The Benefits of ML, AI Use in Managed Care Pharmacy | AMCP Nexus 2023

Adopting evolving computer system tools like artificial intelligence and machine learning in managed care pharmacies have resulted in efficiency when addressing the challenges they are faced with, according to Jessica Hatton, PharmD, BCACP, associate vice president of Pharmacy at CareSource and Nick Trego, PharmD, senior vice president of Clinical Analytics and Client Services at HealthPlan Data Solutions, Inc.

The interpretation and reading of pharmacy benefit manager contracts, as well monitoring 100% of pharmacy claims, pose to be some of the biggest challenges in the managed care pharmacy space.

However, adopting evolving computer system tools like artificial intelligence and machine learning in managed care pharmacies have resulted in efficiency when addressing these challenges, according to Jessica Hatton, PharmD, BCACP, associate vice president of Pharmacy at CareSource and Nick Trego, PharmD, senior vice president of Clinical Analytics and Client Services at HealthPlan Data Solutions, Inc.

Both Hatton and Trego presented on these tools and their benefits in the PBM space during this year’s AMCP Nexus conference in Orlando.

Though AI is still developing daily, one of its few subsets – machine learning (ML) — is assisting machines to extract knowledge from data and learn autonomously. These tools together allow users to greatly increase productivity and performance accuracy.

As mentioned, AI and ML are addressing the challenges associated with PBM contract reading.

For example, the volume and complexity of contracts, which can often contain key “adjudication concepts in single sentences,” can be efficiently managed through these technologies, Hatton shared.

Currently, these contracts are done by hand and have a prevalence of being disorganized. These factors lead to avoidance by plans who often limit interactions with contracts overall.

Through experience, the pair shared that integration of AI and ML in contract reading addresses these challenges through:

• Expedited contract review: AI and ML enable the review of entire contracts and amendments within minutes, a vast difference compared to the days or weeks usually required.
• Index key terms or elements: These tools can be trained to identify specific elements within the contract and can find key terms within the document.
• Digital contract repository: AI and ML create a hub for all contracts and amendments, organized by effective date or type.

The next challenge addressed by these tools is the monitoring of 100% of pharmacy claims, a task traditionally associated with a high volume monthly, they’re complex, variable logic and custom plan designs. Currently, PBMs’ methods involve cursory reviews, partial monitoring, or reliance on annual audits.

Hatton and Trego highlighted AI and ML tools are efficient in managing PBM claims through:

• High Throughput Technology Platforms: Which quickly find solutions in 100% of pharmacy claims to ensure accuracy.
• Model with customized plan design: Monitors claims and trains software engine to identify issues from plan setup.
• Near real-time vs. retrospective approach: Identifies errors as they occur versus at the end of year. It also helps avoid repeating unintentional costs by making mid-year changes.
• Fraud, waste and abuse: The tools identify errors specific to FWA by using pre-programmed decision tree logic. This trains the AI model to identify new trends at provider, patient and the pharmacy level.

Hatton also shared real-life situations CareSource experienced where AI and ML solved problems in the pharmacy space.

For example, in an insulin copay caps case study, the implementation of state laws regarding insulin copay limits posed a challenge requiring custom coding changes. An AI and ML system identified errors in copay adjudication, ensuring quick correction, compliance with state laws, and the prevention of future errors.

In another case study regarding COVID testing mandates, the rapid integration of CMS coverage for over-the-counter COVID-19 tests required short notice for coding changes. AI/ML models effectively monitored testing logic, identifying discrepancies and ensuring adherence to the initiative, in result minimizing unnecessary plan costs.

Beyond contract reading and claims monitoring, the pair shared that AI and ML also find use in member engagement, customer service, and predictive modeling for targeted outreach. In addition, these tools can enhance member interactions, connect with the most at-risk patients based on social determinants of health, assist in formulary management and base modeling on a plan’s unique population makeup.