5, 50 and The Role of Precision Targeting in Care Management

5, 50 and The Role of Precision Targeting in Care Management

by Michael Hollenbeck, Proskriptive CEO

Let’s say for a moment that you’re watching your favorite football team play their arch nemesis.  The stakes are huge, and team pride is on the line.  The ball is snapped, the wide receiver goes long, the quarterback drops back, the blockers hold off the defense… and then the quarterback turns to the sideline and throws a 40 yard pass to a fan in row 29 of the lower deck of the stadium.  “What the heck!?” you think to yourself. Now imagine this happens 8 of 10 times the quarterback is given the ball.  Frustration, confusion, bewilderment and even anger would be likely emotions for anyone watching such odd behavior.  The quarterback’s actions are just so contradictory to the goal of winning the game, it doesn’t make sense.

Back in the healthcare world we know that 5% of members account for 50% of the total healthcare spend.  If we try to lower these costs by managing historically high utilizers with the expectation that we’re preventing tomorrow’s costly problems, then like the quarterback, we are merely hurling away any hope of winning our own very important game.
Here is another critical number to consider, 80.  If the care management strategy for your population health program is anchored to managing historically high cost members, you are throwing away 80% of your opportunity to bend the cost curve.  An internal study of 3 separate managed cohorts showed that on average only 20% of the previous year’s high utilizers were represented in the following year’s high utilizers.  I’ve heard statistics cited that range between 15% and 40% for the same analysis, but in all cases, we’re hurling far too many opportunities to deliver impactful interventions into row 29 when our members are depending on us.
Care managers guided by precision member targeting technology, on average, reduce costs by more than double their peers managing historically high utilizers, or those with polychronic conditions.  That’s approximately $4,000 additional dollars per managed member each year.
As you look for ways to reduce costs and increase productivity, improving your care manager’s ability to reach out to the most impactable members before they become high utilizers may just deliver ROI numbers everyone can get behind.  Your members will enjoy improved health AND decreased healthcare costs.
Touchdown!  Now, that’s way more satisfying.

6 thoughts on “5, 50 and The Role of Precision Targeting in Care Management

  1. D'Mar Phillips says:

    Hi Michael,

    What do you think works? Historically, our healthcare system is treatment based more so than healing based. Do you think that is a possible area to review. For example, diagnosis of diabetes leads to treatment not healing of diabetes. Just as an example to solicit your opinion. Thank you.

    1. Proskriptive says:

      Hi D’Mar,
      Thanks for your comment, and I’ll offer my opinion for what it’s worth. People and organizations usually do what they are incented to do. Payment rules for the US healthcare system rewards providers for treating the sick and afflicted vastly more than it does to focus on preventing sickness and affliction. This leads to a very procedure motivated system vs. one that innovates around preventing and intervening as early as possible to avoid unneeded costs and suffering.
      At Proskriptive, we’ve shifted our focus from delivering products for healthcare providers and payers to innovation on behalf of employers to help identify and prevent unnecessary utilization of healthcare services. Employers incentives are far more aligned with their employees’ wellbeing than with the profit motivations of payers and providers. There are many wonderful diabetes programs that concretely reduce costs substantially while providing a far better quality of life for the individual and a much more productive employee for their employers. If we can identify diabetics or pre-diabetics and care for them in the right setting then consumers of healthcare services win.


  2. Michelle Stone-Moore says:

    Great answer Michael! So, you target diagnosis specifically or risk of based on surrounding trends for that population? Or are you driven by socioeconomic and /or ethnic morbidity trends? I would imagine that there are many factors you can involve and filter out, but the trick is picking the most impactful. Thoughts on this? Michelle

    1. Michael Hollenbeck says:

      The answer to the data feed question is a good, better, best, situation. Much is made of social determinant data (for good reason) but you don’t need sd data to get a great start on impactability modeling that can improve each of your care manager’s cost savings per member substantially. When we introduce sd data to the mix you can definitely improve the accuracy of impactability scores as well as delivering more tailored interventions that will drive better outcomes.

  3. Michelle Stone-Moore says:

    Hi Michael,
    I really loved this post and agree thst by only targeting 20% of the population the goal of true population health management will never truly be met. Is there a “magic bullet” that can truly cover each people group? Interested in your thoughts,

    1. Proskriptive says:

      Thanks for the question Michelle,
      As a vendor in the space I should probably say “yes, there’s a magic bullet and it’s us” but there really is no magic bullet per se. The good news is that it’s possible to do a lot better with data assets you probable already have. This is how we’ve been able to yield a great deal of improvement for our customers. The first step is to identify a specific diagnosis or condition set that you’d like to target for improvement. We’ve been at this a while so we have a quality library of risk models that are highly tuned to identify “risk of prospective utilization” for a given condition set.
      Where our approach differs is that we’re not just looking to measure future risk, we’re trying to identify risk we can actually impact. To that end we create a 2nd “priority” model that analyzes each members impactability within that risk spectrum. This measure scrubs their history to see if the condition is 1)clinically impactable and then 2)estimates a $ opportunity value to measure the delta between what their utilization costs would be with and without intervention.
      There are many other tricks of the trade that we utilize to improve targeting accuracy, but we’ve seen great success with this approach.


Got Something To Say:

Your email address will not be published. Required fields are marked *