership. In reality, these fears have a much smaller negative impact than reputational damage or harm to others because of an AI that showed bias and wasn’t inclusive.
Becky Ferraro Inclusive Design Lead, Deloitte
Send in the robots? It’s already happening..
Stroke patients undergoing rehabilitation are already using robots to practise assisted walking and hand grasping, and prosthetic robotics are enabling those with limb difference to pick up objects with their thoughts. Lab robots are testing hundreds of clinical samples every day. In February, a £350,000 robot was delivered to Cumbria’s West Cumberland Hospital to improve pharmacy staff efficiency. Aiming to take pressure off their dispensing service and allow staff more time on the wards, the robot is programmed to provide medication to hospitals and community clinics, managing and distributing large amounts with its robotically controlled arms. Elsewhere, in Bristol, robots have been enlisted for behind-the-scenes tasks including drug-packing, moving food, and assisting with operations in theatre. Pictured: Pepper, the humanoid robot, has been used as part of an experiment in Bristol to see how robots might interact with patients. Photo: www.roboterly.com
Latest podcast Human first, designer second
In the latest episode of ‘Exploration’, our guest Rama Gheerawo – inclusive design innovator at the Royal College of Art – talks to us about human-centric healthcare design today and juggling its multiple demands. Rama, who has used his skillset to address societal issues around ageing, healthcare, ability and diversity, speaks of the importance of healthcare placement, localisation and access, getting comfortable with uncertainty, non-binary thinking and leadership models based around empathy, clarity and creativity.
Listen via iTunes, by searching Hoare Lea
The ability for AI to efficiently and effectively manage the huge amounts of data in healthcare will not only free up valuable time and resources but also directly improve health outcomes.
Sam Shooter Director, Hoare Lea
Key trends in the inclusive healthcare experience
Digital inclusion and assisted digital strategies are increasingly important in healthcare. Digital provides an opportunity to connect with more people in multiple environments and allows people to receive care and advice within their own space. However, it’s important to remember that not everyone has the access, skills, or ability to engage with digital.
Therefore, there are two approaches that need to be considered. The first is an omnichannel approach, meaning people can interact with a product or service in several ways that best suit them. This could be digital, physical or telephony. Importantly, we need to ensure people can move between channels as circumstances change, and we need to design for flexibility. The other benefit of designing across channels is, of course, reliability, which is, again, increasingly critical in healthcare when we are talking about quality of life and risk of harm. If your digital channel breaks, what is the back-up? How can people continue to access the care that they need?
The second point is about providing an opportunity for people to upskill. Just because someone struggles to use digital products, it doesn’t mean they don’t want to, especially when the ability to interact with them could drastically increase their quality of life and the communication and care available to them. This doesn’t have to be something you include in the direct design, but it can be something you can easily signpost to, or allow a customer to request.
Co-design is a recognised and utilised tool being applied to many healthcare products and services. People are increasingly aware of the need to include the customer in the design process.
However, many people default to using the ‘co-design’ label on what is, in fact, a focus group centred around the problem. The benefit of co-design actually comes from including customers in the solution as well as listening to their needs and requirements. The simple switch people can make is using tools from the design process in ideation activity sessions and challenging themselves on whether they are solely relying on asking about the experience rather than acting on the experience.
It’s important to design for and with the entire care ecosystem; the patient, care providers, healthcare professionals, family and friends. Without taking this approach you can fail to identify barriers that could take away from the benefits of a product or service.
“Now is the time to bring inclusive design and designers to the table,” agrees Dr Melanie Flory. “To be truly human-centred means the broadest range of diverse stakeholders, from clinician and patient to carers and other support networks, must be included in informing and shaping design ideas and processes from ideation to innovation. Integrating frontline users, and their input and feedback, throughout the design process is vital to successful human-centric AI innovation and assisted care.
“Empathy-driven AI innovation is intelligence-based innovation.”
LET’S TALK