Electronic Health Records

Adoption & Implementation News

How can predictive analytics improve medication adherence?

By Kyle Murphy, PhD

- With commercial and public insurers looking to reward healthcare organizations and providers based on patient outcomes, more and more attention is being paid to methods for staying ahead of medical conditions developing into something more serious and costly. Payers are focusing on a number of areas where novel approaches to care coordination and delivery could improve care and reduce costs.

One shining example is medication non-adherence, which costs the healthcare system close to $300 billion annually in avoidable costs and according to Clifford Jones, CEO of Allazo Health, represents an opportunity to demonstrate the potential of healthcare analytics. “Without applying predictive analytics to this, they are just doing more of the same stuff and that hasn’t solved the problem yet,” he says.

The reason for this high cost of medication non-adherence stems from its impact down the line in the form of emergency visits and readmissions that could be avoided through targeted interventions. “First of all, improving medication adherence helps reduce future healthcare costs,” continues Jones, “because each diabetic patient that you get adherent to their medication who wasn’t saves a little under $4,000 per year in healthcare costs, which is a huge benefit.”

Considering the penalties being levied against hospitals for preventable readmissions, focusing on medication non-adherence can prevent medical conditions from transforming into something worse and exposing them to penalties.

Additionally, providers stand to gain in other ways, such as accreditation and quality scores. “A lot of providers, especially ones who run their own pharmacies, are looking for EHNAC certification. Medication adherence is an important part of that. Also, there’s an impact on HEDIS scores,” says Jones.

For insurers, there are also tangible benefits in the form of star ratings, which can dramatically affect their bottom lines:

If you reach different thresholds, then you get a star rating for your plan and consumers choose plans based on those star ratings. Also, there are restrictions on marketing if you don’t meet certain star ratings. And then there are bonus payments that the government pays you for reaching star ratings that can be tens of millions of dollars if you’re a big Medicare plan.

As with healthcare analytics in generally, part of the problem in the way of taking advantage of predictive analytics to improve medication non-adherence involves getting at the necessary patient data, the highest quality of which exists in the EMR system.

“The EMR data is great — it’s just harder to get and it’s more costly to integrate,” Jones explains. “But it does make it better. You’re getting more clinical insight. You’re also getting more real-time things because while pharmacy claims are relatively fast and medical claims can lag some, EMR data can be faster. That’s another key benefit.”

The value of integrating this data, however, is already paying off for accountable care organizations (ACOs) whose organizational structure allows for diverse forms of patient data to come together.

“It helps when their interests are aligned, for example through an ACO,” observes Jones. “The ACO is great because you can get more data with the other parties. With the health plan, you can get the medical claims, pharmacy claims, intervention data, and demographics on the patients. With the ACO, you can also get the EMR data, so you can do more with that.”

While medication non-adherence may appear to benefit mostly individual patients, its impact over time should prove more far-reaching as more and more knowledge about which interventions work for which types of patient. “By choosing the right way to help a particular patient, you can across the population generate a greater uplift in medication adherence,” says Jones.

Continue to site...