Dr. Dan Low loves his job. It shows in almost everything he does.
Low is a pediatric anesthesiologist — he’s in charge of giving children all the medication and anesthesia they need during surgeries.
But in the past few years, Low has also taken on a new job. He is the founder and CEO of a new health technology startup, MDMetrix, that’s aiming to help doctors improve their work by giving them easy access to their own patients’ data.
In a nutshell: A/B testing analytics for healthcare.
The idea came from a painful experience in Low’s professional life. In 2016, his hospital, Seattle Children’s, started using a new drug in one of its most common surgeries. Low’s boss asked him to figure out if the new drug was working better than the old one.
“It seemed like a really easy problem to solve, initially, because in the last five years the hospital has become 100 percent digitized,” Low said on the most recent episode of GeekWire’s Health Tech podcast. “So that data lives somewhere. It exists [in] a digital format. We thought it would be really easy to just call up someone and ask them to surface that data.”
It was the opposite of easy. The process started with Low sitting down with an analyst, who translated his clinical question into a computer program. The analyst queried the hospital’s database and sent Low a huge spreadsheet of raw data.
Hear Low’s full story on the most recent episode of our Health Tech podcast. Listen in the player below or subscribe to Health Tech in your favorite podcast app.
“Inevitably, what comes back is not quite what I think it would be, so then there’s a bit of back-and-forth,” Low said.
When he has the correct data, Low then has to go to a different hospital employee: A statistician. Low has to then re-explain what all the data is and what question he’s trying to answer. Eventually, the statistician analyses the data and sends Low a report.
The whole process can take 15 to 20 weeks, and that isn’t because Low’s hospital is doing a bad job: It’s a problem that persists all over the healthcare industry.
It turns out the new drug was about as effective as the old one and the hospital didn’t change anything. But for Low, the experience pointed to a need for a different kind of change.
“I don’t want to make an intuitive decision whether to adopt something in my personal practice of medicine. I want to make a data-based decision. And often that data is not there — except it is. All the data is there if you know how to look for it,” Low said.
So why does the industry have this problem in the first place?
“Healthcare data isn’t necessarily easy to use,” said Jim Harding. Harding is a longtime technology executive who has led teams at Amazon and is now the CEO of healthcare data company MultiScale Health Networks.
“It’s highly regulated, for one thing, and people that work with are somewhat scared about it when they come into the industry,” he said.
National laws regulating how healthcare data can be stored, accessed and shared make working with the data much more difficult than would be in other industries, Harding said. But there’s another problem: The basic form of the databases.
“Historically, the databases that some of the more popular [Electronic Health Records] are based on are hierarchical databases, not relational,” Harding said. Hierarchical databases are harder to work with and there is very little analytics software that works on this kind of data.
That technical and legal complexity means building products on healthcare data is extraordinarily difficult, but as the industry becomes more concerned with improving outcomes that difficulty is starting to be overcome.
MultiScale is one example. The company has built a cloud-based healthcare data platform that lets hospitals securely store data where it is easily accessible, and even securely build applications on top of it. The company is focused on improving operational data, information that care teams of doctors and nurses need in the moment as they are working with patients.
“We can’t change what we can see,” MultiScale Chief Product Officer Stacey Kinkead said. “The care teams themselves can’t change where they can’t see. So we allow them to see the data, communicate and collaborate on the data and then problem solve all together to get that patient better. To get them into that room, to get them that CT scan. So much of it in healthcare right now is so basic.”
So why not use that data to evaluate how well medical care works? That was Low’s question. He started asking around the hospital and eventually teamed up with some of Seattle Children’s executive leadership, including the organization’s chief data officer and veteran tech executive Ryan Sousa.
In spring of 2016, Low sketched out his vision for MDMetrix on the back of a Starbucks napkin. He teamed up with his co-founder and CTO, software engineer Matthew Goos, and got to work designing the software. Straight from the start, Low had a rule about the software’s design.
“It has to be as easy as VRBO,” he said. At the time, his family was planning a vacation and his 12-year-old son was able to book a rental home through VRBO using a set of criteria: Three bedrooms, allows pets, and so forth.
“So accessing the [hospital] database — although there’s thousands of thousands of patients — can you segment that as easily?” Low asked. It turns out you could.
“Say I’m interested in 4 to 16 16-year-olds who had this type of surgery, who had this drug, who has looked after between these dates. And I want to know what the outcome metrics were. I want to know how long they stayed in hospital afterwards. I want to know what the narcotic requirement was afterwards,” Low said. “That’s what we now have the ability to do.”
MDMetrix’s software has launched in Seattle Children’s Hospital and the startup has become Seattle Children’s first spin-out company.
But throughout all of this, Low still hasn’t quit his day job — and he doesn’t want to.
“I didn’t want to be a CEO,” Low said. He just wants to do his job better. In fact, he’s looking for someone else to take over running MDMetrix so he can focus on being an anesthesiologist, helping kids every day.
Kinkead said it’s doctors like Low that are going to transform the healthcare industry, not technologists.
“Do you want to disrupt healthcare? I don’t think so,” she said. “We don’t. We want to allow the providers to disrupt themselves. I get that healthcare is in a change-or-die type of situation. But the last thing you want is a bunch of arrogant technologists coming in from the outside and beating their chests and saying, ‘we’re going to disrupt you.’ I mean, it took me three years to understand what healthcare is. You don’t want us coming in and disrupting — but what we do want is to provide that bridge or provide that new way that allows them to transform themselves.”
Editor’s Note: Dr. Dan Low works at Seattle Children’s Hospital and MDMetrix is a Seattle Children’s spin out. Seattle Children’s is also the sponsor of this season of Health Tech. The organization was not involved in the creation of this story beyond approving our request to interview Dan on the hospital’s campus.