There are very few things in healthcare we can all agree on. Providers, policy wonks, do-gooders, business folks-we come at this $3 trillion dollar problem attempting to “bend the cost curve” and “provide patient-centered” care. Big Data, Data Science, PCMH, ACO, HMO, PPO, MSO: the acronyms and catch all phrases have gotten so predictable and empty that a person with decades of healthcare experience can still find themselves in a conversation that’s disturbingly coded. At the top of all of our lists is the importance of “data analytics.” But, what exactly is data analytics? Is my data analytics the same as your data analytics? The following will define what I think data analytics is and what it isn’t and how we should be embracing story more than numbers.
"We need to embrace a new era of data analytics wherein we capture all the soft factors"
What Data Analytics Isn’t
Do me a favor-write down your definition of data analytics. Done? Okay, how many of you mentioned any of the following: statistics, programming, machine learning, mining, interfaces like Tableau, Alteryx, SQL, SAS, predictive algorithms? It’s fine if you did. Those are all components or tools that one would use for data analytics, but what I see as the biggest voodoo in our current healthcare data zeitgeist is that those tools are dominating the definition. This misnomer is perpetuated by the fact that data product developers define analytics in such a way as to sell you their stuff. So then, of course, we all think the tools the keys to data-driven decision making. “Just build the platform and we’ll all be data savvy!” Data analytics is not defined by the tools just like a guitar is not defined by its strings. No one has ever said that Jimi Hendrix had the most amazing strings on his guitar. Data analytics, therefore, is not defined by data nor is it defined by analytics. So, then, what is data analytics.
What Data Analytics Is
When was the last time you were at a presentation that displayed some deck while an expert basically regurgitated the slides? They showed you graphs and charts and reports. They were attempting to do something that is the basis of data analytics. They were trying to tell a story. They were failing because there was no relationship built between the information and the real-world application. Without tangible relatable factors, data is noise without story. The trick is that the people who are typically the most adept at understanding how best to extract data are so left-brain dominant that they have a hard time telling the story between the numbers. The data extracts hone in on points that speak to statistical significance (p-values) and confidence intervals (control charts) but leave out the soft stuff. For instance, what is the family structure of those people in the hospital whose average length of stay for congestive heart failure was longer than average? That’s not as easy to pull because we haven’t prioritized the information enough to capture it. And that’s the change that’s needed.
Capturing Information For Stories
We need to embrace a new era of data analytics where in we capture all the soft factors. Instead of cleaning the raw data, we need to dirty it up. For instance, when we look at pharmacy data we should look at how far from a pharmacy someone lives, how many cars they have, who can drive in their family, and what primary languages are spoken in homes. We need to look at all the information that surrounds the people we are trying to understand. In coordination with recording more ancillary information we need our data-oriented people, whether scientists or managers, to take story telling classes. Writing instructors should be training our healthcare leaders and data scientists to approach data analytics more like Thoreau than like Google.