Is Any Part of Population Health Easy?
This March I had the chance to attend the HIMSS conference in Las Vegas. I spent the trip wishing I had a pair of Tempur-Pedic mattresses in my shoes but found my consolation in talking with providers and vendors. One conversation with a vendor representative stands out in my memory. We’d been chatting about data aggregation, and the rep made the claim that data aggregation was easy for their company.
My audible response was, “Really? That’s impressive. Most people I talk to have difficulties with that.” My internal response was, “Nope. There’s got to be more to this story. Data aggregation is certainly doable, but it’s never easy.” That idea led me to wonder to myself, “Is anything in the field of population health easy to do?”
The short answer is: Probably not. The long answer involves a closer look at the different pillars of population health, the road blocks and complications within each pillar, and the humanity behind the technology.
Vendor executives and provider leaders defined six pillars of population health during KLAS’ Keystone Summit in 2016. I consider the first two pillars, data aggregation and data analysis, to be the foundation of population health IT. But “foundational” doesn’t mean “simple.” Even these beginning phases in a population health strategy come with extreme difficulties.
A Lack of Standards
If your provider organization has a single EMR system, I hope it’s easy for you to aggregate your EMR data. But you’ll also want to bring in claims data from payers, including CMS. Each payer sends data in a slightly different format. If they’re lucky, providers can convince the payers to use a standardized format.
But even that isn’t the end of the road. Once the data is aggregated, providers have to deal with duplicate information. Questions like, “How many of our facility’s eight John Smith accounts are actually for the same patient?” are as common as they are frustrating.
Those questions are further complicated by diagnosis codes. In addition to different data formats, payers may use different diagnosis codes for the same conditions. This could leave providers asking, “Do these different diagnosis codes prove that our two John Smiths are separate people, or is there only one John Smith with the same diagnosis coded in two different ways?” I’m getting a headache just thinking about this.
Too Many Stratification Options
Even if my vendor friend from earlier actually could sidestep the difficulties I mentioned and easily aggregate a provider organization’s data, the organization would still have hurdles. For one thing, aggregated data isn’t always accurate data. No one can perform effective data analysis—the second pillar of population health IT—without effective data.
Even accurate data can be stratified in a paralyzingly large number of ways. There are many different registries providers can use, such as those for diabetes, COPD, and other chronic conditions. But an organization could also choose to target their most expensive patients. That’s not to mention meeting metrics for value-based contracts, choosing the best risk-score system, or deciphering the other options that are dumped in providers’ laps.
With so many ways to slice and dice data and no how-to manual for finding guaranteed success, many provider organizations feel overwhelmed. Maybe that’s why most of them start small. Choosing one or two chronic diseases to target and hoping to expand from there is hard enough.
Difficult but Possible
Luckily, there are some bright spots in the industry. Most provider organizations that do well with data aggregation and analysis are thriving because they work with great vendors. These vendors aren’t exceptional because they can magically simplify everything; they’re exceptional because they are engaged, happy to help, and cooperative partners. Such success should give us hope. When provider organizations get the guidance and expertise that they need from vendors, the foundation of population health can be built.