AI Can Predict Which Members Will Struggle, But Can It Change Their Behavior?

Published on August 27, 2026.

Healthcare organizations are swimming in member data. With AI and predictive analytics, it’s much easier identify patterns in that data that suggest members at risk of diminished performance, such as a missed refill or skipped doctor’s appointment.  

But prediction is only the beginning.  

Knowing a member may be likely to struggle doesn’t tell a plan why. And it doesn’t help with understanding what will motivate that individual to complete a healthy behavior, like filling and taking their medicine.   

For health plans looking to improve specific behaviors, the bigger opportunity is connecting what technology can predict with what we know about human behavior—then using both to make it easier for members to take the next step toward healthier outcomes. 

What can AI tell us about medication nonadherence? 

AI can help health plans make sense of large amounts of member data and recognize patterns that may be difficult to identify at scale. These insights can help identify members whose adherence may be declining, prioritize those who could benefit most from intervention and inform the timing, frequency and channel for outreach. 

That’s a helpful start. Plans can identify emerging risk and intervene sooner, before adherence dips below 80% PDC. 

But a risk signal is not an explanation of behavior. 

As AdhereHealth Chief Experience Officer, Dr. Chandra Y. Osborn, explains in a recent AI Journal article, “AI can help determine who to engage and when. Behavioral science helps determine how to engage them in ways that actually work.” 

Simply knowing who’s at risk doesn’t solve the problem 

Consider two members who have both stopped refilling their medication. On paper, their behavior may look nearly identical. However, the same adherence signal can represent very different circumstances. 

For example, one may have experienced side effects and decided the medication isn’t worth taking. Another may have moved, changed pharmacies, started caring for a family member or simply fallen out of a routine that previously helped them take their medication consistently. 

That’s why identifying risk is only one part of an effective adherence strategy. Plans also need a way to understand what is driving that risk for the individual member. 

How does human interaction boost effective technology? 

Human interaction helps uncover what the data cannot. 

A conversation with a member can reveal challenges that aren’t readily visible in claims or other data. These can range from logistical barriers such as a lack of transportation to the pharmacy, to behavioral challenges like confusion about why a medication is still necessary.  

As Dr. Osborn shares, “Technology can flag these situations. But only human conversations can uncover and address them.” 

Once a barrier is understood, the response can become more meaningful. That might include education, connection to financial or social resources, refill coordination, pharmacy support or communication with a provider. 

What actually changes medication-taking behavior? 

It’s critical to understand the barrier a member is facing. But even then, telling someone what they should do may not be enough to change what they actually do. 

Taking medicine happens within the context of everyday life, where caregiving responsibilities, mental health, financial stress and other factors are competing for attention.  

Behavioral techniques can help address those realities by focusing support on how people make decisions and what makes an action easier or harder to complete. 

An intervention should fit the barrier rather than expecting the member to fit the intervention. 

And because barriers change, behavior change cannot be treated as a one-time achievement. Continuous monitoring helps plans recognize when a previously adherent member begins showing new signs of risk and create another opportunity to understand what changed and provide support. 

What should health plans consider with AI-enabled adherence strategies? 

As AI becomes a larger part of healthcare operations, the strongest strategies, especially around medication adherence, will connect predictive insights with behavioral science, clinical expertise, human engagement and ongoing monitoring.   

AI can make it easier to see where strategies may break down. The greater opportunity is using that insight to understand what is standing in a member’s way and make the next action easier to take. 

Gather more insights to drive adherence and outcomes-based plan performance 

In her article with the AI Journal, Dr. Osborn, AdhereHealth CXO, explores where AI can—and cannot—make a difference in medication adherence. She explains how AI can work behind the scenes to strengthen human engagement, integrate behavioral science and help make the right action easier for members.