The NHS is broken. After years of chronic underfunding, the waiting list has grown to more than 7.5 million people, with more than 40% waiting longer than the 18-week target. The public health sector is also struggling to recruit. As of December 2023, there are over 110,000 secondary care jobs available in the UK, including 10% of all nursing jobs.
Technology is seen as a solution to many NHS problems. There has been much discussion about how data and technology, such as artificial intelligence (AI), can facilitate patient self-service and personalization, increase efficiency, and free up time for doctors and nurses.
Speaking on a panel at the recent Women in Data event in London, Ming Tan, chief data and analytics officer at NHS England, said that while the health service is under extreme pressure, it is also He said they needed to find money to invest in services. technology.
“We need to invest in data and infrastructure,” she said. “That’s why a lot of our investments right now are not just putting a Band-Aid on something that’s broken, but perhaps understandably, rather than just putting a Band-Aid on something that’s broken, we’re investing in long-term things that will help us reach our goals. We’re focusing on that, because there’s too much broken stuff.”
The NHS also needs to have a healthy dialogue with the public about its use of data. Personalized medicine has great potential to improve treatments, but requires collecting additional, better data about individuals.
“[Without that] “We’re not going to look at whether we’re reaching out to inequalities. We don’t understand the population, we don’t understand the determinants of health. We’re not going to solve those problems. We cannot do that,” Mr Tan said.
“It all comes down to public trust. Data is a real opportunity, but we also need to explain to the public why data is so important.”
gain the trust of the people
Building this trust will become even more important as the NHS moves forward with plans for a federated data platform (FDP).
FDP is being developed by NHS trusts in England and 42 integrated care boards across the NHS to connect existing patient data in a safe and secure way, reducing the time it takes to review and find information about patients. will be used.
Mr Tang said FDP would build an underlying cloud-based data infrastructure at the heart of the NHS, serving as a national instance as well as providing individual versions of the software for integrated care boards. He explained.
“The key is to have connective tissue that can be shared appropriately when needed,” she said. “If you’ve ever been on a hospital ward, you’ve seen nurses with pieces of paper in their pockets scribbled down on something. For handover, information is passed from one paper to another.
“We’re talking about workflow management, decision support, and making sure that we make the lives of frontline clinicians and administrative staff easier, so they have more time for patients. We’re talking about that,” Mr Tan said.
She said there had been much discussion about how AI would transform the NHS, but the organization was still in the early stages of implementing the technology. There are many research teams looking to leverage technology, including looking at productivity improvements. For example, simplify the process of creating discharge forms.
“This is a human implementation of AI,” she said. “This is really important because you learn to test and iterate instead of jumping in and thinking that AI is going to replace your job. It’s not.”
Data preparation
Before the NHS can begin a major project, a lot of ground work is needed to ensure its data is suitable for AI.
“People forget that you have to think carefully,” Tan said. “Next, it’s important to have ethics and processes in place to get the most value out of what you do with AI.”
Sophie Williams, Principal Data Scientist at Barts Life Sciences, also took part in the panel discussion and agreed that building the right data foundations for NHS AI systems is critical.
“In terms of diversity and coverage, the UK probably has the best health data in the world,” she says, meaning there is huge potential for personalized care. Stated. But there needs to be a process to ensure that the data fed into AI models is fair, accurate, covers everyone, and is an appropriate dataset.
“When statistics are misused, they can tell a different story than the actual truth,” Williams said. “You have to be very careful about that. What works in one population won’t work in another.”
Efficiency and cost savings
Once AI is deployed, there is great potential for efficiency and cost savings. Another panelist, Helen O’Neill, CEO of Hertility, He cited the example of a London hospital that spends a million pounds just to send letters.
Digital self-service is another key aspect where technology can further improve the NHS, allowing people to enter their personal data, symptoms and conditions.
Leveraging the strengths of different groups means giving data- and digital-enabled people the ability to advance reservations and enter information before arrival.
“An average nine-minute visit is not enough time to ask all the relevant questions and for the patient to answer them all and get a diagnosis,” O’Neill says.
“And what you get is this fundamental culture of blame. [complaining] The doctor didn’t prescribe it to you, but the doctor saw 18 people in the past few hours. Expecting humans to give us answers based on so many complex symptoms is fundamentally flawed. ”
Asking basic questions rather than diagnosing symptoms not only saves appointments from being wasted, but also gives you the opportunity to collect more truthful data. O’Neill said Hertility was able to collect such a truthful data set by using a digital interface to provide encouragement, validate the reason for data collection, and explain how the data will be used. I did.
“When we look into the eyes of a stranger, we tend to give different answers, especially when it comes to intimate issues,” she added.
Digitization and self-service will also make the NHS more convenient for modern life. Mr Tan pointed out that for many people who work multiple jobs, trying to take half a day off just to go to the doctor is no longer part of our world.
One of the major hurdles Tan and her team need to overcome when it comes to AI development is talent acquisition and retention. “The hard part is we’re thin,” she says. “We don’t pay them properly for those skills.
“We need to be a little more creative in how we get people to work with us,” Tan said. “So, we do rotations, we join Girls in Data, we participate in practicums, but we also develop ourselves. We are considering it.”
