TriplePundit • AI Is Tackling One of Healthcare’s Biggest Problems: Missed Appointments

In the race to build artificial intelligence that detects cancers and diagnoses diseases, few are paying attention to the quiet crises that erode healthcare systems from within. Research from the United Kingdom shows that missed appointments, also referred to as “no-shows,” can cost almost £2 billion annually, extend waiting lists, and in some cases, cost lives.

No-shows usually don’t mean a patient didn’t want to go. People frequently miss appointments because they can’t get there, have to be at work, received unclear scheduling information or forgot about them. Deep Medical, a U.K.-based healthtech startup, is showing that AI can make a real impact on issues like this.

It started as a data project at University College London in 2018. “We were looking at predicting no-shows to radiology appointments,” said David Hanbury, co-founder and chief AI and strategy officer at Deep Medical. “Loads of people were trying to do the sexier side of AI in terms of predicting cancers in CT scans. Very few were looking at the problem of trying to solve operational issues in healthcare.”

The scale of the problem is staggering. In 2024 alone, the U.K. National Health Service experienced an estimated 11.8 million missed appointments, according to research from Esendex, a company that offers automated message services for healthcare providers and other businesses.

“It’s a huge problem,” Hanbury said. “If you miss two appointments in a row and have a long-term condition and a mental health condition, your chance of dying in that year goes up eightfold.”

Deep Medical’s technology draws on 15 years of hospital data and about 200 predictive factors — everything from weather forecasts to public transport access — to predict which patients are most likely to miss appointments and why.

“It would be very complicated for a human to work out,” Hanbury said. “[Missed appointments] are not just because it’s Monday at 9 a.m. but because it’s an 80-year-old patient with a bad knee, it’s predicted to be very cold and it’s 9 a.m. on a Monday.”

When Deep Medical’s system identifies a likely no-show, it intervenes with a personalized response. That might mean sending three reminder messages instead of two, rewording texts based on behavioral cues or offering free transport.

“We have a partnership with Uber Health,” Hanbury said. “So an option is to send a free Uber to patients in order to try and get them to come to the appointment.”

None of that personalization requires patient records, said Benyamin Deldar, Deep Medical’s co-founder and chief growth officer. “Initially, we were looking at medical problems that patients had, but that’s not really ethical for us to use,” he said. “We’ve been able to build really accurate models by not touching the patient records at all.”

Crucially, Deep Medical designed its system by listening to patients. “We were calling up the patients that the AI model was predicting were likely to miss their appointments,” Deldar recalled. “We heard stories like, ‘You’re sending me a text message, and I’ve got visual impairment. I can’t see what you’re telling me.’ Or, ‘I can’t afford to take the time off work.’”

These insights help shape the next iterations of the technology, like the flexibility to rebook missed slots quickly. Deep Medical’s backup booking feature, an AI-informed equivalent of airline overbooking, fills appointment slots that might have otherwise been no-shows. “We unlock an extra, on average, about 45 percent of the remaining no-shows,” Deldar said.

The impact extends beyond logistics. “It isn’t a capacity issue, it’s an operational issue,” he said. “There are 6 million people waiting for a hospital appointment in the U.K., and 12 million appointments that go missed or canceled under short notice and are just not filled. So really, it’s a matching of supply and demand that we’re helping with.”

Deep Medical works with Cisco and Uber to manage its communication and transport infrastructure, building an automated system that takes off of overstretched administrators’ plates. That seamlessness is central to the vision.

“The best technology is the one that fits into our lives as it is,” Deldar said. “You can book a ticket from Dubai and go to Kenya tomorrow, and you don’t need to speak to anybody. The technology exists, so our approach is about using best practices in other industries and translating that into healthcare.”

The startup’s early partners include National Health Service (NHS) hospitals in Essex, Surrey and Hertfordshire. “In hospital systems which have a no-show rate of 8 percent, we have managed to bring that to under 6 percent,” Deldar said.

Mid and South Essex, one of Deep Medical’s early NHS partners, found the pilot results “hugely encouraging,” said Program Director Erica White. “It has helped us to make the best use of our resources as a service, reducing no-shows and supporting the movement of patients off the waiting list,” White added. “Critical to this is being able to communicate with our patients at scale, across multiple channels.”

Part of the tool’s success might be because Deep Medical’s business model is structured to reward outcomes, not just software. “We don’t get paid unless we drive up revenues by capacity that we’ve been able to contribute to our clients,” Deldar said.

Still, one intervention alone cannot address the myriad of reasons behind no-show appointments. While the AI model significantly reduces missed appointments, decreasing them by 30 percent over six months in a 2024 trial, it does not eliminate them entirely.

The company’s next step is expansion across the Atlantic. “We started off in the NHS, which has one of the lowest no-show rates in the world of 8 percent,” Deldar said. “As we translate this over to the United States, where Medicaid has up to 30 to 40 percent no-show rates, being able to halve that underutilization will have a huge impact.”

The team is currently finalizing contracts with a U.S. healthcare provider that serves primarily Medicaid and uninsured patients. “They have a large no-show rate problem: 40 percent of their patients are just not turning up,” Deldar said. “There’s a big opportunity there to really drive impact.”

Deep Medical represents a shift in what counts as innovation in health tech. Instead of chasing futuristic diagnostics, it’s reclaiming wasted time and capacity in the system we already have. As Deldar puts it: “If you’re on TikTok or Instagram, you’re getting a targeted advert, so why should we not be bringing that technology to healthcare where we can improve outcomes?”

Featured image credit: Nappy/Unsplash