What’s holding back big data in healthcare?

With the countless benefits big data could bring to healthcare, why are we still struggling to adopt it?


Healthcare has come a long way since the days of barber-surgeons and medicinal leeching. The last century has seen incredible advancements in medical science, technology, and pharmaceuticals. Today, the benefits big data has to offer the healthcare industry are truly immeasurable.

Big data has added a new level of understanding to an industry where even the smallest observation can be the literal difference between life and death. As Tom Andriola, Chief Information Officer with the University of California explains, "we now have a tremendous amount of data, let's call it 'digital footprints', around our patients. Their name's from the electronic health records, the medical images that are captured in digital form, as well as a significant amount of genetic information that some patients now come in with. This is information they may be getting through their Fitbit or health apps on phones."

This abundance of data has the potential to fix one of the foremost problems in healthcare - waste. The Organisation for Economic Cooperation (OECD) estimates one-fifth of global healthcare funding is wasted, and no other country wastes nearly as much money on healthcare as the United States of America. According to the Commonwealth Fund, the US spends about $3 trillion a year on healthcare, so finding areas of waste might save US citizens billions of dollars.

Visit Innovation Enterprise's Big Data Analytics in Healthcare Summit in San Francisco on October 15–16, 2018

And using big data to save money doesn't necessarily mean worsening patient care either. More often than not, it means the exact opposite, "All of that data is being brought to bear and really doing a lot through machine learning. Not only are we using machine learning models for diagnosis, but also in the creation of pathways for patients such that we can understand the trajectory that they may be on. For example, as you know, one of the big health issues that the United States is struggling with is the growing percentage of people who are living with a chronic disease, such as type 2 diabetes. It actually has very little to do with genetics and a lot more to do with social determinants and choices, like our diet and whether we smoke or exercise regularly. This means we can use the data to understand which patients are at higher risk and predict who might be headed for a visit to one of our clinics."

However, despite these obvious benefits, unlike other industries like retail or entertainment, the healthcare industry has been notoriously slow in adopting data-driven solutions, and for frustratingly obvious reasons. Tom Andriola further illuminates:

"I think there's a first battle that industries struggle and grapple with before they start to adopt these types of technologies to drive decision making. This is, what is the value of data and the insights you can glean from it versus our professional experience. It's pretty much a risk-free process with, Amazon when it recommends my next movie or an item for my Amazon Prime wish-list, despite the tremendous amount of data that goes into it.

"However, there's very little impact if that recommendation is wrong. Either I want to buy it or I don't, right? But they learn from every one of those decisions and it makes their next recommendation better. It's a little bit of a different scenario when you're talking about a doctor-patient relationship and a doctor trusting data over himself. How do we feel about a recommendation or diagnosis coming out of a machine learning algorithm and allowing it to make the decision about whether to admit a patient into a hospital bed or change their medication pattern or regimen?

"There's just a much greater implication. Same thing if you think about education and a teacher or professor deciding to change the lesson plan based on the data is telling that professor."

These issues extend past simply asking doctors to take a back seat when it comes to the diagnosis of patients. The other concern is how you accumulate all this data in the first place, especially considering health records are, generally speaking, confidential.

"A tremendous amount of activity goes on around using the medical data around patients." Andriola continues," To create these models for better diagnosis, to talk about medical images and the ability to see on a medical image potential cancers. These models are getting very good with respect to being as good as the radiologist doing it. However, there are other people who say 'wait a minute, let's step back and talk about the privacy associated with this. Did you know that you used 3,000 medical images to train that model? Were those 3,000 patients informed their medical images we're going to be used to train up a model?'

"Where does privacy come into this and where does patient advocacy fit into this model? At university, these are the types of conversations that we are able to start having, one that is not just about progress but also the implications of progress."

However, despite these implications and fears, the reality is, the healthcare industry is inevitably moving towards a more data-driven system. As data gathering becomes both more efficient and interpretable, a concerted collaborative effort between healthcare professionals and data scientists is needed if we are ever to fully reap the bounty of rewards big data has to offer.


Read next:

Why We Need Data Visualization To Understand Unstructured Data