Can Data Help Solve Longstanding Challenges in Medicaid?

Which approaches are most effective at improving individual wellbeing and population health? What financial investments yield the greatest results in enabling greater longevity and quality of life for Medicaid members? Which interventions and investments fall short in achieving intended goals? 

As state Medicaid agencies face growing pressure to demonstrate and improve the effectiveness of their programs and initiatives, the ability to answer questions like these is becoming essential. That’s especially true when it comes to understanding the value of large, highly visible programs where expenses can really stack up.

Significant work is already being done to measure aspects of care quality and outcomes among beneficiaries of Medicaid and the Children’s Health Insurance Program (CHIP). For instance, the Centers for Medicare and Medicaid Services (CMS) requires states to report, on an annual basis, a set of core quality measures to strengthen health outcomes among children and adults. 

When dealing with complex and fragile populations, however, it becomes more challenging to weave data into a fuller picture of what’s working, what’s not working and how adjustments could contribute to better results. 

More challenging — but not impossible. 

Tackling an ‘Unsolvable’ Challenge

For every state, caring for those with severe and persistent mental illness (SPMI) has long been a significant challenge. These individuals represent at least 20% of Medicaid members, with the unstable 60+% of this population costing more than four times as much as their counterparts without severe mental illness. 

Despite higher spending, many physical disorders are three to four times more prevalent in this population. What’s more, among those with SPMI, the severity of those disorders is worse — and life expectancy is, on average, 10 to 15 years shorter.

These statistics aren’t new. In fact, for many years, the SPMI population has been regarded as an unsolvable challenge. That view is not only inaccurate, but also increasingly untenable as CMS asks states to deliver more equitable outcomes for these members by moving toward integrative care across the bio-psycho-social continuum. 

Through our collaboration with one particular state Medicaid program, Gainwell Technologies’ analytics have demonstrated that there are ways to improve outcomes and reduce costs for the SPMI population. 

Using Programmatic Analytics to Evaluate and Improve Outcomes

Before engaging Gainwell, the state had established an innovative intervention — a behavioral health (BH) health home to provide integrative care coordination for those with SPMI. The goal: to start moving the needle for these Medicaid members.

With the BH health home up and running, the state wanted to know: Are this program's care coordination services really making an important difference for those who use them? 

To answer this question, the state asked Gainwell to make a quantum leap in program evaluation analytics, beyond the standard metrics they were already collecting. Specifically, the state wanted a longitudinal assessment of overall program effectiveness. Moreover, the state’s team wanted a degree of robustness and replicability we’d not much seen in the reports that other states have published about their programs’ performance. As part of the required robustness, the state’s team asked explicitly for “error bars” — that is, estimates of the confidence the state could place in the findings. That way, the state could move forward with the solid results, not the shaky ones.

We rose to this challenge by creating an analytics solution that surfaces the impact of multiple drivers on four-year risk of all-causes mortality and on total cost of care, for those with SPMI who were referred to the health home. We also reused the same set of drivers in evaluating both outcomes — so that the state could see how the same drivers are in play in both longevity and cost. The result was a first: an overarching programmatic view of the state’s innovation program, which Medicaid agencies can’t achieve with industry-standard analytics.

The good news? Gainwell’s programmatic analytics clearly characterized how the state’s innovative BH health home is, in fact, starting to move a historically entrenched needle for the SPMI population. Beyond that, the solution surfaced additional insights about the interdependence of physical and psychological wellbeing for those with SPMI. And it pointed to opportunities for this program to provide more comprehensive care for the SPMI population, and the financial case for doing so. Using our solution — which runs in the cloud — we delivered a comprehensive understanding of the various drivers that impact health outcomes and TCOC among the SPMI population.

Findings of the Health Outcome Estimator

Four years is an unusually short time to evaluate mortality risk (consider, for example, that 10 years is the period commonly used with heart disease). However, in a population with life expectancy shortened by 10 to 15 years, four years is a long time. As a result, the Health Outcome Estimator yielded several findings in which the state can have solid confidence.

