Building the ACO Engine

The Northern Virginia-based Inova Health System with a total of 1,753 licensed beds, one of many healthcare organizations on a path to meaningful use, is in the process of liberating the data silos of its six community hospitals to inch its way toward becoming an accountable care organization (ACO). Inova enlisted the help of the Redwood Shores, Calif.-based Oracle Health Sciences, whose suite of interoperability and analytics products formed the informatics foundation of a data services hub within Inova to cull patient information for clinical care and research purposes. HCI Associate Editor Jennifer Prestigiacomo spoke with Inova’s Chief Medical Information Officer Ryan Bosch, M.D., about the strategies and challenges surrounding the journey toward accountable care. Here are excerpts from that interview:

Healthcare Informatics: What is Inova currently doing to link its silos of data?

Ryan Bosch, M.D.: We are really excited about our allergy product and our allergy work. It’s going to sound like a little, teeny thing, but when you have all those different systems, a lot of those systems have an allergy list in it. When a patient shows up to our ER, that patient expects that we enter the allergies [across all our systems]. During that visit, that might be the only system they touch. One of our hospitals only runs McKesson, so the patient might always be in McKesson. The next year they may move, and get admitted to another Inova hospital, so we have to think at a more enterprise level. And the Inova intelligence engine allows us to consume disparate pieces of data and liberate them from their core source, and then create a source of truth for that patient, so when we pull up the Inova intelligence engine we can say, “What allergies did Mr. Smith have?” And we’d have a consolidated view of what they were.

So the master patient index is in its simplest state a source of truth, for a unique identification method of that patient. It’s like a medical record number with a wizard in it that helps it match and make sure that the 10,000 John Smiths that we have, we know which one of them are the same and we have algorithms that allow us to make those connections and bridges and algorithms that tell us, yeah these probably are same, but we’re still not comfortable enough based on our internal audit to bridge them, so we’re going to keep them as individuals now. Conceptually, it allows us to know and identify an individual across our entire footprint.

HCI: What have been some challenges so far?

Bosch: We call our project inside the shop here the Rosetta Stone because we are taking Greek, Cyrillic, hieroglyphics, and trying to find translations. Because the legacy data was written in whatever language they chose to write it in, without any plan ever to push it to anybody, we have 15 years of data written in hieroglyphics. I can go extract it, but it doesn’t mean anything. So how do I extract it in a way that translates it too, not only in today’s terms—because we’re in an explosive interoperable environment, where there are actually more and more definitions, “you should use this language when you’re talking about medications, and the FDB [First Databank] language when you’re talking about allergies, and SNOMED for medical problems, and ICD-9 when you’re talking about billing and diagnosis? So, the real challenge is, the IT folks would call it, the migration table. How do you use a Rosetta Stone to consume and extract the data and place it in a new storage where it has the tags and the categories and it has context now to be used in a meaningful way?

HCI: How will healthcare systems like Inova move toward the ACO model of care?

Bosch: The whole accountable care model really requires an intelligent delivery of healthcare. You need an intelligent engine underneath it. You need to know what you did, how it helped the patient, the care situation, so you can iterate on it and do better. It’s a feedback loop. To become an accountable care organization, it implies that as a healthcare delivery system, we have insight into what we’ve done, where we did it, how much it cost us, how much it cost the patient, how much benefit it provided? There are just so many different layers of feedback or analytics that it requires—because ultimately the ACO model is trying to take some of the risk of financial care to the patient away from the third party payers and push that risk to different places. So, they’d like health systems to accept some of that risk in the care of the patient, but to do that, anyone who wants to be an ACO has to know really granularly what their costs, expenses, outcomes, and quality are. To get to that stage of being an ACO requires this internal capacity to know your internal data really well.