Beating The Stats

Getting a bead on what is a “good” outcome from mental health treatment is difficult. A recent study found that 61% of youth with primary anxiety disorder treated with long-term cognitive behavioral therapy (CBT) showed total remission. Is this a “good” outcome?

In his presentation, Unpacking The Executive’s Toolbox For Building Collaborative Care Models With Kaiser Permanente at the 2024 OPEN MINDS Operational Excellence Institute, Stuart Buttlaire, Ph.D., the Regional Director Of Behavioral Health and Addiction Medicine of Kaiser, presented three different behavioral/primary care collaborative care models tested at Kaiser. The collaborative care for consumers with SMI, which has a pharmacist as the lead clinician, found that optimal psychotropic medication adherence was up 6% to 66% of consumers, there was a 6% decrease in psychiatrist visits, and over 75% of the consumers attended an intake visit and at least one follow-up visit.

Kaiser’s intensive community treatment model, focused on consumers with health-related social needs, results in a 26% decrease in emergency department visits, a 34% decrease in inpatient stays, and a 0.8-day decrease in inpatient psychiatric length of stay compared to one year prior to the start of the program. The question is what is good? Architect William McDonough famously said “Don’t let perfect be the enemy of good”–words to live by in health care. And in my closing keynote at that institute, The Excellence Gap In Health Care, Why It’s Great To Be Good!, I pointed out that since health care system performance is not great, there is a great opportunity for provider organizations that can demonstrate ‘better’ performance. 

We had a look at two organizations investing in initiatives to improve performance at the 2024 OPEN MINDS Performance Management Institute. In the session, Harnessing The Power Of Predictive Outcomes & Analytics In Measurement-Based Care, executives of two provider organizations—Adelphoi Innovation Solutions and Devereux Advanced Behavioral Health presented their approaches to data-driven outcomes management. The session featured Shawn Peck, Vice President, and Karyn Pratt, Vice President, Marketing & Strategy Development of Adelphoi Innovative Solutions; and Dr. Patricia Hillis-Clark, Psy.D., Executive Director, and Martha Zhan, National Finance Adviser of Devereux Advanced Behavioral Health.

Adelphoi Innovative Solutions is a $70 million nonprofit organization headquartered in Latrobe, Pennsylvania. They operate 40 different service lines in 30 counties focused on at-risk children and families, juvenile justice, child welfare, and behavioral health education. Devereux Advanced Behavioral Health is one of the country’s largest nonprofits with annual revenue of over $516 million. They serve children and adolescents with autism and intellectual developmental disability (I/DD) services, as well as provide youth and family services at more than 250 treatment locations in 11 states.

Adelphoi has deployed a predictive analytics solution to children with the right treatment and the right program. They are using the platform, FirstMatch, to facilitate decisions about individual consumer care and formalize treatment planning. With the platform, they are now able to track consumers, collect data, and measure outcomes.

Ms. Pratt noted that the platform has improved outcomes for their clients, but it has also helped them identify incompatible treatment matches (children they should not be trying to help), monitor fidelity of treatment programs, and measure program quality. During the past two years, they have seen a 13% increase in both their program completion rates and in their remain-out-of-care rates.

Devereux has an organizational approach to outcome improvement–their Integrative Outcomes Reporting Model. Their platform combines core assessments and metrics across three key areas: clinical and educational, care and compassion, and safety and quality. The Devereux platform uses a data-driven risk stratification algorithm to match youth to therapy resources. It also tracks and reports incident data, providing a clear and comprehensive view of consumer safety incidents in real time filtered by state, center, incident, time-period, and program area.

Both organizations had “good” results that measurably drove better business outcomes. Ms. Pratt said that although it took Adelphoi a long time to build their model, they are “seeing the fruits of all of this work.” Mr. Peck referenced the “garbage in, garbage out” axiom to stress the necessity of accurate data. “Whatever you put is what it’s (the machine learning) going to work with,” he said. He added that it takes time to create a validated database to make the switch from macro to micro-outcomes. “Without data analysis, I’m spit-balling…and shooting in the dark.”

Ms. Zahn said at Devereux they are improving treatment with some of the most complex cases using predictive analytics. Also, by being transparent about their performance data with payers, they have been able to increase their reimbursement rates by 25% over the past few years. And the data-supported, evidence-based practices are positioning them well for value-based reimbursement. She said, “(The data analytics) we have implemented have tremendous operational and financial implications. We’ve used the data to drive business results.”