• Center on Health Equity & Access
  • Clinical
  • Health Care Cost
  • Health Care Delivery
  • Insurance
  • Policy
  • Technology
  • Value-Based Care

AI, ML Can Mitigate Risks, Improve Efficiency Across Managed Care Space

News
Article

A session at AMCP Nexus on the use of artificial intelligence (AI) and machine learning (ML) in managed care highlights some real-world examples of successful implementation of these technologies.

guy pointing at AI

Artificial intelligence (AI) can be used in many ways across the managed care pharmacy space, including managing new initiative implementation, monitoring pharmacy claims, and more.

A session at the Academy of Managed Care Pharmacy (AMCP)’s annual Nexus meeting showcased some of the practical uses of artificial intelligence (AI) and machine learning (ML) in the managed care pharmacy space, including some real-world examples of successful implementation of these technologies.

Nick Trego, PharmD, senior vice president of clinical analytics and client services at HealthPlan Data Solutions, began the session with an overview on what AI is, how ML is different, and what major technologies in our society use AI regularly—from personal assistants, like Apple’s Siri, to chatbots, like ChatGPT, to smart home devices, like smart thermostats.

“The easiest way to think about machine learning and AI is AI is like physics. It's like the whole subject; the theory and the methods. And then machine learning is a subset like Newton's law. It's an application of AI—like an array of physics—but it's not the whole AI,” he explained.

Jessica Hatton, PharmD, BCACP, associate vice president, pharmacy, CareSource, chronicled some of the major ways that AI and ML can be applied to managed care pharmacy. For instance, AI can be used for:

  • Pharmacy benefit manager (PBM) contract reading and interpretation
  • More efficient monitoring of pharmacy claims
  • Member engagement and customer service
  • Enhancements for predictive modeling for targeted outreach
  • Formulary management

The speakers went over 4 real-world case studies of AI/ML in managed care practices, outlining the problems that the technologies mitigated and the overall impact they had on the efficiency of the practice.

The first case study was the implementation of an ML model to autonomously identify errors for insulin claim adjudication at the invoice level. The model was a response to several states looking to establish laws for insulin co-pay caps as a means to lower patient out-of-pocket (OOP) costs. The model was able to quickly correct insulin co-pay errors, remain in compliance with state law, enable money returns directly to members that were being overcharged, and ensure proper coding for the system to prevent future errors.

Case study 2 regarded National Average Drug Acquisition Cost (NADAC) pricing methodology. In light of a new initiative to implement NADAC in the pricing logic for independent pharmacies, an AI/ML model was created to monitor implementation of the initiative and identify discrepancies in reimbursement. The model was able to capture $1 million in avoided costs and mitigate risks associated with program implementation.

The third case study involved an AI/ML model that was trained to monitor a new current exposure method limit on a National Drug Code (NDC) level. The model ensured adherence to new plan designs, avoided unnecessary costs resulting in over $25,000 in plan savings, and ensured a continuous alert system for high-cost NDCs was in place to prevent future errors.

The final case study was about COVID-19 testing mandates. During the pandemic, CMS implemented coverage for a maximum of 8 OTC COVID-19 tests per month at no OOP cost to the patient. Practices were given short notice to deploy new changes for the entire Medicaid population. An ML model was trained to monitor new COVID-19 testing logic to autonomously identify discrepancies in reimbursement. Although the model didn’t detect any errors, it flagged 1 pharmacy to abnormal utilization, ensuring that the COVID-19 initiative was being followed as intended and minimizing waste.

Trego concluded that although AI and ML have the potential to transform the managed care space, concerns about model accuracy are valid as some AI technologies, like ChatGPT, have been known to give wrong information on occasion. He recommended still having a human double checking that the AI/ML model is working correctly.

“[AI] is still not replacing human intelligence,” Trego said. “So, I think to make sure whatever you intended to uphold makes sense, [you should be] double checking it. We do it internally just to make sure that the system is working the way you expect it to.”

Related Videos
Screenshot of Adam Colborn, JD during an interview
Screenshot of Susan Wescott, RPh, MBA
Screenshot of an interview with Adam Colborn, JD
Screenshot of Stephanie Hsia, PharmD
Screenshot of an interview with James Chambers, PhD
Screenshot of an interview with Megan Ehret, PharmD
Interview screenshot with Megan Ehret, PharmD
Screenshot of an interview with Susan Wescott, RPh, MBA
Screenshot of an interview with Nadine Barrett, PhD
Female doctor in coat with stethoscope on blue background - Pixel-Shot - stock.adobe.com
Related Content
© 2024 MJH Life Sciences
AJMC®
All rights reserved.