How did we create healthcare's first, proven referral system?

Sure, I'm slow. It took us 30+ years to turn the Treatment Outcome Package (TOP ©) into a powerful, predictive assessment tool and outcome delivery system. I'll show you how to turn yours (like maybe the PHQ-9 and GAD-7, you think?) into something that can deliver the same eye-popping, best-therapist-predicting system like ours. With unlimited resources it should take about 4 years of R&D.

1)    MULTI-DIMENSIONALITY (9 months of R&D)

Ensure your assessment is multi-dimensional and has the confirmatory factor analytic construct validity statistics that meet or exceed TOP’s stats presented in multiple peer-reviewed journals. You’ll need multiple samples of tens of thousands of patients across all levels of care. Remember, I said unlimited resources...

Here is a table from one of our peer-reviewed papers. You'll need your system to have numbers that are as good or better than ours. HINT: Nothing in the behavioral health outcome field has.

Why it’s important

Without exceptional validity you cannot predict the future (like predicting with what subgroup of therapists a patient will have the best outcomes, or which clients are at the highest risk of hospitalization). Those who take short cuts like cobbling together the PHQ-9 and GAD-7 will fail. It's already been tried. Below is a quote from the peer-reviewed results:

“The results of exploratory and confirmatory factor analyses, as well as the Omega hierarchical test, suggested a unidimensional construct such that both instruments were part of a common general factor, as the subscales of PHQ-9 and GAD-7 separately had negligible independent effects to explain the variance of the construct.” (Teymoori, Gorbunova, Haghish, et al., 2020)

In English, this means that the PHQ-9 and GAD-7 are measuring the SAME thing. There is no multi-dimensionality gained by using both tools. That's a lot of questions to measure one thing!

2)    RISK-ADJUSTMENT (14 months)

Next, take your new multi-dimensional tool and risk-adjust (RA) it. Why? The consequences couldn’t be starker. Without risk-adjustment, what looks like poor care (e.g., minimal change ) could, in reality, be exceptional care (e.g., a client with multiple previous psychiatric hospitalizations on a slow, deteriorating course). If you are in the business of delivering or managing care, there may be nothing more important! You don't want to discourage your providers from refusing to take the harder cases -- those that need your help the most!

These RA models need to assess the contribution to “variance explained” from both patient and provider influences. In 2016, our previous models explained an exceptional 53% of future, end-of-treatment outcome variance (see table above). Our latest models can explain up to 71% of the variance. No other system has that much predictive power and is why our POD Grade ® Matching system (see JAMA Psychiatry publication) roughly doubled care from providers whose outcomes were already impresive.

3)    SCORE CARDS (6 months)

Next, with your new tool and RA models, create algorithms to fairly compare providers on each outcome dimension. You’ll need to update these score cards for each provider monthly as our published research shows that it’s the most recent outcomes from each provider that predict their near-future skill. We've done this for multiple state child welfare departments, creating score cards on each of their deep-end providers, and for solo practitioners who treat the worried well, and most every provider type in-between. We have yet to find the proverbial super-shrink who's exceptional at everything. We all have our strengths and weaknesses. I'm, for example, below average at treating substance abuse. I earned those stripes by having little training in it, and almost no real-life experience (as my family were mostly Mennonites). We all have our blind spots, our warts, and those things that make us unique and exceptional.

Above is a real report card for the State of Connecticut's child welfare providers. Provider "G" is exceptional at treating Depression, Psychosis and Sleep issues. But, do we really need outcome data to tell us this? Of course we do...

The problem is that we are all very poor at guessing what we are good at. How could we really know? It requires data on how well we perform compared to other providers! That's what outcome data and the feedback it provides delivers.

4)    MATCHING ALGORITHMS (18-36 months)

Next, with scorecards created, various matching algorithms need to be tested to see which ones produce the best results. This requires a randomized clinical trial like our three-year federal PCORI study which compared the outcomes of five different TOP matching algorithms.

Now that you’ve taken 4+ years to create a system as powerful as TOP, you’ll need to wait. US Patent No. 7,873,525 precludes others from using or selling a matching system like ours until 8/20/2028. Until then, we’d love to have you as a customer! 

Visit us at www.OutcomeReferrals.com or www.MatchedTherapists.com for more information.