Healthcare needs to streamline before it can cope with AI

Healthcare doesn’t need a tech revolution, it needs a process revolution — and there is a lot more foundational work to be done before hospitals would be able to benefit from the addition of generative AI. That is what a panel of healthcare investors and entrepreneurs told us on our webinar this week, looking at the future of hospitals.

With almost every healthcare system in the world under strain because of increased demand, funding shortages and lack of staff, eyes have been turning to technology as a possible way out. Could AI replace clinicians? Or improve diagnostics? Could hospital-grade wearable healthcare monitors mean more people could be treated in their homes rather than in hospital beds?

Around 56% of healthcare provider executives polled by Bain & Company and KLAS Research last year said technology was in their top three priorities, and 80% had increased spending on it.

But simply throwing more technology at hospitals isn’t in itself going to result in improvements. So far, tech spending in hospitals has focused on electronic medical records, says Will Caldwell, chief executive of SaVia Health, a startup that provides workflow software for hospitals. Electronic medical records have increased efficiency in billing and documentation, but, says Caldwell, what hasn’t happened in healthcare yet is the process revolution that has happened in most other industries.

“There is usually one best way to, say, treat a patient with sepsis,” says Caldwell. “The problem is the variation in the care that’s being provided. The question is, how do we create a process or a tool that gives the physician something that helps them adhere to those steps. So, if a patient comes in with sepsis, here are the 14 things that people need to do, and the order they need to happen. How do you use technology to digitise that and push that into the workflow?” he says.

“We’ve done that in every other industry but we haven’t done it in healthcare.”

Healthcare systems are already notoriously difficult to sell into — with sales cycles of upwards of a year — even without the need for a shift in their thinking around the adoption of technology.

Gustavo Cavenaghi, head of investments at Brazil-based Kortex Ventures, recalls a typical example of this tricky sales cycle: “I went to one of the biggest diagnostics companies in the country to understand how they did things. They literally did it on paper. They printed out the exam and then the doctor with a ruler and pen marked it up.”

Cavenaghi thought it would be a no-brainer for the company to work with the startup Kortex was backing, which digitised exactly this kind of manual diagnostics work.

“It [still] took them one year to do a proof of concept to prove the technology was worth it,” he says. “Investing in healthcare is very different than doing venture capital in any other vertical. We don’t see exponential growth, such as you’ve seen in fintech and consumer goods. It’s a whole different animal.”

We asked our panel which, in addition to Caldwell and Cavenaghi, included Tas Gohir, senior intellectual property, innovation and commercial research manager at Guy’s and St. Thomas’ NHS Foundation Trust in the UK, and Eric King, investment director at Intel Capital, to help us unravel the choke-points for healthcare startups and investors.

This is what we learned from our webinar, Innovate to heal: The future of hospital tech:

1. The “hospital in the home” is coming — and it’s good for doctors as well as budgets

The “hospital in the home” concept has been a theme in hospital tech for some time, with remote monitoring technology potentially allowing more patients to be treated in their homes rather than in a healthcare setting.

There’s consensus that more and more of the simpler care will migrate out to the community — some estimates are for around 20% of hospital beds to be in homes in the future — although hospitals will remain important for the complicated and acute cases.

The most obvious driver for the shift is cost.

“Everything that surrounds a hospital is expensive. Labour is expensive, the assets are expensive. Everybody’s time is very expensive. So de-hospitalisation and bringing care to the home takes the burden off of the entire healthcare system,” says Cavenaghi.

Koretex Ventures has backed a company called Isa Labs, which is playing into this thesis. The Sao Paulo-based company started out as a home health testing kit company but has migrated to offering all types of hospital treatments in the home.

“The major clients are healthcare insurers. And the reason why they’re paying them is because it’s a lot cheaper to treat that patient at home or to bring care to the hospital patient than to do that inside the hospital,” Cavenaghi says.

But there may be other more human reasons that moving care out of the hospital is a good thing. It might actually restore some of the bond between the physician and patient, says Caldwell, who is himself still a practicing physician.

“We’re in a stage where we’re all burning out. At three in the morning you’re in the ER taking care of a patient who is intoxicated and they’re not very nice to you. We get bitter. We get cynical,” he says. “When you go to someone’s home, all of a sudden you’re seeing them in their own environment.”

