Dr. Navdeep Tangri, developer of the Kidney Failure Risk Equation (KFRE), which uses laboratory data and demographic information to accurately predict the risk of renal failure in patients with chronic kidney disease (CKD), was recently appointed to the Viewics advisory board. Dr. Tangri is a medical researcher in the Chronic Disease Innovation Center at Seven Oaks General Hospital and an associate professor in the University of Manitoba Faculty of Medicine. His research program focuses on improving clinical decision making for patients with advanced CKD.
I was intrigued by his background and wanted to know more about him, and how he came to develop the KFRE. Here is part two of my interview with him. (Click here to read part one.)
Heidi: With your background, how do you see machine learning, data science, and predictive analytics impacting healthcare delivery, and do you have any examples?
Tangri: I think to date we’ve had some innovation in EMRs, but it stopped short of its true potential. For some reason, EMRs have progressed to the point where they’ve become really good data warehouses, but they’ve not made efforts to become intelligent. So they house tons of data, but they don’t analyze it to any ability. I think that machine learning, data science, and predictive analytics can really turn data warehouses into intelligent clinical decision support systems.
The data that’s in the lab information system (LIS) and the EMR can be used to identify patients and risk-stratify them into different blocks. A slightly cynical view that doctors take sometimes is that there’s a percentage of patients who are going to do really poorly no matter what you do. And then there’s a percentage of patients who are going to do really well no matter what you do. For example, if you take someone who’s young and who’s got a viral illness or even a small bacterial illness, they’re just going to get better. And then there’s somebody who’s, say, 100 years old and has multi-organ failure; no matter what you do as a doctor, they’re not going to do well. And there’s these people in the middle, where your healthcare delivery can actually change their trajectory. It’s really important for doctors to identify those people with the modifiable risk and actually modify it for the better.
Heidi: Would you say that that’s the majority of people?
Tangri: I think that’s a good chunk of people who get admitted to hospitals. And a good chunk of people who are probably in ambulatory care, where if you knew, for example, that this older patient had a very high risk of readmission in the next 30 days, you may want to arrange some home care or close follow-up for that patient.
The other thing that I think has to move in sync with all of this is reimbursement, pay-for-performance, and those types of considerations. For example, if the payer — an insurance company or Medicare — penalizes readmission in 30 days, that should become the driving force for people like us to develop solutions that tell the hospital that patient X, patient B, and patient J are at very high risk of readmission. Change their care pathway to make sure that they get very close follow-up when they go home. So, I think that those are the two things that need to work together, and that’s how I see machine learning, data science, and predictive analytics making an impact. The only way to identify patient X, patient B, and patient J is to use the data that’s in the EMR and the LIS.
Heidi: Could you tell me a little bit about this private venture group that you’re involved with that’s doing research on analytics for other things besides kidney disease?
Tangri: One of the things that we are working on is adverse events that happen to patients and how to detect them early. We’re working on seeing if there’s information in the free text that’s written by doctors and nurses and whether it contains clues to detect these adverse events. We’re trying to design a natural language processing algorithm that will go through all the nodes and try to find strings of words and phrases that portend a poor prognosis, that’s not captured in traditional data.
Heidi: Wow, that’s amazing. What’s the group called?
Tangri: The overall group is called the Chronic Disease Innovation Center. It’s a group out of our hospital. So, that’s one of our projects. Another project that we have takes biomarkers that we identify as important predictors of risk and tries to bring them to the bedside. We’re trying to create like a small, let’s say a laptop-sized machine that can measure very specific biomarkers with a finger prick. So, there are different technology applications that we’re working on. The whole group is based on being nimble. In addition, we do health economic consulting, because one of the directors of our group is a health economist. We do a lot of different things that are trying to solve these problems in medicine.
Heidi: So, what attracted you to join the Viewics advisory board, and what do you hope to accomplish as a board member?
Tangri: When you guys originally reached out to me, it was to get my advice on the risk equation and how to implement it. Obviously, I was thrilled and happy to help. But I felt that it would have stopped short of the potential of our relationship, especially given the synergies between my interest in my work and your company. That’s why I really wanted to join the board; because I could see right from our first conversation that Viewics is also squarely focused on providing predictive analytics and using data science, using lab and EMR-based data. And that’s the world I live in on a daily basis. So I thought this was sort of a perfect marriage in that sense.
“I could see right from our first conversation that Viewics is also squarely focused on providing predictive analytics and using data science, using lab and EMR-based data.”
Heidi: I think so, too!
Tangri: Thank you. Something I also hope to accomplish as a board member is to continue to identify opportunities where we can leverage the Viewics platform to develop innovative products that no one else is doing in the industry.
Heidi: Bringing people like you — innovators and experts in your field — into our organization is the only way that we’re going to be able to innovate and get better. And, like you, we’re very interested in coming up with creative solutions to these problems. The IBMs of the world can come up with solutions that do what ours do, but they originally started out as doing something else. At Viewics, it’s a completely different focus than trying to be one of the big guys and do everything. So, it’s very cool and I’m so glad that you’ve joined us and that you see the value of the company. Could you tell us a little bit, in your own words, what value you see Viewics providing to the healthcare industry overall?
Tangri: The 50,000-foot view would be that I hope that Viewics will become an essential component of an EMR and LIS to any major healthcare provider. To me, the goal is that you would have an EMR, you have an LIS, and then the healthcare providers think, well, of course I have to have an intelligence system. I think that’s really the goal, and that’s the value I see Viewics providing. I think that solution is missing. And I think the more innovative solutions we create, the more indispensable this is going to seem.
We’re very excited to have Dr. Tangri on our advisory board. With his knowledge of machine learning, data science, and predictive analytics, along with his experience developing the Kidney Failure Risk Equation, he will assist us in driving measurable value for our customers. We look forward to a long, productive relationship with him.
Click here to read the press release.