Measuring value — How to harness data in an outcomes-driven paradigm

Value-based care arrangements are placing healthcare institutions in unfamiliar situations, introducing most providers to risk for the first time. As providers begin to shoulder risk, it is critical hospital leaders determine the appropriate metrics to measure physician performance, track costs and use data to bring about continuous quality improvement and other operational enhancements.

This content is sponsored by 3M Health Information Systems

Establishing and tracking metrics necessary to manage risk and improve patient outcomes can be challenging for organizations, and some may end up going down a quality improvement path that ends in an unintended outcome. That is, hospitals may have an outcome in mind, but initiate a value program that yields a different result. For example, a hospital may implement an enhanced process measure, such as a diabetes process indicator, as opposed to an outcomes improvement measure, like hospitalization rates.

"When measures are misaligned with the outcomes we want, it provides a muddy signal," explains Gordon Moore, MD, senior medical director of clinical strategy and value-based care at 3M Health Information Systems. "We [head] in a direction that isn't the ultimate goal."

As organizations set up and track the appropriate value measurements, they should look to models currently working in the value-based care environment.

Value-based care models include pay-for-performance; bundled payments; patient-centered medical homes; accountable care organizations; upside-only and downside shared savings; and capitation. Currently, around 15 percent of provider organizations' business operates within value-based agreements, says Jason Burke, vice president of data informatics at 3M Health Information Systems. Some commercial payers are pushing this trend, with more than 50 percent of their contracts operating within a risk-based model. Mr. Burke foresees the majority of contracts moving in the direction of value over the next three to seven years.

Quality misalignment in value-based care
Delivering better care at a lower cost is not a new phenomenon in healthcare. Steve Delaronde, manager of consulting and analytics for population and payment solutions at 3M Health Information Systems, recalls when payers began implementing capitated payment models in the '90s to incentivize healthcare organizations to provide more cost-effective care.

Because there were no quality targets associated with these delivery models, participating organizations often failed to improve outcomes while lowering costs, largely because "quality" and "outcomes" weren't explicitly defined. Mr. Delaronde believes healthcare leaders can learn from previous experiments in managed care to better inform value-based care models today, especially when it comes to setting clear definitions of quality measurements.

"The managed care backlash of the 1990s taught us that quality and outcomes had to be clearly defined as a specific goal and measured," he says. "Unfortunately, quality is not nearly as easy to define or measure as cost."

Misalignment of quality measures impedes success within the value-based care paradigm. An organization may be focused on a clinical outcome, but mired in tracking and reporting measures that don't contribute to improving outcomes and instead distract physicians from patient care. When entering a value-based contract with a payer, a health system agrees on key quality and performance measures. The payer will assess the health's system performance against these measures to determine reimbursement. For example, a payer may look at a health system's mortality and patient experience measures.

Quality measures in contracts rarely mirror each other from one organization to the next, however. In fact, GAO researchers discovered quality measurement misalignment between public and private payers as well as across state and regional healthcare programs. While 48 state and regional healthcare programs — including state Medicaid programs, commercial health plans and regional collaboratives — utilized a total of 509 quality measures, only 20 percent of the measures were used in more than one program, according to a 2013 study by Bailit Health Purchasing.

This misalignment creates a surplus of administrative burden on providers as they attempt to report on measures enforced by different payers. Annually, internists, family physicians, cardiologists and orthopedists spend more than $15.4 billion to report quality measures to payers, based on a Health Affairs study. Because payers use different quality measurements to determine outcomes, providers receive incomparable performance assessments, meaning no two payers will rate the provider the same. Quality measurement misalignment may inhibit providers' ability to drive continuous quality improvement under a value-based care model due to increased administrative burden as well as inconsistent performance assessments.

Missing links in data measurement and management
The Triple Aim sums up the definition of value: better outcomes, lower costs and an enhanced patient experience.

"The Triple Aim is very clear that the numerator of the value equation is a combination of quality of care and the patient experience," explains Mr. Delaronde. "Compared to the denominator, which is price, the metrics in the numerator are the great stumbling block to successfully defining and measuring value."

Common quality metrics include the following:
• Mortality
• Complications
• Healthcare experience
• Patients' physical and mental ability to perform activities
• Risk-adjusted total cost of care
• Hospitalization rates across all conditions, adjusted to illness burden

While organizations likely know to focus on metrics such as those above, compiling the data necessary to measure them is more difficult. Healthcare organizations possess enormous amounts of data, but most lack the tools or capabilities required to meaningfully access, organize and deploy that data to inform patient care.