This can help build empathy and understanding between doctor and patient.

“There are a lot of benefits to getting things out of the hospital. Not just technology, home nursing and home care but physicians. It helped me be less cynical about when I do go in at two or three in the morning because it’s now much more personal.”

2. Doctors aren’t the biggest problem when it comes to selling to hospitals

Healthcare professionals have a reputation for being overly cautious when it comes to adopting new technologies.

“I can see the frustrations when companies are coming to approaching us with new technologies,” says Gohir. “I would like to see more technologies taken up for patient benefit. But there is a reluctance and that does need to be tackled. There’s something about the culture needing to change within a medical environment.”

But Caldwell says this is changing.

“When I first started, physicians were cantankerous. We didn’t want to change. We’ve seen that completely go away and healthcare clinicians today are hungry for standardisation,” he says. “When I go in and I try to sell a tool the clinicians are easy.”

3. The finance office is a much more substantial hurdle

“You have to be able to show that you’re going to deliver a near term return on investment. If you don’t, you’re going to get killed in the CFOs office,” says King at Intel Capital.

Anyone wanting to sell technology to a hospital needs to be able to show rapid returns on investment, he says.

“You need to show how you can deliver strong ROI in the near term in the healthcare industry. And and then you have a real shot at making a difference.”

Gohir says startups wanting to sell into the UK’s National Health Service need to think about the issues that keep senior hospital managers awake at night and work to solve those.

“Currently in the UK, it’s long waiting lists, financial constraints,” he says. “Are you thinking critically about solving those problems and not just adding another layer of cost? It is difficult for NHS organisations, any hospitals, to take on new tech. It has to be replacing something else, not just adding to what we already have. Otherwise, you’re adding layers of costs.”

4. The biggest roadblock is the IT department

Hospital IT departments are under pressure, and don’t always have the bandwidth to take on new technology, says King.

“There’s so much financial pressure in these provider organisations that IT organisations have shrunk over the last few years. You have to be a really high priority to get implemented even if they clinically want you,” he says.

And then there’s a related problem. A lot has already been invested in hospital tech but not necessarily in a way that has been helpful to doctors.

“There’s a huge disconnect right now between what the IT functions and hospital think they’re doing to help clinicians with process and what’s actually happening. We have to figure out how to break that down,” says Caldwell.

“We’ve had decades of EMR [electronic medical record] implementation and yet we still hear clinicians say we don’t have the process tools or the clinical decision support tools that we need,” he says. The upshot is that IT departments are working flat out but doctors still feel dissatisfied.

5. Hospitals aren’t ready yet for generative AI

Generative AI is the watchword on everyone’s lips right now, and there’s been much talk of AI helping do everything from improving patient diagnoses to writing doctors’ notes to finding the right people for clinical trials.

But the honest assessment of people who are close to healthcare tech is that hospitals just aren’t ready for this yet.

“I get that AI is changing everything. But healthcare, at least in my point of view, might be a step behind,” says Cavenaghi. “First we need to we need to build processes and we need to digitise those processes. We haven’t got there yet. You have to have the first layer, and what we’re seeing, at least in in Brazil and even in Mexico, is that we lack a lot of infrastructure.”

King is bullish about AI being able to eventually bring in solutions such as virtual intensive care units, where a smaller number of staff can monitor more patients, possibly remotely, with the help of machines. But, he agrees with Cavenaghi that the data infrastructure is still not ready.

“You have to have the data in a usable format, with normalised and time-synchronised data. And this is in that category of things that people think is already happening. But, by and large, it is not happening and is kind of that next wave coming up for the hospital of the future.”

Caldwell, on the other hand, says it comes down to getting that process revolution going in hospitals. “I think AI could be very effective. It can help reduce costs, certainly labour cost. But until we have a digital way, or a better way to create process to get you from point A to point B, only then are you going to be able to plug in these sorts of analytics or tech tools like AI.”

Watch the full webinar here:

This webinar is part of GCV’s The Next Wave series of webinars. We run a webinar on the second Wednesday of every month, alternating between advice for CVC practitioners and deep dives into specific investment areas. Our next webinar will be CVCs and the art of the follow-on on May 8, 2024. Register here to secure your place.