A new way to boost medication adherence: reading patient “vibes” with AI
Two-thirds of Americans take prescription medications, but about half of them won’t follow their treatment plan correctly within the first year. Poor medication adherence (e.g. patients not taking their pills as prescribed) cost the U.S. healthcare system $500 billion annually and contribute to countless preventable hospitalizations.
Smart pill bottles, reminder apps, and calendar alerts have all tried to improve medication adherence, but they’ve mostly fallen short by focusing on forgetfulness, a symptom, rather than its common root cause: emotional barriers.
The American Medical Association (AMA) has identified that four of the top eight barriers to medication adherence are emotional in nature. Fear, anxiety, and feeling overwhelmed after receiving a diagnosis or complex treatment plan can drive patients to abandon their medications more than forgetfulness.
Enter Pleio, a company specializing in medication adherence, which recently released OLLIE, an AI system designed to identify these hidden emotional obstacles through conversation analysis.
“By analyzing the ‘vibe’ of patient conversations, OLLIE helps our GoodStarters address the underlying issues of fear, confusion, and isolation that often drive non-adherence,” Michael Oleksiw, CEO of Pleio, shared in an email. “When we combine OLLIE’s emotional intelligence with genuine human connection, we’re not just improving adherence statistics—we’re fundamentally changing how patients experience their healthcare journey, creating support that resonates with their lived reality rather than just their clinical profile.”
AI-enhanced human behavior improvements
Oleksiw emphasizes that OLLIE doesn’t replace human interaction, it enhances it. The system works behind the scenes during conversations between patients and Pleio’s “GoodStarters,” trained peers who provide support to patients starting new medications. As these conversations happen, OLLIE analyzes cues that might indicate emotional barriers to adherence.
“OLLIE excels at identifying emotional cues that traditional analytics might miss. Take a patient dealing with a chronic condition like diabetes. They might express subtle signs of emotional fatigue — slight hesitations, micro-expressions of frustration, or language that suggests diminishing hope. At times they may feel lonely, isolated or even over-stimulated, orover-reminded,” Oleksiw explains. “In a hypothetical interaction with a GoodStarter, OLLIE might detect that a patient’s reduced medication adherence isn’t just about forgetfulness, but stems from a deeper emotional overwhelm. OLLIE could flag that this patient might benefit from motivational support, maybe a positive affirmation, less information or perhaps a more empathetic communication approach.”
Finding friction points and tipping points
One of OLLIE’s key capabilities is identifying what Pleio calls “friction points” and “tipping points” in patient experiences, obstacles that impede adherence and moments that could lead to positive or negative behavior changes.
“A friction point might be the overwhelming complexity of medication schedules,” Oleksiw offers as an example. “OLLIE could identify that the patient’s language suggests increasing anxiety around managing multiple medications.”
These insights allow GoodStarters to provide personalized, timely interventions before patients abandon their treatment plans.
“A tipping point might be when this anxiety transforms into potential dis-engagement,” he continues. “OLLIE could provide an actionable insight like: ‘Patient exhibits high anxiety around medication complexity. Recommend: Simplify instructions, offer digital medication tracking support, connect with a care coordinator who can provide personalized guidance.'”
He emphasizes that their approach maintains human oversight throughout the process. “Our AI never acts autonomously or automatically,” Oleksiw stresses. “We always have a ‘human in the loop.’ AI learns and suggests, but our humans oversee and implement.”
Related Content: CompassionIT – Starting with a May 20 virtual summit (available now on demand) on embracing innovation with empathy, DHI’s special focus on the evolving relationship between compassion and technology in healthcare includes articles and special events.
Empathy at scale
When asked how Pleio plans to measure OLLIE’s success, Oleksiw introduces an intriguing concept: “We would love to see success measured as ROE—return on empathy.”
For now, they track more conventional metrics, including patient satisfaction through “positivity scores” and engagement measurements like call depth and opt-in extensions. They also monitor traditional medication adherence metrics such as persistence, proportion of days covered, and increase in days on therapy.
Oleksiw envisions a future where AI-powered conversational analytics create what he calls “empathy at scale.” the ability to provide personalized, emotionally intelligent support to patients at a system-wide level.
“We envision OLLIE evolving to not just analyze conversations, but to help predict potential patient challenges before they emerge,” he says. “Imagine AI that can help customize patient support journeys, identifying unique emotional needs and connecting patients with precisely the right type of support at the right moment.”
Despite his enthusiasm for AI’s potential, Oleksiw maintains that technology should augment rather than replace meaningful human connections in healthcare.
“We do see Ollie automating human connection fully delivered by technology (chatbot) to address a patient’s transactional needs with a high degree of efficiency and personalization. At the same time, we believe that only a true human connection can deliver emotional intelligence to meet the uniquely human emotional needs of patients,” he says. “Our vision isn’t about replacing human connection, but making that connection deeper, more meaningful, and more accessible.”
In a field where the adoption of new technology continues to accelerate, Pleio’s implementation of AI is an important reminder that the most impactful advancements are often the ones that make it easier for us to communicate with each other.