Artificial intelligence (AI) is reshaping health care, including how medications are used. On November 13, 2025, PQA hosted PQA Convenes: Artificial Intelligence in Medication Use Quality to bring together PQA members, health technology leaders, and other stakeholders for a discussion on how AI and machine learning are being used to identify, understand, and engage patients in their medication use.
This is the third blog in a four-part series on the event, and it covers the third session, “Cross-Sector Collaboration and Future Innovation in AI.” Moderator Loren Kirk was joined by Justin Bioc of Devoted Health, Yoona Kim of Arine, Helen Kourlas of Healthfirst, Hiva Pourarsalan of Aetna, Mercy Care Advantage, Richard Waithe of Pyrls, Todd Whitehurst of Sanofi and Elven Xiao of MedWatchers.
This panel featured short discussions from cross-sector collaborations on how AI was used to achieve organizational goals in medication use quality. The collaborations were a testament to how the intersection of AI and medication use quality is rapidly evolving, moving from theoretical concepts to practical, large-scale use. At the core of each collaboration, AI was used not as a replacement for clinical judgement, but as a powerful engine to enhance human efforts and scale quality interventions.
Enhancing Quality, Not Replacing Care
One collaboration between a health care service and technology organization and a large national payer showcased a successful integration of AI to enhance patient-pharmacist interactions. This was achieved by developing rubrics specific to quality improvement areas, such as medication adherence and appropriate use. An AI model was then trained on these rubrics, with human oversight and input, to accurately identify moments in calls where pharmacists could improve their communication with patients.
Importantly, the integration of AI was focused on augmenting, not replacing, pharmacists. This approach enhanced existing human workflows by improving the quality of services provided, a framework that was well-received by the national payor.
Prior to using AI, the quality assurance (QA) process for patient-pharmacist interactions was manual and resource-intensive, with feedback taking several weeks to reach pharmacists. The introduction of an AI-powered, human-vetted QA system enabled all calls to be reviewed, feedback to reach the pharmacists in real time. The immediate, actionable feedback resulted in significant improvements in the services provided and, ultimately, in improved medication use quality scores.
Deepening Support for Holistic Health
A biopharmaceutical company is reimagining traditional patient support programs by shifting focus from single-medication support to a holistic life-management platform that supports people living with diabetes. The platform is built on the understanding that adherence and outcomes are influenced by far more than just medication-specific factors.
This broader approach focuses on all the medications a patient is taking, including those from competitors, to ensure proper use, medication safety, access, and adherence. The platform also integrates support for crucial lifestyle factors like diet, exercise, sleep, and stress, and includes programs for addressing depression and smoking cessation. This comprehensive range of interventions are delivered by AI-augmented health coaches.
By analyzing interactions between health coaches and patients, the organization identified that a significant portion, up to 70% of communications, consist of low-risk conversations, or small talk. Low-risk conversations were found to be time consuming and limit the scalability of valuable clinical interventions.
Blending AI and health coaches’ communications with patients seamlessly in a transparent way increases the opportunities for health coaches to address patients’ health holistically and enables more effective interactions, all while ensuring human expertise remains central to the interventions.
Smarter Targeting and Measurable Impact
Another collaboration between a payor and AI-powered medication management solutions company highlighted the power of AI to do more with less through an improved identification of outreach opportunities. The collaboration began by addressing persistent challenges in medication adherence despite traditional efforts such as increased provider training, addressing cultural and language barriers, and building trust with members. Due to the measurable success during a one-year pilot, the collaboration evolved into a multi-year effort rooted in transparency and shared goals.
A sophisticated AI platform that integrates with existing workflows was used to analyze hundreds of data parameters, including patient cost share, demographics, social behaviors, and clinical history. This analysis was used to predict the likelihood of adherence and identify the specific interventions most likely to succeed in addressing member barriers.
This capability enabled the health payor to focus its resources on identifying and intervening with members with the greatest need. This targeted approach led to substantial impact, reflected in an improvement of upwards of five percentage points in medication adherence scores within the first year. For the payer’s care team, the AI platform enabled better allocation of limited resources. For their members, this increase in medication adherence scores meant improved access to necessary medications.
Reactions
A panel of subject matter experts synthesized and reacted to the successful collaborations discussed, stressing that solutions should never begin with how AI can be used, but rather clearly defining the problem and deeply understanding the existing workflows that aim to address it. AI should be considered as one of many potential tools that may become part of the solution.
Successful implementation of AI-powered solutions starts with integrating AI into existing workflows rather than disrupting them. Implementation also relies on adoption by the care teams, which can be achieved by demonstrating how AI enables them to perform more meaningful clinical tasks and by giving them ownership to override or verify the AI’s outputs. Lastly, transparency with care teams and patients about the data inputs and the use of the outputs is vital to establishing trust.
In summary, this panel established two key points:
- AI is currently used as a tool to augment human efforts rather than replacing them.
- Successful AI implementation rests on clearly defining the problem, integrating AI within existing workflows, securing care team adoption, and ensuring transparency.
PQA Convenes: Artificial Intelligence in Medication Use Quality was made possible by the generous support of Arine, Merck, Pfizer and PQS by Innovaccer. PQA does not endorse, recommend or favor any organization, or its products or services. PQA general funds also supported this event.