Skip to Content
MIT News feature

Catching Errors

A portable pill scanner invented by Ívar Helgason, SM ’08, María Rúnarsdóttir, MBA ’08, and Gauti Reynisson, MBA ’10, flags medication mistakes at the hospital bedside.
December 18, 2014

On a winter day in 2013 at the Admiraal De Ruyter Hospital in Vlissingen, the Netherlands, a veteran nurse named Bastian Teichert was about to administer a dose of medicine to a patient with epilepsy. But when he used a bedside pill-scanning device that the hospital’s psychiatric ward was piloting, he got a warning that something was amiss, even though nurses had checked—and double-checked—the medication.

Alumni Ívar Helgason, María Rúnarsdóttir, and Gauti Reynisson founded Mint Solutions to commercialize the pill-scanning device they began developing at MIT.

At first, Teichert chalked it up to a technical error. But upon closer examination, he discovered that the pharmacy had given the hospital 500-milligram tablets when the prescription called for a 300-­milligram dose. “The pills looked really alike,” he says. “So the patient had been getting the wrong dosage for a few days.”

Teichert estimates that in the two years his hospital has been using the bedside pill scanner, known as MedEye, he’s seen it flag about 40 errors. Its inventors, Ívar Helgason, SM ’08, María Rúnarsdóttir, MBA ’08, and Gauti Reynisson, MBA ’10, say that in hospital pilot tests in the Netherlands, the device has caught errors—some critical—in roughly one out of every 10 scans. “There are too many mistakes,” says Teichert. “The MedEye is an extra control on medication, so you give the right medications to the right person, at the right time, in the right doses.”

According to the 2008 Commonwealth Fund International Health Policy Survey, about 6 percent of chronically ill adults in Dutch health-care facilities reported being given the wrong medication or dosage while being treated. A 2002 study found errors in 19 percent of doses given to patients at 36 U.S. hospitals and skilled nursing facilities. And a 2006 report from the Institute of Medicine noted that a U.S. hospital patient, on average, is subject to at least one medication error each day, with most errors occurring during the prescribing and administering stages. The report estimates that these errors lead to about 400,000 preventable cases of injury or death each year, resulting in roughly $3.5 billion in extra medical costs.

Such troubling statistics are what prompted the three Icelandic alumni to develop their portable medication verification device, which relies on computer vision technology. “It’s a complicated chain of events that leads up to medication mistakes. But the bedside is the last possible place to stop these mistakes,” says Helgason, who cofounded Mint Solutions with his partner, Rúnarsdóttir, and Reynisson to develop and commercialize MedEye.

To use the device, nurses scan a patient’s wristband with a handheld scanner to access electronic records and then push the prescribed pills into ­MedEye’s sliding tray. Inside, a small camera quickly scans the pills, recognizing size, shape, color, and markings (such as letters, numbers, and logos). Algorithms then seek out a match for the pills in ­MedEye’s database—which has descriptions of nearly all pills in circulation—and cross-reference the scanned pills with the patient’s medical records. Within seconds, results pop up on software that runs on a connected laptop. Green or red boxes appear next to the medication names to indicate whether they’re right or wrong.

“We want the device to be the nurse’s best friend,” says Reynisson, 38, who is Mint’s CEO and oversees the startup’s Amsterdam headquarters. (­Helgason and Rúnarsdóttir live in Reykjavík, where ­Helgason, 43, runs the startup’s research and development efforts; Rúnarsdóttir, 36, now serves on Mint’s board.)

Helgason, who has an MD, and Reynisson, a software developer, met in the early 2000s as employees of the Reykjavík-based biopharmaceutical company deCode Genetics. Both went on to work at TM Software, where they helped Dutch and German hospitals set up electronic prescription systems and other ways of making medication errors less likely.

They found that although hospitals depend on nurses to catch errors caused by such things as illegible handwriting or confusion between similarly named drugs, mistakes slip through because nurses are busy and often overburdened, and there’s a lot to keep track of. Patients at the Admiraal De Ruyter Hospital, for instance, sometimes get up to 10 different pills at once, and their prescriptions can change daily.

Systems that require nurses to scan both a patient’s wristband and the bar codes on each pill container were lauded as solutions; in the United States, the FDA ruled in 2004 that certain drugs must have such codes. But the major hurdle, Reynisson says, was—and still is—implementing the systems. It’s especially difficult (and expensive) for small and medium-size hospitals to get the necessary software and scanners to work together, he says: often they’re made by different companies. And bar codes on pill bottles are sometimes difficult to scan, frustrating nurses and forcing them to bypass the system.

