Investing in big data? It might benefit your hospital to think smaller

In less than a decade, internet searches for the term "big data" exploded, growing nearly 100 percent between 2011 and 2016 alone. That trend makes sense if the promises big data advocates ensure will actually materialize — the ability to glean meaningful insights from statistics and numbers that in the past would've amounted to not much more than digital noise. The promise of big data is not only being able to see pictures where they couldn't be seen before, but that there would be pictures that we couldn't have imagined were there.

For hospitals, even small degrees of movement toward that reality could be revolutionary. But what if most healthcare organizations are overlooking the myriad opportunities presented by those smaller degrees of shift, leaving valuable insights that could change care delivery and improve population health unclaimed because they're too focused on the "big" part of big data?

"My point of view for a while has been that hospitals shouldn't do big data — they should do small data," says Munzoor Shaikh, director of healthcare transaction services and population health for West Monroe Partners in Chicago.

Mr. Shaikh says a big data mindset can set hospitals up for data inundation. Collating patient records, financial records, population health records, wearable and fitness tracking records, among many other datasets and throwing them into a huge digital information repository where it is sorted and aggregated is impractical for the vast majority of health systems.

"It takes a lot of organizational effort to do big data and projects that require data warehouses," Mr. Shaikh says. "They're huge, complex, very expensive, and, historically speaking, often end up in failure."

While that's a great vision for a future in which building that technological infrastructure is more affordable and realistic, chunking data projects in smaller but more measurable goals is likely to be more beneficial — both for patients and hospitals.

For example, it is significantly more challenging for a CIO or hospital data administrator to justify $15 million in funding for a five-year project that requires a data warehouse than a $250,000 data-driven disease management program that takes three months to roll out. The less ambitious project might not only be easier to pull off financially and logisitically, Mr. Shaikh says but could produce a more readily demonstrable impact.

"I think the types of things that really have promise now are based in small problems and small data that add up," Mr. Shaikh says. He offered an example of a behavioral health data experiment where a healthcare company combined biometric and behavioral data to try to determine the predictors of depression.

The outcome the company anticipated was not the one they got, according to Mr. Shaikh. They hypothesized the most significant factors contributing to depression would be social and demographic, but their study demonstrated depression in the patient population was also highly correlated to certain clinical markers. That unexpected insight was a valuable metric the hospital could use to make targeted population helath improvements.

"If you can roll out a goal-oriented data project like that and prove it's actually generated value, you achieve momentum for the next small data problem, and the next," Mr. Shaikh says.

Setting sights on smaller projects can also help hospitals nail down the basics of managing large quantities of data. The basic tenets, such as appropriate data stewardship, data governance and data quality, are overlooked for the shinier, more-buzzed about terms, like analytics and big data.

"I think a clear strategy that's driven by your clinical and business needs is what's really necessary," Mr. Shaikh says. Organizations lose sight of how the big data infrastructure they're ready to sink capital into will actually return measurable benefit to the organization in a reasonable amount of time. It's the most important question a business should ask itself, he says, and it's one that generally goes unasked.

"How are you actually going to serve patients with this? You have to ground the vision," Mr. Shaikh says. "The dream is real and it's right, but you have to ground it in reality. And that reality has to do with the basics."

More articles on big data:

The importance of assessing critical IT system investments 
Mining for data gold: Transforming healthcare with natural language processing 
5 Tips for getting the most out of your healthcare data analytics program 

© Copyright ASC COMMUNICATIONS 2019. Interested in LINKING to or REPRINTING this content? View our policies by clicking here.

 

Top 40 Articles from the Past 6 Months