Healthcare Analytics, Population Health Management, Healthcare Big Data

Analytics in Action News

How can predictive data analytics improve medication adherence?

By Jennifer Bresnick

- Medication adherence is one of the most complex and costly problems facing the healthcare system today.  With approximately one third of patients failing to take their medications as prescribed, wasting $290 billion a year in avoidable costs, primary care providers have turned to clinical data and predictive analytics to figure out why patients are skimping on their pills and how to stop them.

While automated telephone reminders and letters in the mail are common tactics employed my managed care organizations (MCO) to urge appropriate refills and chide patients into compliance, they don’t work quite as well as healthcare providers could hope.  How can harnessing the power of predictive analytics produce targeted interventions that provide a meaningful effect on the problem of non-adherent patients?

According to a survey of managed care organizations conducted by AllazoHealth, the vast majority of MCOs are currently targeting patients experiencing heart failure, hyperlipidemia, hypertension, and diabetes for medication adherence interventions.   More than half are planning to expand their programs to include COPD, asthma, and depression.  Trigger-based reminders, sent out to patients when they are late to filling a prescription, for example, are currently in widespread use, but a mere 7% are currently using predictive analytics technology that prioritizes patients who are at significant risk for becoming non-adherent to their treatment regimens.

“Predictive analytics has shown, very concretely, that it is most effective, in terms of driving a greater change in medication adherence for patients,” said Clifford Jones, CEO of AllazoHealth, to HealthITAnalytics.  “It can be used in conjunction with trigger-based targeting.  But you have to make sure that you’re always capturing the data and measuring the impact of interventions to drive continual improvement.”

“When you roll out one of these intervention programs, you actually end up intervening with patients multiple times,” he explained.  “You need to be really careful to ensure that you track that data, and can differentiate that impact, so that you continue to make your program better over time. The reality is that a lot of programs aren’t doing that.  The MCOs that we talk to, by and large, don’t know which interventions are being effective for their patients.  It’s difficult, because you have to control the way that you decide who to intervene with, and the way that you track that data.”

Only a third of MCOs believe that their adherence programs are very effective.  The remaining organizations think they’re only having a “moderate” impact, failing to intervene with patients before they become non-adherent and spending far too much money on interventions for people who don’t actually need them.  Sixty-percent of MCOs participating in the survey said they plan to expand their predictive analytics platforms to help them with the process while trimming excess costs.


Automated phone calls and direct mailings are popular initiatives, but many MCOs believe they are ineffective for the patients most at risk of falling behind on their prescriptions: the elderly and patients of lower socioeconomic status.  “For our elderly patients, automated reminders don’t work,” said one small MCO.  “They prefer personal interaction with the pharmacist.”

“Not all interventions work for all patients,” agrees Jones.  “Some patients won’t respond well to any intervention, and some patients will respond better to anything.  That’s where it’s really important to have effective targeting in terms of who you’re delivering those interventions to so that you can have the biggest impact.  You don’t have to roll out every type of intervention under the sun.  You could start with a more focused set of interventions, especially if you want to do things in-house and expand your capability over time.  But it’s important to be really targeted and structured in the way that you determine who to intervene with.”

“Whether or not you view this as the patient’s responsibility or the provider’s responsibility, it’s costing a lot of money for anybody who’s taking on risk for those patients,” he added.  “So you have to do everything that you can.”


Join 25,000 of your peers

Register for free to get access to all our articles, webcasts, white papers and exclusive interviews.

Our privacy policy

no, thanks