Health Data Management: Value-Based Care Driving Need for Predictive Analytics
Health Data Management’s Greg Slabodkin spoke with our CEO, and several other healthcare analytics trailblazers, about how analytics are being used to help with alternative payment models like value-based purchasing. Check out the article below.
Value-Based Care Driving Need for Predictive Analytics
As the healthcare industry looks to transition to value-based care and new payment models that reward quality instead of quantity, healthcare organizations are seeking technologies that can help them improve patient outcomes and ultimately lower costs.
One such area is predictive analytics, which provides organizations with the ability to identify patients who are at risk for complications or expensive treatments. Some organizations are already using predictive analytics to identify patients who are at risk for readmission.
Others are using analytics tools to stratify the risks of patient populations as part of population health management, enabling them to focus resources on high-cost patients, particularly those with chronic conditions for whom the right interventions could save the most money.
Growing market
The market for health analytics is projected to nearly triple in value over the next five years, growing from $5.7 billion in 2015 to $16.9 billion in 2020, a compound annual growth rate of 24 percent, according to predictions by BCC Research. With the growing use of predictive analytics, BCC Research analyst Neha Maliwal believes the clinical analytics segment—projected to demonstrate the highest compound annual growth rate at 24.8 percent—should be a key market driver.
Currently, most healthcare organizations are extensive users of descriptive analytics, using reporting tools and applications to understand what has happened in the past and to classify and categorize that historical data. However, Maliwal believes the industry is moving to increasingly adopt predictive analytics—the process of learning from historical data to make predictions about the future.
“Predictive analytics is an area that has not been explored to its full capacity inhealthcare,and going forward, it is expected to reduce healthcare costs,” says Maliwal. He believes predictive analytics’ greatest impact will come in identifying patients with chronic conditions, enabling organizations to provide better care and to better control costs.
The use predictive analytics is still new for many healthcare organizations, a sentiment supported by a 2015 survey of 271 healthcare professionals conducted by advisory firm KPMG that found only 10 percent of organizations are using data and analytics at their highest potential, and 21 percent indicated that they are still in their “infancy” when it comes to data/analytics capabilities.
Likewise, a separate 2015 survey by healthcare technology vendor Jvion revealed that only 15 percent of hospitals are using some kind of advanced predictive modeling, defined as “the application of machine learning algorithms to find patterns within data to predict patient-level risk.”
Todd Schlesinger, vice president at Jvion, which develops software designed to predict and prevent patient-level disease and financial losses, believes the survey findings point to a growing need within the provider community for solutions that help prevent patient illness through real-time predictions. “With so much changing in the industry, providers are hungry for analytics that will help them improve health outcomes while reducing risk and waste across the system,” he says.
Patient care payoffs
Community Medical Center in Falls City, Neb., has successfully implemented predictive analytics as part of a population health program designed to target high-risk diabetes patients.
“We took this large group of diabetics that we had and we risk-scored them,” says Ryan Geiler, assistant manager and clinical analyst of family medicine at Community Medical Center, a rural health clinic with five providers. “We called the highest risk patients first, and within the first month, we diagnosed 20 new thyroid diseases.”
Patients with diabetes have a higher incidence of thyroid disorders, and having diabetes increases the likelihood of also having heart disease, Geiler notes. Thanks to the center’s population health program, he adds that they also saved one particular diabetic patient’s life by leveraging analytics capabilities to predict an adverse heart event that ended up requiring triple bypass surgery because of a major blockage of her coronary artery.
Since the initial success with its diabetic population, Community Medical Center has added other patient populations including screening women for cervical cancer and scheduling well-child visits as well as flu immunization campaigns.
“We have about a 25 percent Medicaid population, and what we found when we ran the numbers was that our Medicaid patients were not doing a very good job of making sure they brought their children in for well-child checks,” Geiler says. “We’re up 76 percent this year on the number of immunizations that we’ve given.”