First, the Health Outcome Estimator showed that participation in the BH health home reduced four-year mortality for those with SPMI who received care coordination at least one time (about half of those referred). Using this estimator, we also found that, just like increased complexity of physical disorders, increased psychiatric complexity is associated with increased four-year mortality risk for those with SPMI. What’s more, this estimator showed that being seen and treated for non-lethal disorders is associated with reduced four-year mortality risk for those with SPMI.

Beyond the value of staying the course with its current BH health home intervention, from this single estimator, the state gained actionable insights about how to shape interventions for the SPMI population. For example, the findings suggest that simple complaints — like headache or digestive upset — can be gateways to other needed, sometimes life-saving diagnoses and treatments. 

The findings also indicate benefit to increasing the focus on earlier, decisive intervention into those life-threatening physical disorders that are more prevalent among the SPMI population (like diabetes, heart disease, liver disease, among others). Because these individuals often have long neglected their own healthcare, their disorders may have advanced to more severe stages.

Findings of the Cost Outcome Estimator

Through the Cost Outcome Estimator, we learned that for those with SPMI, psychiatric complexity alone exerts significant upward pressure on TCOC, in addition to the pressure exerted by complexity of physical disorders. 

Because members with SPMI often neglect their own health for years or even decades, care costs increase when they start receiving care coordination and begin catching up with their medical needs. However, once these members engage with the BH health home, their health status stabilizes and improves — and TCOC, no longer in runaway, drops by more than 5% per year.

Again, from a single estimator, the state gained actionable insights about how the BH health home was working to transform runaway cost growth into steady cost reduction. Further, the estimator supports better financial planning by illuminating an expectable burst of members’ “catch-up” cost on program entry.

Programmatic Analytics: A New ‘Power Tool’ for Medicaid Agencies

These two estimators provide a strong first demonstration for a program’s robust, longitudinal equity improvement and total cost reduction. Such a foundation is key to justifying large, expensive programs for the SPMI population. 

Yet these two estimators are just part of Gainwell’s programmatic analytics capability. Serving the SPMI population is just one example where programmatic analytics can produce fresh insights into the drivers of multifaceted challenges, the potential solutions to those challenges, and the tangible mechanisms for measuring and continually improving outcomes.

Indeed, the approach we took with this state generalizes to other interventions for complex, fragile populations. In addition to evaluating programs, the same estimators can help in assessing the performance of MCOs, hospitals, practices, providers and other intermediaries in the delivery of care. They can even be applied to individual members for triage, stratification and care personalization. Gainwell’s programmatic analytics offer a new, qualitatively more comprehensive power-tool.

This post’s title asks, Can Data Help Solve Longstanding Challenges in Medicaid? We believe the answer is a resounding “yes,” not only for the challenges presented by those with SPMI but also for many of the challenges Medicaid agencies face. To learn more about the programmatic analytics we’ve already delivered — and explore opportunities for your organization — contact us to speak with a behavioral health expert.

About the Authors

Carl Frankel, Ph.D., is Gainwell’s Behavioral Health Solutions lead, bringing decades of integrative experience across multiple key, “full-stack” disciplines. He has spent the last 10 years doing healthcare data engineering and data science, at scale, in both AWS and Azure clouds. Mid-career, he received doctoral training as a clinical psychologist, and spent a decade in academia as a published emotion science researcher using longitudinal methods. Earlier, he spent more than 10 years in Silicon Valley as a software engineer and architect, authoring five patents. Pragmatic Marketing^® trained, Dr. Frankel is committed to listening for clients’ pain points and working collaboratively to develop programmatic analytics innovations that lead to continuously improving, cost-effective programs and interventions for Medicaid populations with complex needs.

Gary Call, M.D., is senior vice president and chief medical officer for Gainwell. He collaborates strongly with Gainwell clients and leads our clinical program development and execution. Dr. Call has more than 25 years of experience in the practice of medicine and managed care. Prior to coming to Gainwell, he most recently served as the corporate vice president of clinical programs at Molina Healthcare. Dr. Call graduated from the University of Washington School of Medicine and completed his residency training at the University of Utah. He is a board-certified family physician.