July 26, 2018
The pair say the new model “suggests agencies should be able to materially reduce visits per episode while ensuring high-risk patients get more care than before.”
Original article published in Predictive Modeling News Periodical (www.predictivemodelingnews.com)
Medalogix, a healthcare data science company based in Nashville, reports that its data scientists “identified several trends that were surprising” when they studied the results of a pilot program testing its newest product, Medalogix Care. “Based on analysis of multiple patient populations, Medalogix found a clear trend of diminishing returns and reduced visit effectiveness for some patients earlier in the home health episode of care than what was expected,” a statement from the company says. “With the industry average number of visits provided per episode ranging between 15 and 19, this will create resource capacity for an agency’s higher-risk patient population.” Medalogix Care is a home health utilization management solution that “leverages predictive analytics and optimization technology” to provide guidance on patient care plans to providers, the statement adds, based on “historical analysis of millions of patient records, care plans and outcomes.” Each patient’s clinical and functional assessment is analyzed; the solution then provides “an objective determination of how many home health visits the patient may need to achieve the desired outcome.” Early simulations of the pilot data “suggest that a home health agency should be able to reduce visits by a significant amount while achieving the same or better outcomes,” the companies say.
Adds Elliott Wood, CEO and President at Medalogix: “With the ability to optimize episodes and resources, patients with high-risk needs are likely to get more care than what they’ve experienced historically. Machine learning technology allows us to do more for patients who need additional clinical attention.” The pilot began in October 2017 with long-time customer Encompass Health Home Health and Hospice. “We spend a lot of time and resources ensuring the data we collect is accurate,” reports Bud Langham, Chief Clinical Officer there. “Building a ‘just right’ care plan is very difficult, because there are literally hundreds of factors to consider for each patient.” He adds: “Having evidence-based recommendations and statistical support to augment our clinical decision making is critical as we continue to move toward a value-based system.” And, says Luke James, Chief Strategy Officer, having those capabilities “has the potential to change the conversation with our strategic partners.” Encompass has partnered with Medalogix in the past to identify risks in its patient populations, he adds; the addition of the new solution “allows us to take risk in a measured fashion and have confidence in our approach.” It will also come in handy as the Centers for Medicare and Medicaid Services, through the Home Health Groupings Model, “fundamentally changes how agencies are reimbursed,” Medalogix adds in the statement. Its platform is integrated with “several home health and hospice electronic medical records,” the company says, “and is the only portfolio of predictive analytics products focused on improving patient outcomes in the home health space.” Predictive Modeling News talked to Wood about using analytics to better target care and about how the home health population is changing.
Predictive Modeling News: Can you tell us more about the surprising data on diminishing returns? What types of visits were behind the trend? What types of patients?
Elliott Wood: We are looking specifically at patients who are using home health services with a skilled need. These patients are receiving a blend of skilled nursing, physical therapy, occupational therapy, speech therapy and home health aides. The most surprising element of the analysis was the distance between the point of diminishing returns of total visits for a patient in an episode and the number of visits patients receive on average. However, the wrong assumption to make with the data is that every patient simply needs fewer visits. Every patient needs a more precise assessment of his or her actual need. That is where technology and data science can help. Our goal is to be more precise about care a patient needs, so patients who need fewer visits get fewer visits, which frees up the capacity of a provider to do more visits for patients who need more visits.
PMN: Do most home health agencies have the data necessary for detailed predictive analytics? What type of data do too many lack? Is interoperability a problem?
EW: Yes! Agencies have the data, but that is not really the challenge. The challenge exists with having a good modeling process and cultural disruption — for example:
- having a clear understanding of what problem you are trying to solve
- developing the appropriate data understanding and structure for the data
- having the skillset and the technology to develop performant models
- having the skillset to engineer or re-engineer clinical workflow to incorporate the information into the process so that it can make a difference.
This is where a bunch of people miss the mark.
PMN: Do home health agencies face financial pressures under state and federal benefits programs? How can predictive analytics help, in addition to planning for enough episodes?
EW: Yes. Home health agencies have faced financial pressure for several years in a row. Even at the end of a several-year reduction in reimbursement, CMS is now in the process of studying and understanding the initial pilot of home health value-based purchasing, as well as defining HHGM with a technical expert panel. Either of these have broad, sweeping financial implications to the industry. In addition to the changes CMS is considering for traditional Medicare populations, managed care enrollment continues to outpace traditional Medicare enrollment, and managed care has traditionally come with lower reimbursement rates for home health than traditional Medicare.
PMN: What role does home care play in accountable care plans and other value-based programs for private payers? How can home health agencies use analytics to maximize that role?
EW: We believe home health has a critical role to play in value-based care. As the most cost-effective care setting by far, we continue to see our customers consistently drive improvement in key outcomes for their patients. Additionally, we can tell you from the data, patients in home health today have higher risk for hospitalization than they used to, which means they are sicker. This also means these value-based networks and referral sources are trusting home health agencies with their high acuity patient populations earlier in the process. We believe Accountable Care Organizations and value-based networks will have to recognize this, and begin to view home health as a strategic asset and partner, specifically for keeping patients out of the hospital and in their homes.