Revolutionizing Real-World Evidence: Role of AI in Supporting Pharma Market Access

Real-world evidence is one of the backbones of health data for understanding the safety and efficacy outcomes of medical products, alongside clinical trials. While clinical trials compile data in their own reports and conduct their own analyses of the results, RWE can be more sporadic, widespread, and disorganized, making it very difficult to aggregate for practical use. According to Deloitte, artificial intelligence’s addition to pharma’s data compilation methods has granted companies a simpler way to analyze these large data sets and apply them to the market, saving up to 60% in drug development costs and accelerating time to market access by up to 30%. AI has a strong role to play in accomplishing the insurmountable task of compiling and interpreting real-world evidence in Pharma. [8]

Real-World Evidence and Its Purpose

Real-world evidence consists of data from patients and healthcare providers outside of a trial setting, such as in electronic health records, MedSafety reports, national health surveys, and claims data. RWE provides additional insight via real-world uses that can’t be obtained in a controlled trial. Its data can be used to determine patient outcomes in general or specific populations for pricing and reimbursement and patient-centric approaches for R&D. [5]

RWE helps influence payer negotiations and reimbursement decisions by providing real-world examples of patient outcomes and focusing the reimbursement on data with the most beneficial health outcomes. It also plays a role in formulary placements, as it provides evidence for the use and reimbursement of certain medications in a hospital or clinic. Recently it’s seen more use in regulatory affairs to support the approval of new medication products, since approvals used to mainly rely on RCT data for safety and efficacy. RCTs are also able to use real-world data as a synthetic treatment arm for well understood diseases when trial patients were difficult to enroll [7]. Generally, real-world evidence provides a strong foundation for which multiple aspects of pharma market access are built upon.

However, there are a few problems that can arise with trying to collect this data practically:

• Physical vs. Digital Records: In terms of medical records, not all documents are recorded or kept digitally, and even the ones that can be stored digitally like physician notes may be written in plain language instead of a clear structure, as that isn’t the main priority for healthcare providers when treating their patients. Without full digital information, important pieces of patient information may be left out. [5]
• Multiple Record Keeping Formats: Regardless of this disorganization, some hospitals use multiple health record vendors within their own network, so maintaining structure in one area does not equate to organization on a grand scale.[5]
• Volume: The sheer amount of data that would need to be collected and analyzed to interpret patterns in the desired patient populations can’t simply be handled by a team of experts in a reasonable amount of time. There are multiple data rich sources of real-world evidence that are sizable enough to require their own separate teams to pull useful data from them. [5]

Artificial Intelligence and Real-world Data

Artificial Intelligence provides digital tools with the ability to go through thousands of data pieces in a fraction of the time and analyze them for the desired outcomes with consistency. AI can account for the volume of real-world data in its computing speed. Not only can AI compile the data it's given, it can also be trained to look for patterns or specific outcomes and give overall interpretations of that data in a more organized and digestible format so it can be applied to clinical decision-making and drug development. Because these AIs can be trained to detect these specific types of information, they are strongly tailored to the roles Pharma companies give them. They are forced to follow data protection and security laws as well, which minimizes the need for human interaction and therefore poses a lower security risk for confidential patient information. [4]

Examples of known AI-driven tools for these purposes include ReimbursementAI, MedAffairs AI, ACMA Predict™, and Truveta Data.

• ReimbursementAI: This tool uses market data and policy changes to predict how difficult or time consuming products may be when facing prior authorization requirements in specific regions, allowing manufacturers to adjust their field reimbursement teams and analyze relevant information when building market access strategies. [1]
• MedAffairs AI: This AI trains upon external and internal data for information relating to medical affairs including medical information, regulatory affairs and compliance, and medical strategies. [2]

Conclusion:

Real-world evidence is essential to the research and development of new medications, pricing and reimbursement based on patient health outcomes, and tailoring treatments to target populations. Unfortunately, the lack of a consistent structure between multiple systems of electronic health records and physical documentation, as well as the countless amounts of data points across all sources of real-world data, make it extremely impractical to sort through for the information pharma organizations need for tactical strategies, and it can delay market access for newly approved medical products. Artificial intelligence and machine learning are able to combat these challenges in ways that are fine-tuned to achieve certain goals. Companies are only scratching the surface of what AI can do to support market access.

References:

1. ACMA Life Sciences. Field Reimbursement & Prior Auth AI. Accessed January 19, 2025. https://acmalifesciences.org/acma-ai/reimbursement-ai

2. ACMA Life Sciences. MedAffairs AI. Accessed January 19, 2025. https://acmalifesciences.org/acma-ai/medaffairs-ai

3. ACMA Life Sciences. ACMA Predict. Accessed January 19, 2025. https://acmalifesciences.org/acma-ai/acma-predict

4. Astera. AI and Real-World Evidence (RWE): Extracting Insights from Real-World Health Data. May 2024.

https://www.astera.com/type/blog/ai-and-real-world-evidence/#:~:text=The%20transformative%20effect%20of%20Artificial,patient%20behavior%20and%20treatment%20outcomes.

5. LinkedIn. How Pharma Can Use AI to Leverage Real World Evidence for Data-driven Healthcare. Oct. 2022.

https://www.linkedin.com/pulse/how-pharma-can-use-ai-leverage-real-world-evidence-healthcare-bates

6. Truveta. EHR Data and Analytics. Accessed January 19, 2025. https://www.truveta.com

7. Dang, Amit. “Real-World Evidence: A Primer.” Pharmaceutical medicine vol. 37,1 (2023): 25-36. doi:10.1007/s40290-022-00456-6 

https://pmc.ncbi.nlm.nih.gov/articles/PMC9815890/

8. Pharmacy Times. Data-Driven Development: How Real-World Data and AI Are Transforming Clinical Trials. Jul. 2024.

https://www.pharmacytimes.com/view/data-driven-development-how-real-world-data-and-ai-are-transforming-clinical-trials