Transforming Behavioral Health Care With Predictive Analytics

When contemplating data and analytics technologies – particularly highly technical systems like SAP – most people think in terms of supply chain management, sales, accounting, data governance and other enterprise business processes. Useful, valuable applications that, truthfully, may not be very exciting to the average individual. But more often than you might think, we get the chance to leverage our technology in pursuit of something truly extraordinary.

Our recent partnership with HarrisLogic, a technology and clinical services company, is an example of exactly that. They selected EV Technologies to assist in modernizing and advancing analytics for their proprietary technology, Stella, a platform which allows multiple, disparate entities to share criminal justice data and protected health information and supports behavioral health care decision making. Working together, we developed a tool so successful that it has been recognized with a Digital Transformation Award by the 2018 SAP Innovation Awards.

Reducing Recidivism and Crisis Care Costs With SAP

HarrisLogic was founded on the belief that technology can improve mental health care. Specifically, their platforms focus on suicide prevention, crisis response, jail diversion, and detection and coordinated response to super-utilizers. Their Crisis Response and Jail Diversion platforms streamline and manage crisis hotlines, mobile crisis outreach efforts, telemental health and facility-based mental health initiatives, offering an invaluable solution to behavioral health programs that operate in resource-scarce environments.

“We translate science to practice and bring evidence based tools into mental healthcare: improving service delivery, saving lives, and changing the paradigm of how we treat mental health.” Hudson Harris, Chief Engagement Officer, HarrisLogic

We began working with HarrisLogic to help them make better use of their data at the operational level. Together, we sought to leverage their decades of clinical expertise and client records and our data and analytics capabilities to 1) develop a data-driven clinical support tool and 2) accurately predict which clients would be high utilizers of resources or would run the risk of recidivating (returning to prison after release).

High utilizers and recidivists typically comprise the highest cost for insurers and state and local governments. In leveraging a thoughtful user experience and predictive analytics, we were able to reduce and make more efficient use of community tax dollars on crisis care spending while saving lives, breaking the high utilizer cycle, reducing jail population and recidivism rates and helping to prevent repeat offenses.

Using SAP HANA and SAP Predictive Analytics to Identify At-Risk Individuals

To achieve our goals, we proposed a migration to an in-memory platform, SAP HANA, a new user experience, and the integration of a set of predictive algorithms that HarrisLogic had always hoped would be possible but didn’t quite have the internal tools or team to apply.

Together, we launched an eight-week pilot program that leveraged SAP Predictive Analytics to use 29 months of historical data – gathered from ERs, jails, 911 calls, community services, courts, hospitals, primary care, and other sources – and attempted to determine if we could accurately predict what took place in month 30. The results were astoundingly successful. In the pilot alone, we achieved:

  • Predictions which were 99% accurate
  • 89% average rate of crisis diversion in major metro areas
  • 25% reduction in behavioral health crisis spend

The results proved that with the right data and analytics, you can detect warnings and triggers that influence someone’s behavioral health. The next step was to take the existing Stella experience and rebuild it with a user experience that engaged and empowered its government user base while providing fast, accurate and most importantly, live predictive analytics. The results of our pilot needed to be repeatable, and the live stream of data needed to be presented in an easily digestible manner.

Crafting User Experience to Empower Behavioral Health Clinicians

We had overcome our first hurdle and achieved our initial goals. Now, the challenge was to present vast amounts of live data in a way that put it into context for the user. Good design and user experience would give shape to the data and help drive better decision making by the clinicians accessing it. Getting the UX right was crucial for everything from saving costs, increasing comprehension, reducing decision-making time, deterring litigation, increasing trust, reducing training costs and more.

We partnered with our highly skilled UX partners at Integrity to give Stella 3.0 a facelift in the existing framework that was lighter and more responsive, allowing on-the-go access to Mobile Crisis Units.

“This tool will allow me to completely redesign the clinical process and provide the right amount of care at the right time” – Executive Director of Mental Health Provider

Predictive Analytics on HarrisLogic and EVT Stella application

The updated application allows for up-to-the-minute prediction on incoming jail records, flags important factors for clinicians and enables real-time decision support for accurate resource allocation. Its easy-to-use interface allows users to look at individual client care as well as use data to make higher level business decisions. An advanced search function also allows users to pull up records that match specific criteria, such as individuals who have just committed a violent crime. Additionally, the revised workflows and integration of a Predictive Recidivism component gives clinicians information necessary to provide the best possible care.

In the end, we demonstrated with confidence health outcomes that would allow a caregiver to provide more effective, targeted care. The live data stream provides in-the-moment decision support by pairing mental health data with criminal justice and other contextual data. This additional information reduces the utilization of higher levels of care like hospitalization and commitment to psych wards, which are in fact much more expensive and much less beneficial to the individual than lower level options of individually-focused care.

Achieving Award-Winning Results

The results of our hard work have been both impressive and immensely gratifying. In addition to the aforementioned statistics from the pilot, this program has resulted in both business and social value. Namely:

  • 4:1 Return on Investment
  • 5% Reduction in recidivism
  • 22% Lower daily average jailed
  • 25% Reduction in behavioral health crisis costs
  • Proactive prevention of repeat offenses
  • More efficient use of community tax dollars
  • Lives saved by preventing suicides and breaking the high utilizer cycle
  • Clinicians given access to important predictive mental health factors in real time
  • Clinicians empowered by clinical decision support tools with evidence-based best practices

Through our engagement with key business stakeholders, EV Technologies successfully addressed a business and social problem with technology. Our successful prototyping and creation of a set of predictive models improved the quality of service clinicians are delivering to clients and provided HarrisLogic with a roadmap and vision for future iterations that can scale infinitely with their data. Thanks to these successes, EV Technologies will be at the forefront of HarrisLogic’s next great innovation in the fight to help save lives by applying their technology to suicide prevention with the Zero Suicide Institute.

The 2018 SAP Innovation Awards has recognized the success of this program with an award in the category of Digital Transformation. EV Technologies and HarrisLogic will be speaking at the annual SAPPHIRE NOW +ASUG conference on Wednesday, June 6, at 11:30 ET. If you’ll be attending and are interested in hearing more about how we used SAP HANA, SAP Predictive Analytics, and SAP BusinessObjects to transform behavioral health care, be sure to add our session to your agenda.

 View our case study or watch the presentation from SAPPHIRE NOW  for more technical detail on how we harnessed the power of SAP to improve the face of behavioral health care.

 

One thought on “Transforming Behavioral Health Care With Predictive Analytics

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.