Geiler credits the center’s NextGen Healthcare electronic health record system for enabling proactive outreach to its patient populations. “I’ve worked on multiple EHR systems, and one of the strengths of NextGen is that it has real-time SQL information available, so at any given time, I can go into the database and see whatever I want to on any of our patients, even if they were seen five minutes ago,” he observes. “It sounds like something every EHR should be able to do, but I can tell you that every EHR cannot do that. A lot of them act as a data repository, where you might have to wait two days before you get live data.”
The population health tracking software is also from NextGen, says Geiler, who developed his own risk-scoring algorithm, enabling care team members to discover gaps in patient care, identify high-risk patients, as well as those requiring preventive care. “99 percent of what I’ve done was using NextGen’s native population health program straight out of the box,” he comments.
For those healthcare organizations looking to implement predictive analytics for population health management, Geiler says that figuring out the workflow and the process is much more difficult than implementing the technology. While predictive analytics is a relatively new field in healthcare, he believes the sooner healthcare organizations embrace it the better.
“There is a fiscal incentive to do this now,” Geiler says, referring to the fact that by 2018 over 50 percent of Medicare fee-for-service payments will be rewarding for quality and value and aligning Medicare Advantage and Medicaid to do the same. “If you don’t start to do this now, next year you’re going to be in complete panic mode.”
Specific uses
Clay Richards, CEO of post-acute care management vendor naviHealth, believes in the value of predictive analytics. According to Richards, his company manages patients when they are discharged from hospitals, determining their risk of readmission and post-acute care needs, focusing specifically on improving care transitions and health outcomes.
Serving nearly 2 million health plan members and more than 75 partner hospitals and physician groups, naviHealth manages the entire continuum of post-acute care, helping to determine the appropriate care plan for patients after they’re discharged, using the company’s predictive analytics and evidence-based protocols to assess the “functional status” of patients.
“Whether they are in a skilled nursing facility or home health, we basically determine what is the right care setting for patients, and then we use our clinicians to continue to manage the care of those patients as they transition from hospital to post-acute and then back the community,” says Richards. “CMS has recently changed the conditions of participation for hospitals, requiring them as part of the discharge planning process to take more of a quality-based approach, looking at factors such as risk stratification and predictive analytics to determine, from a clinical standpoint, where patients should go.”
His company’s outcomes prediction tool—LiveSafe—has compiled 800,000 episodes of care, supporting discharge planning and helping to create individual care protocols designed to reduce hospital readmissions and focus on the cost effectiveness of post-acute care.
With post-acute spending estimated to exceed $200 billion per year and growing at approximately 6 percent annually, Richards argues that reducing these costs has never been more important for health systems, payers, and providers. Consequences will only loom larger with emerging value-based payment models, including bundled payment initiatives and Medicare Advantage.
Using naviHealth’s benchmarking data, one health plan—Priority Health in Grand Rapids, Mich.—has achieved a 15 percent decrease in skilled nursing facility days per 1,000 beneficiaries, reduced average length of stay, and reported a 13 percent reduction in its per-member, per-month costs in its Advantage program, Richards reports.
Other naviHealth customers have reported similar results, Richards says. “We’re able to get patients to the appropriate level of care, which results in not only lower utilization rates but also lower readmission rates and higher patient satisfaction.”
Another company in this space isMedalogix, a home health analytics vendor. The company’sTouchsolution is a population health management solution that specifically analyzes home health clinical data to identify patients who are at the greatest risk for hospital readmissions.
Touch leverages a customized predictive model to generate patient risk rankings using real-time EHR data. Its capabilities enable clinicians to view which patients are most at risk for being transferring off census before their 60-day care episode is completed, and which patients would benefit from an additional care episode.
“The model is highly predictive in being able to identify patients who have elevated probabilities of hospital readmission and helping home health care agencies reduce that risk,” says Dan Hogan, CEO of Medalogix. “You will start seeing large swaths of budget dedicated to the acquisition of predictive tools like this, especially in large hospital systems.”
Originally published inHealth Data Management.