Getting to the Heart of the Matter: Big Data in Healthcare

May 23, 2016 - by Art Shectman

“It is a capital mistake to theorize before one has data.”      -- Arthur Conan Doyle

This is Part 2. See Part 1.

Big data can be enormously useful to American healthcare. It can make sense of the enormous volume of data the healthcare system generates, and it can use that data to provide actionable insights into options for improved healthcare delivery.

There are, however, obstacles to implementing a big data approach that go beyond the simple issue of the enormous amount of data the system can produce.

Some of those obstacles are the expected ones that arise whenever a big change is afoot. Providers, for example, have entrenched habits and established workflows, and they don’t take kindly to disruption, especially disruption that lacks immediate, individual and tangible benefit.

But that’s a problem inherent in any big change, and it’s not insurmountable. New systems take some getting used to, yet practitioners will adjust as they always have.

Other problems go a little deeper, but they, too, can be solved.

First, the size of healthcare data is not its only distinction. It’s also distinguished by the vast variety of kinds of data the system generates, an almost unholy mix of text, numbers and images in every format imaginable. In itself, this doesn’t make for a fatal problem, but it does mean that systems must be exceptionally robust, fast, efficient and adaptable.

We're building the next generation of health information management systems with our client Vigilias to create a modern telemedicine practice.

At Elephant Ventures, we're building the next generation of health information management systems with our client Vigilias in order to create a modern telemedicine practice. Our healthcare system requires access to quality data and the ability to move data among systems accurately. We're applying technology and process innovation to make it happen.

And that's only the beginning. Add in the health data that the Internet of Things will generate, all flowing from a variety of wearable and portable monitoring devices that don't necessarily speak the same language, and the demands on the system become even greater.

Managing vast amounts of data is an area in which we’ve had considerable experience, and the problems of working with disparate data are problems that Elephant Ventures has been called on to solve in several contexts. The solutions always call for wide experience and deep expertise, but the real point is that this problem, even if it looks overwhelming from the outside, is hardly intractable.

If the systems in use are not fully up to the task, however, problems are likely to be reflected in provider acceptance, and that’s a true cornerstone of the big data approach. It’s not just that medical practice frequently needs information immediately. Not everything medical is an emergency, but providers in general are already pressed for time. If the systems can’t keep up, providers will actively resist.

We know that we can solve that problem by designing and implementing the right technologies, but that brings us to a problem that stems from the nature of the American healthcare system itself: It’s a fragmented system with multiple independent participants.

Healthcare data is generated by a long list of sources, including providers, hospitals, pharmacies, insurers and, through Medicaid and Medicare, the state and federal governments. Each of those sources has its own systems for dealing with the data it collects, and those systems weren’t built to interact with each other.

As one would expect, that makes interoperability an ongoing problem. The situation is further complicated by the fact that each player has a vested interest in its own systems and its own data. If that data has value – and it does – it’s asking a lot of one of the big health insurers, for example, to open its doors to the competition and give free access to data that’s the product of significant private investment.

That complication finds its ultimate expression in the form of the Kaiser organization, an organization that’s involved in multiple aspects of the healthcare system. It’s a provider, a payer, a pharmacy and a manager, and it has been a pioneer in technology as one of the first companies to adopt electronic medical records. It has also developed a comprehensive system, HealthConnect, that aims to integrate its diverse functions into one central repository. With all that experience, Kaiser could undoubtedly be a source of useful insights into the interface of big data and healthcare, but there’s a catch: Kaiser is a private organization, and we don’t even have the benefit of access to the kind of information a public company must provide.

Big data has one more river to cross before it can really come into its own in healthcare, and this crossing has the makings of a long-term project. While it has two distinct branches, the question here revolves around the use to which the data is ultimately put.

On the one hand, the issue is privacy. When it comes to sharing information, no field is more regulated than healthcare, and potential intrusions on that privacy attract enormous public interest. It is, in fact, a highly emotional subject, and public opposition to the collection of health-related data can be fervent even if appropriately powerful safeguards are in place. This will always be an issue, and the lesson for big data companies is not just to make privacy a priority, but to be seen to be making privacy a priority.

On the other, self-interest can lead to companies’ putting data to their own less-than-noble uses. For example, an MRI manufacturer might comb the data for likely candidates for its machines, and the right marketing might lead to MRIs for patients with, at best, marginal need for the procedure. Patient benefit isn’t the foremost concern. Questions of value to the system as a whole don’t matter. The only value being generated is limited to what the manufacturer gains by expanding its market.

All in all, this makes for a complicated situation, but it’s certainly no more complicated than the immense healthcare system itself. Whatever the obstacles, big data offers more than enough to healthcare to make the battle absolutely and unquestionably worthwhile.