This algorithm is making diabetes more manageable

For diabetics, getting insulin doses right is a serious matter. The Quin app can predict when and how much they need

Cyndi Williams never meant to get into healthcare. A chemical engineer by training and a one-time software engineer by trade, she’d spent much of her career on the business side of tech, working with Fortune 500 companies in retail, media, telecoms and finance. But in 2014, an unexpected conversation with a colleague and mentee at global software consultancy ThoughtWorks changed her trajectory.

During one of their bi-weekly mentoring sessions, Isabella Degen revealed that she was one of the 400,000 people across the UK living with Type 1 Diabetes, a disease that required her to inject insulin multiple times a day, working from crude formulas and personal experience to determine how much to dose and when. She was leaving the company to create an app to make that easier for others in her situation – and she needed help.

Williams’ interest was piqued. “She made me aware of the challenges that exist for self-management, and the hundreds of decisions a day people are making to keep themselves alive and well,” she says. “I saw this incredible possibility to bring what we were using in modern consumer technology, retail and media, and applying it to this very specific human need.”

That year, Williams and Degan founded Quin. The app tracks users’ data and contextualises it in order to produce highly personalised recommendations that take the guesswork out of regulating insulin levels.

“It's all about curation and really putting the right information in front of people at the right time to stimulate engagement,” says Williams, who serves as CEO. “We have all these really fancy technologies that we've used in entertainment, media and retail to really engage people – we just need to apply them.”

Williams is banking on this hyper-personalisation model to set Quin apart from competitors in the burgeoning mobile health market, predicted to be worth $189 billion by 2025. And, so far, the market seems enthusiastic about her approach. Since its launch in 2014, Quin has received £3.6 million in investor funding, and more than 17,000 users have downloaded the iOS app since it was released in the UK and Ireland last October.

Quin – a portmanteau of “quantifying intuition” – relies on a combination of predictive algorithms and personal data, which can be collected automatically or added manually. To get started, users input information about their food intake, insulin doses, activity and blood sugar levels over the course of their day. Once the app has enough information about how certain factors affect the user’s blood sugar levels (Williams suggests three weeks of daily usage to see benefits), it’s able to suggest specific insulin dosage amounts to regulate them, and graph how the user’s blood sugar levels are likely to change over the next five hours.

“So, say I'm going to eat a burger; I've eaten three burgers in the last year and I've dosed anywhere from two to six units of insulin for those burgers,” Williams says. “Quin will ask, ‘What's your starting blood glucose? How much insulin do you have on board already? How active have you been in the last 24 hours?’ and then say ‘OK, four units is probably the best for you right now, given all these other factors that we know about you.’”

Part of the secret to designing an app people actually want to use, Williams explains, is creating the product with them: since 2014, Quin has consulted more than 300 diabetics as part of their research programme.

“We had to spend time [asking], Why are people using apps? Do they want an app? What would an app that they would use do?” she says. “The people we're working with are taking a drug that, if they take too much, it can kill them; if they take too little, it can kill them. There are very serious, potentially catastrophic risks here, so we have to be respectful of that.”

As Quin prepares to launch in the US at the end of this summer, the company is looking into measuring and tracking other physiological, psychological and behavioural data, such as sleep and stress, to provide more holistic recommendations. But where social media and entertainment apps are designed to keep users checking in as much as possible, Williams hopes that, over time, Quin users will feel confident enough in the app’s recommendations that they find themselves opening the app less and less.

“We actually don't want to be an app that's just sucking you in all the time. We just want to be useful at the moments when you need to make decisions,” she says. “If we can give you two or three hours where you can just be like, ‘OK, I'm going to be fine based upon what I see on Quin, I'm not even going to check in…’ then that's a wonderful thing.”


More great stories from WIRED

This article was originally published by WIRED UK