Can Medication Adherence Be Predicted?

Jan 17, 2013
Express Scripts has developed a set of predictive models to predict medication nonadherence. More than 400 factors are used to identify who's most at risk.
Tags
  • Accountable Care Organization (ACO)

Answer: Absolutely. As long as you have enough data and you know where to look.

Last week, Express Scripts announced our new solution to improve medication adherence, ScreenRx(SM). The engine behind ScreenRx is our set of predictive models that identifies in advance which patients are most at-risk for falling off their prescribed therapies.

Think for a moment about the data we receive when an Express Scripts member fills a prescription. We know quite a bit about that patient—where he lives, who his physician is, what pharmacy(s) he uses, who else is on his benefit plan, his gender, marital status, etc. We also know a fair amount about the pharmacy—where it’s located, how adherent its customers are, and how effective it is at pointing its customers to lower cost medications. And of course we know about the medication—how it interacts with other drugs the patient may be taking, how affordable it is, and what other patients’ experiences with it have been.

As you can see, there's a mountain of data—more than 400 key factors—in just a single prescription fill. And when we compare these same factors across the hundreds of millions of prescriptions we manage each year, we start to see patterns that enable us to predict nonadherence before it even occurs.

So which characteristics correlate with medication adherence?

Among other things, you are more likely to follow your doctor’s orders if:

  • You’re a woman
  • You’re over the age of 65
  • You don’t have young children in the house
  • Your medication is delivered directly to your home
  • Your copay is less than $0.50 per day
  • You live with a partner (bonus points if your partner is adherent)
  • Your gender is the same as your physician’s
  • You have a history of being adherent to other therapies

It should be noted that no single factor by itself is all that predictive. But when we look at the entire constellation of 400 factors associated with a single prescription, we get a pretty clear picture of which patients might need extra assistance to stay on a path to better health.

...I’m looking at you, young fathers.

comments powered by Disqus