Protocol Manual: Health Care

Pivoting to COVID-19: How a weekend sprint helped one startup boost adoption

“All this infrastructure out there now that doesn't really exist for health care," says AccuRX CEO Jacob Haddad on the health tech API ecosystem.

Doctor holding a phone

AccuRx is now used by 98% of GP practices in the U.K.

Photo: National Cancer Institute/Unsplash

At the beginning of March, the team at AccuRx, a patient communication company, realized that COVID-19 could cause a big problem for doctors in the U.K. "It was seen as a foreign thing," CEO Jacob Haddad told Protocol, but when talking to doctors one Friday morning, he realized they "were starting to get worried." It became clear that they needed a video appointment solution — and AccuRx, which already provided SMS communication services to U.K. doctors' surgeries, was well-placed to build it.

That afternoon, the team built a "hacky prototype," as Haddad called it, and by Monday morning, they had a live product. "By Monday afternoon, we turned it on for half of the practices in the country," Haddad said. "And then over the next three weeks, we had the other half of the practices in the country, essentially, sign up." AccuRx is now used by 98% of GP practices in the U.K.

Speaking with Protocol, Haddad discussed how the company managed to roll out video consultations so quickly, his views on the future of health care, and how AccuRx overcame institutional roadblocks.

This interview has been edited for clarity and length.

What does AccuRx do?

We think health care is a communication industry. That's what everyone on the front line is doing every day: sending referrals, coordinating care, getting information to patients, getting second opinions. The value in health care is added in communication.

Our vision is that anyone involved in a patient's care can communicate with each other, which is very uncontroversial — no one thinks that's a bad idea, but it's really bloody hard to do. You've got everything from legacy technology, to the network effects problem of how do you get all these different people across different organizations using us, through to there being different incentives in the system.

We've started in [primary care] practices, making it very easy for them to communicate with their patients through SMS messaging. And then over the past few months, we've really enhanced that as a result of moving fast in response to COVID, so everything from video calls, to patients being able to respond, to being able to send documents to patients and all sorts of things like that.

How did you build the video tool so quickly?

We had an incredibly large and strong amount of infrastructure already: Everything from how we deploy releases, to how we update our software, how we integrate with the medical record, all of that sort of stuff we had in place. It was more a case of joining up a third-party video service API with our existing messaging, and then doing some user testing to make sure that the flow works. The main constraints were actually, at the time, most NHS computers didn't have webcams and microphones. [In] the first version, you could only do it by sending yourself a link to your mobile phone. So you would initiate it on the computer, but the clinician would take part from their personal smartphone. That was how we started

How common is it to use third-party APIs in health tech?

It's not common in health care. You can't move as fast, as you can't just plug something in and go. A lot of that weekend wasn't just spent building the tech, it was looking at where's the data going, how is it encrypted, is any of it stored? (The answer's no.) You've got to have those considerations, and that massively limits the amount of third-party services that you can plug into.

There isn't a rich ecosystem. If you were going to build another Uber, you'd use Stripe, Twilio and Google Maps — all this infrastructure out there now that doesn't really exist for health care. We were just quite fortunate that video was this one space where, because it's encrypted but none of the data is stored, the third party has no other information about the patient.

Is that third-party ecosystem developing at all?

To be honest, not really at the moment. We'd love to provide APIs to third parties, but we've got to focus on our own product first and get that standing on its own two feet. One of the challenges so many health tech products have is commercialization. At the point where more health tech products can commercialize, I'm sure there'll be a much bigger market for third-party infrastructure. But at the moment there isn't, because there isn't really a vibrant market of health tech. [There's] some consumer health tech, but in terms of real enterprise systems, a lot of it's quite legacy: classic field sales, long implementation, infrequent releases. It doesn't work like SaaS works today.

The U.K.'s NHS has a reputation for being slow and unwieldy, yet you've managed to get into 98% of primary-care practices. How did you get around those institutional roadblocks?

We're big believers in bottom-up, self-service adoption. We don't think anyone knows better what their user needs are than users themselves. The NHS isn't bad at adopting technology, it's bad at adopting bad technology. But if you can actually empower frontline staff to set themselves up and pick the tech they need, we found that they'll do it. Doing self-service in GP practices has worked relatively well, because they've got more autonomy. Doing that in hospitals is going to be trickier for us, but we're already at over 15% of hospitals in the country now using us, and zero were before March, so that's also moving quickly, too.

Did you ever bother with a top-down approach?

We've tried working at that level with the Department of Health and NHSx [the NHS' IT division], but we've really struggled to make any progress at any sort of pace. We're very happy to and keen to engage at that level. But every time we've done it, we've just gone back to self-service — [it's] the best way to get software that works in the hands of users, in the hands of patients.

What's going to stick around from this forced telehealth transition?

I think specific parts might decline a bit, video consulting being one example. In general, the shift to more remote and in particular, more asynchronous communication — which video isn't — is definitely going to stick around. Video gets a lot of attention at the political level: It's seen as this very futuristic technology. But in reality, having asynchronous ways to communicate is often far more disruptive — and not just disruptive, but useful. So I'm sure that's going to stick around long after.

Do you have plans to bring AccuRx to the U.S.?

Our mission to bring patients and their health care teams together is very much a global one, be it developed health systems like the U.S. and Canada and Western Europe, but also developing ones, too. Now, there's a lot deeper for us to get in the U.K.: We're still a relatively early-stage company. And so it's going to be a couple of years before we're able to [expand internationally] without massively distracting from the problems we're solving here. Health care is just so full of inefficiency to solve that actually one of the biggest challenges is really having a good focus on the right areas and the right problems.

The U.S. health system is very different from the U.K.'s. Will that inform your product thinking at all when you expand to the U.S.?

A lot of health systems are very different, but in particular I'd agree, here and the U.S. The way we think about it is there are some things that are much more common between health systems: typically the types of problems people have with their health, how they want to engage with the health system, things like that.

And then there are some things that are very different. A big one is incentives: In the U.S., primary-care clinics are very much incentivized to have more activity, because they get paid for it. Whereas in the U.K. it's the opposite, you want to have less because you don't get paid on activity, you just get paid for the patients registered. I think that will definitely shape the type of products.

It is not clear yet in the U.S. how, for example, resolving a patient's query remotely rather than having their appointment is going to be reimbursed. That's something that we're interested to not just see play out, but be a part of that evolution. Those are some of the sorts of things we're thinking about between health care systems: the incentives, the behavior, the culture around health care regulation.

Next in Protocol's Health Care Manual: A special edition of Protocol's Braintrust: What it's like to pivot a health startup in a pandemic?

More from Protocol Manual: Health Care