"We have more than enough data to measure processes, improve outcomes, manage care, facilitate communication, coordinate care, and reduce costs," says Mr. Delaronde, but as much as 80 percent of healthcare data is unstructured. This means it does not adhere to a pre-defined data model and is unorganized. Unstructured data presents a vast opportunity in healthcare to "collect, organize and harness the data elements available in all data sources to achieve the objectives of the Triple Aim," says Mr. Delaronde.

The lack of standardization in data measurement also means organizations find it difficult to compare quality across enterprises, or even within the same organization. Mr. Burke notes many 3M clients say they want to assess how they're doing against an adopted market standard so they can risk-adjust their data on financial and quality care levels.

Speed and interoperability represent additional technological hurdles. When organizations fail to process information quickly, providers are unable to use the information to influence medical decisions in real-time. Without real-time data, all insights are retrospective and thus not driving change. Additionally, many health systems try to purchase enterprisewide EHRs, but run into obstacles when trying to collaborate with affiliated practices in the ambulatory setting due to interoperability issues between clinical settings.

"The ability to consume that data and make it useful and rank the opportunities so [providers] can identify unnecessary cost drivers is critical," says Dr. Moore. "It's a struggle that we're working on quite aggressively, because that is the future."

Making sense of quality data
To harness data for actionable insights, providers are seeking partnerships with organizations that can help them compare massive amounts of data as well as provide tools for accessing unstructured data.

Mr. Burke says it is imperative for providers to pick the right partner and approach for their institutions. "You can't jump from vendor to vendor and approach to approach, because you will dilute the data," he cautions.

Providing the resources for the following approaches, vendor partnerships allow organizations to harness data for comparisons and insights.

1. Leverage data to inform quality measurements. The industry would benefit from organizations sharing best practices and case studies. Hospitals and payers could then develop standard quality measurements that drive actionable insights and support quality improvement initiatives.

3M aims to do this by leveraging massive amounts of healthcare data to pinpoint which quality factors lead to different outcomes. The company then recommends systematic improvements that will yield improved outcomes.

"Our approach is to marry technologies and bring those together in a data-rich learning environment that is meaningful to outcome change," explains Dr. Moore. The company develops financial as well as episodic, preventable and risk modules.

2. Use artificial intelligence and big data to drive insights. Organizations must be able to apply structured data in real-time, and pull from unstructured data to act on insights. If providers could search keywords in unstructured data, like "shortness of breath" or "difficulty walking," they would be able to pinpoint underlying issues that structured data sources currently don't address.

Artificial intelligence helps providers comprehend unstructured data. Machines have the ability to analyze vast quantities of data by identifying patterns, such as the combination of symptoms that show a patient is at risk of sepsis. "The deep learning gives us access to comprehend things that are incomprehensible at a human scale," explains Dr. Moore. However, Mr. Burke says approaching AI with a data science mindset alone won't necessarily yield the desired answer. Effective use of AI tools requires data-driven science as well as expert knowledge and a clear understanding of context.

That's why 3M leverages natural language processing technology to combine unstructured and structured data. The combination of the two data sources creates a clearer picture of clinical scenarios, because AI and big data both depend on vast amounts of captured data. The more data, the more value the technology creates.

"Big data is like a fire hose to someone who just wants a glass of water," explains Mr. Delaronde. "Unless we understand how to harness data and use it to solve real healthcare issues, just having it doesn't help."

How to stay ahead of value-based measurement
In the era of value-based medicine, providers and payers must be willing to try various care delivery, reimbursement and measurement models. One method will not work for every institution, so the industry must build tools and metrics that can conform to an array of approaches.

Further, organizations should hold unwavering focus on the core work that will most likely result in population health improvement. To advance this work, leverage technologies that report metrics most closely aligned with improving population health management, says Dr. Moore. Ensure these technologies are proven and defensible in the industry, though.

"Defensibility is a key problem and a key piece of the decision," explains Mr. Burke. "There are many ways to risk-adjust healthcare data in the market, but if you can't defend [the method] with physician groups or financial groups, there's no value for it."

Ultimately, organizations must be able to access, integrate and analyze data to measure processes, outcomes and costs.

"We must move in the direction of what I would call 'intelligent simplification,'" concludes Mr. Delaronde. "If we ever get to that one metric or value that can be used to measure our effectiveness in delivering value in healthcare, there will be an immense amount of thought, data points, intelligence and technology behind it."

Copyright © 2024 Becker's Healthcare. All Rights Reserved. Privacy Policy. Cookie Policy. Linking and Reprinting Policy.

 

Featured Whitepapers

Featured Webinars