Author: Eric Grossman,
CEO of NextHealth Technologies

eric-grossman-picAre predictive analytics a waste of your company’s money? Yes and no…

(Hint: they don’t have to be if you insist on accountability for measurable outcomes.)

Predictive analytics are a capital “B” or a billion-dollar annual investment in health care today with many C-suite executives writing blank checks and still waiting on return on investment or improvement in health outcomes. So, for the majority of investments, the answer is resoundingly “yes” — predictive analytics are a waste of money or at least have a “let’s wait and see” outlook.

Many investments in predictive analytics produce a “list” of who to target to close a gap in care. These lists are predominately focused on high risk members. This is how the story goes – first invest heavily in an enterprise data warehouse over 12–16 months to ensure the organization is ready to manage disparate and longitudinal data sets. Concurrently, invest in a consulting firm with a set of PhD’s to produce model to predict what will happen next. Now assuming you have the “list”, it is handed off to your care managers to close gaps in care. Six months or a year later, the organization measures closed gaps in care before and after the intervention. The business owner asks, “Did it work, and how do you know?”, but there isn’t a good answer. Does this sound familiar? If so, you are not alone.

There are major strategic flaws with this approach. For one, there is absolutely no scientific way to prove that your interventions caused the closed gap in care. The outcome improvement could have been as a result of your interventions, or perhaps due to an increase in the seasonal temperature, or maybe a benefit change that led to a drop in emergency room visits. Therefore, continued investments are no more than an educated guess with no true measure of cause and effect. Second, given the investments are not typically housed in a platform, they are limited in their ability to scale to cost effectively support new use cases or populations.

Predictive analytics are not a waste of money if they are seamlessly integrated with these four “no regret investments”:

  1. Causality — causality allows an initiative to determine and amplify what truly is causing behavior change.  Causality is established using randomized control trials (“RCT”) — the gold standard of cause and effect.
  2. Prescriptive analytics —advanced algorithms are required prescribe what actions to take, when, for whom, and in what frequency based on results from the RCT’s and other machine learning techniques. This type of analytics is essential to optimize limited resources and programs where they are making a measurable difference.
  3. Behavioral economics — creating a more personalized approach requires an in-depth understanding of the types of words that will resonate with your consumer. Deploying carefully crafted messages or “nudges” and measuring the effects are critical to effectively and positively changing consumer behavior.
  4. Integrated solution / governance — orchestrating an outcome or consumer behavior change is difficult and requires an executive governance model and system design aligned around a business outcome. Once an outcome is chosen – e.g., avoidable emergency department reduction – the system and governance model must integrate data ingestion through multichannel consumer engagement in a closed loop manner. Unless the solution is integrated, it is almost impossible to change outcomes and importantly, scale to support new use cases.

Are these four incremental investments really “no regret”? You bet, if orchestrated correctly. Health plans and providers can measurably improve outcomes and offer a truly personalized and “frictionless” member experience.

Today, NextHealth Technologies is in the rare air of companies that are delivering scientifically proven outcomes — in multiple cases, a 25% reduction in targeted outcomes and medical costs after only 10 weeks of consumer engagement. NextHealth even puts our fees at risk based on outcomes because we can measure “lift”. That’s an investment worth making.