What does it mean for an organization to pursue a data-facilitated journey into value-based care? Bill Gillis, CIO of the Boston-based Beth Israel Deaconess Care Organization, shared his learnings and perspectives on that subject, in the opening keynote presentation of the Health IT Summit in Boston on Tuesday morning, sponsored by Healthcare Informatics. Speaking to an audience gathered at the Courtyard Boston Downtown Hotel, Gillis spoke on the topic “Managing the Risk Madness: The Value-Based Return on Data Quality.”
As the description of the session noted, “Value-based and accountable care organizations [ACOs] rely on accurate data, analysis and reporting to provide the critical information the organization needs to understand how it’s providers are performing to meet risk-based contract agreements. The extreme challenge is how difficult it is to accurately capture and report on the numerous and disparate performance metrics individual to each risk-contract across commercial, Medicare and Medicaid markets. The lack of standardization across payers is further complicated by a lack of standardization across EHR vendor systems and quality measures for the healthcare industry overall. This is layered on top of the foundational challenge that much of the information health care organizations collect and store for data analysis is incomplete or imprecise.”
Speaking of the Beth Israel Deaconess Care Organization, or BIDCO, a joint hospital-physician membership organization that has brought physicians and hospitals together to deliver care to 225,000 covered lives across 300-plus geographic locations and 35.6 million annual patient encounters, Gillis told his audience on Tuesday morning that “When I talk to our peers across the country, sometimes I think BIDCO is unique. And I’m envious of folks on a single Epic instance; but even those people have docs or systems in their communities that are on something different. And sometimes, they just give up. The reality is, you can get there. But this data resides in different systems. We have 46 different EHRs [electronic health records], and 150 different installs. There is data in core EHRs, but also in lab, radiology, and infusion systems. You want to try to get your hands on all that data, and that can be a challenge in and of itself. And the data itself can be very messy.”
Looking at the oceans of data available to use on behalf of accountable care and risk-based contracts, Gillis noted that the landscape of that data is extraordinarily complex—and yes, messy. There are essentially four categories of data available for use in that context, he told his audience: structured, coded data; structured, uncoded data; unstructured, coded data; and unstructured, uncoded data. “Structured, coded data,” he said, “is the really good stuff; but it represents 1 or 2 percent of all the data you’re going to get. Meanwhile, 5 percent of the data is structured and uncoded,” and needs intensive mapping and normalization. “The really dirty stuff is unstructured and uncoded data; and 85 percent of our data falls into that category. And we don’t really have a way to use that data. So we really try to focus on the first three categories.” The bottom line for all those working with the data, he said, is that “The burden of the normalization and validation of that data falls on you.”
What about continuity of care documents, or CCDs? “Will CCDs solve our problems?” Gillis asked. “The reality is that every one of those systems stores the data differently, meaning that their CCD payload is different. So you’ve got to take all the data from those CCDs, and normalize it and validate it”—which presents a major challenge for those working in accountable care and risk-based contracts.
“We’ve learned a lot, and we’ve gotten beaten up a lot,” Gillis said of the experiences that he and his colleagues have had so far on the journey. One of the things that he said he hears all the time is the idea that “It’s just interfaces, right? And I can just connect them and extract the data, right? The reality,” he said, “is that you get into a network with varied systems, vendors, payloads, and transport mechanisms—and it turns out that everything is all over the place. It feels a lot like organized chaos, and it’s very challenging. Even if you’re on a unified platform, it’s challenging,” he said, noting that his core hospital organization has been on a unified platform for several years already. “In the ideal world,” he added, “you’ve got the data points connected, got the pipes connected, and the data’s flowing, and you’re proud. But is that data usable? The reality is that no, it’s not. We pulled all the data in five or six years ago, and pulled it apart; it’s pretty messy stuff. The fact is that so much of the data is uncoded and unstructured, and everyone one of the [EHR] systems stores its data differently.” As a result, he said, “It just takes some time to work through” all of the necessary processes.
Making meaningful connections with physician practices
An absolutely key element in all of this, Gillis told his audience, is the need for the leaders of ACOs to reach out very directly and intensively to connect with all those in physician practices affiliated with their ACO organizations—from the physicians in those groups to everyone on their staffs. “The reality of improving data capture starts at the practice level,” he said. “Practicing physicians and their staffs need to capture data better. Ability of useful EHR data capture requires clinical practice workflow transformation to maximize data value.”
What’s more, Gillis said, “That’s challenging, because doctors want to treat their patients,” and find so much of what’s going on around data input and output around quality measures for value-based healthcare to be burdensome, at least at first glance. “But the reality of these risk contracts is that it’s critical that we get these data points back. So you need to get out into the community, meet with the physicians, and help them to understand the journey around getting the data capture to capture these data points to support the risk contract.” He added that he believes that it will be nearly impossible to achieve breakthroughs with all of this on a nationwide level without some kind of national patient identifier.
Meanwhile, Gillis went on to say, broader challenges face all those working in accountable care. “All systems store data differently, and there is no fully embraced HIE [health information exchange] standard. What’s more, an overall, industry-wide vision of interoperability has still not been achieved. And there remains a lack of vendor support, as well as the problem of fragmented access to EHR data.”
At the heart of the journey, he said, are “data mapping, data normalization, and data validation,” using some combination of a broad range of data standards, including “NDC, RxNorm, LOINC, ICD, Snomed, CPT.”
Meanwhile, he noted, “We’re in 10 different risk contracts, and each payer wants to calculate things differently”—another challenge in the data-facilitated journey into value-based care delivery and payment. Each of the payer wants harmonization of calculation of data points, but the vendors aren’t set up to standardize how they organize data. We need standard, unified quality measures that leverage EHR data” on a national level, he emphasized.
Still, Gillis noted, “We’ve achieved real-time performance reporting, taking documentation, getting that back in a 24-hour period, merging that with our data warehouse, and creating real-time performance data points.” Above all, he said, offering advice to attendees, “What you need to do is to develop a sustainable, repeatable process. We had our traditional EHR support folks; we had quality improvement coordinators, who were using claims data to support quality work. And we realized, getting into our ACO work, that we needed to merge those two. And we realized we needed to understand what the practice workflow was like. So we merged those two teams together. Initially, we called them the EHR optimization team, but now call them performance improvement facilitators. They were under me, but are now under the population health group, under the chief medical officer. It’s less about the technology and more about the workflow and understanding the practice side. Improve along the way, make changes in direction. And that,” he said, “is the trajectory: getting yourself on a trajectory is the way to go.”
