You can go to the hospital and see inside your body with different kinds of imagery. Surgery can be performed with pinprick holes and lasers. What used to kill us can now be treated in an afternoon thanks to technology.
But somehow, data science and predictive analytics in healthcare hasn’t changed much in the last 50 years. While the technology and its applications have evolved, sources of data remain limited, with your doctor still making predictions about your future health based on what you report when you’re in the office twice a year. Unfortunately, healthcare has been slow to adopt data science leveraging external sources to describe the social and environmental factors that play a massive role in consumer health outcomes.
In this episode of Identify Revolution, Derek Rucker, co-founder and board member of Carrot Health, talks with host Fred Pfeiffer about technology disruptors in the healthcare industry, the shifting focus on patient privacy, and the power of predictive data to help people stop diseases before they develop.
The democratization of data science allows more people to engage with data in meaningful ways
“Pretty much anyone with fairly basic programming expertise now has the opportunity through open source libraries and publicly available tools to get in and start doing some pretty hardcore data science. This is incredibly awesome and allows so many more eyes on good datasets.
Now the downside is, you know, that also creates a whole lot more bad analysis because anyone can jump in, and until you know what you’re doing, it’s quite possible to make some bad assumptions. So you have to be a little bit circumspect there, but all in all, I think that it’s a net positive.”
The heightened awareness of the inherent bias in data is disrupting the way data is used
“There is no such thing as unbiased data. Every single data set is biased. It’s a thing, and that’s not necessarily a problem, but it’s something that people need to be aware of and take into account when they’re using it to predict the future.
You can only use a data set to represent how it was collected and expect to get good results. If you’re looking to use it and it moves in a completely different direction, you’re going to have to be a little creative. That’s something that people are starting to realize now, especially with an identity-based analysis. So we have to be very careful about how we use datasets that were created in the past, that would then project forward to a future that looks exactly the same, which isn’t necessarily what we want.”
Most tech startups aren’t about inventing something new; they’re about using something existing in a new way
“Most of the models that we applied had been created elsewhere, but we were taking these different approaches that had been used in various other fields and starting to apply them to healthcare data in the healthcare space. If you look at startups across the board, oftentimes, it isn’t necessarily new tech or even, um, you know, new algorithms or new inventions. It’s taking something that has been done elsewhere, putting a different spin on it, and using it in another way.”
Regulation changes will impact the healthcare industry’s data and technology, but it won’t be as bad as everyone thinks
“They are coming, and they are going to affect your business. That being said, I wouldn’t worry about it too much. So long as you’re on top of it and realize that everyone’s going to be playing in the same field, it won’t be too bad.
Clearly, your models’ level of predictive power will go down if you can’t use certain types of data. But I think that’s just something that you need to be aware of, watch for and then execute to and realize, as painful as it is, everyone else is in the same boat. It’s not going to destroy the industry, even if perhaps your results won’t be as good as they used to be, because neither will anyone else’s.”
The quality of tech you have as a startup is nothing compared to the quality of your team
“The biggest challenge and risk that you’re going to face is not going to be technological. It’s going to be people. It’s going to be emotions, egos, and interpersonal relationships, be it within your founding team with your employees or with your customers. That’s where startups fail.
I would rather have a good team, a good, humble team, intelligent, capable, but bad tech and bad business ideas than great technology, brilliant business plan, and an arrogant ego-driven team because the first one will find a way to pivot and meet the market where it needs to be. And the second one will drive the market away, even if they were there, to begin with. “
[8:08] “If you can get 1% of people who are likely to be diabetic to get off their butt, get in there and start working out and change their dietary habits, you have saved tens, if not hundreds of millions of dollars for your insurer. Not to mention the fact that you have given them decades of extra life and a far better quality of life they wouldn’t have had before.”
[9:25] “Who you are, how that affects your shopping basket, and what else you can be advertised to potentially purchase based on what you have purchased already. That’s retail. What about healthcare? How about looking at the characteristics of the healthcare plan you would be most interested in having? What would get you to buy a different plan? What would actually get you to engage with some of these behavior-altering initiatives? How can you actually get that movement?”
[17:25] “One thing I will tell anyone that’s listening to this: Before you go out and build your [data tools], do a really good search ’cause chances are someone’s already built it. Do you have any idea how many times I have created something I’ve been really proud of myself and then realized like a week or two later that someone has already done it and usually better than I did? So don’t create your own tools anymore. There’s so many out there for your use.”