Eager to tackle such problems, Helgason quit his job in 2007 to enroll in the Harvard-MIT Health Sciences and Technology master’s program. He and Rúnarsdóttir, who had begun her MBA at Sloan the year before, were two of about six Icelandic students on campus. “Per capita, though, we may be the biggest international group,” he jokes, noting that Iceland’s population is little more than 320,000. The couple further expanded the local Icelandic ranks when their daughter was born during Helgason’s first semester at MIT.

That fall, one of Helgason’s classes took him to MIT’s Computer Science and Artificial Intelligence Laboratory, where he became convinced that 3-D object recognition could be used to minimize drug errors. If computers could identify objects on the basis of various characteristics, he reasoned, they ought to be able to discern pills, which are, by law, all required to look different.

Rúnarsdóttir was just as excited about the idea as Helgason was. Inspired by MIT’s entrepreneurial culture, they found time between studying and caring for their newborn to brainstorm about how a pill-scanning device might work. “There wasn’t much sleeping that semester,” Helgason says.

Helgason also persuaded his former colleague Reynisson to join their quest to make such a device a commercial reality. “Ívar called me one day and said, ‘Gauti, you have to come to MIT: everyone’s starting companies,’” Reynisson says. Reynisson, who would write the early object-recognition code for MedEye, had been programming since childhood, and when he was growing up in the small coastal village of Ísafjörður in northwest Iceland, he had dreamed of attending MIT. By the time he moved his family to Boston so he could begin at Sloan in 2008, he was fully focused on starting a tech company. “I knew MIT would give me a push to do that, even if traveling with a wife and four kids wasn’t exactly what I had planned as a teenager,” he says.

Leaving a career behind was an economic risk that was soon exacerbated by Iceland’s financial crisis. About a month after Reynisson arrived at MIT, all three of Iceland’s major privately owned banks collapsed, leading, among other things, to the dramatic decline of the króna. “I was in a finance class and the lecturer started the class by showing a graph of the exchange rate of the Icelandic króna to the euro. By the end of the class it had gotten significantly worse,” Reynisson says. “Watching the crash live was a strange experience.”

It also increased his determination to get the business background he’d need to make the MedEye a success. His classes on sales, marketing, and customer service proved especially helpful, he says, in spinning medical hardware from the lab.

In 2009, Reynisson entered Mint Solutions in the MIT $100K Entrepreneurship Competition while Helgason and Rúnarsdóttir were back in Reykjavík designing a prototype. Although the team didn’t win any prizes, Reynisson recalls the competition, which emphasized strong collaboration among classmates, as a defining moment at MIT. Carter Dunn, MBA ’10, a fellow competitor and friend, took such a shine to MedEye that he helped the Mint team refine their pitch and business model and “seemed to spend more time on our project than his own,” Reynisson says. Today, Dunn is Mint Solutions’ chief operating officer.

By 2010, the team was ready to test a MedEye prototype they’d constructed with off-the-shelf parts. They visited a Dutch hospital that needed to identify about 250 small, white pills that all looked similar. “We tried them all in our prototype … and it worked,” Reynisson recalls. Five years later, MedEye has been refined into a product that’s ready to hit the market.

With $6 million from its most recent funding round, Mint Solutions is now working with a Dutch health insurance company to expand MedEye beyond the two hospitals in the Netherlands in which it is now being used. The startup also plans to introduce the device at 15 hospitals across Belgium, the United Kingdom, and Germany. And Reynisson expects that MedEye’s U.S. debut won’t be far behind: “We’re in conversations now with U.S. hospitals on how to best implement the device there,” he says.

Keep Reading

Most Popular

Large language models can do jaw-dropping things. But nobody knows exactly why.

And that's a problem. Figuring it out is one of the biggest scientific puzzles of our time and a crucial step towards controlling more powerful future models.

OpenAI teases an amazing new generative video model called Sora

The firm is sharing Sora with a small group of safety testers but the rest of us will have to wait to learn more.

Google’s Gemini is now in everything. Here’s how you can try it out.

Gmail, Docs, and more will now come with Gemini baked in. But Europeans will have to wait before they can download the app.

This baby with a head camera helped teach an AI how kids learn language

A neural network trained on the experiences of a single young child managed to learn one of the core components of language: how to match words to the objects they represent.

Stay connected

Illustration by Rose Wong

Get the latest updates from
MIT Technology Review

Discover special offers, top stories, upcoming events, and more.

Thank you for submitting your email!

Explore more newsletters

It looks like something went wrong.

We’re having trouble saving your preferences. Try refreshing this page and updating them one more time. If you continue to get this message, reach out to us at customer-service@technologyreview.com with a list of newsletters you’d like to receive.