In the last few years, AI has evolved from a pipe dream into a world of possibilities and now has transitioned into a catalyst for change. AI will revolutionize how we monitor, document, and analyze information which in turn will change how we approach Government Pricing compliance. But while manufacturers are using AI for compliance monitoring, Government Agencies are doing the same.
AI harnesses the ability to analyze and evaluate large data sets to identify patterns, inconsistencies and outliers making it easier to move from sample-based oversight to data-driven continuous monitoring. Manufacturers can utilize AI to identify and mitigate compliance risks in real time effectively transitioning audit preparation from a reactive, time consuming effort to proactive audit-ready approach. Government Agencies can utilize AI to transition from spot-checking, periodic reviews to continuous monitoring.
Background
Traditionally, if there was a Government audit, a company would get notice that an audit was going to occur, time to gather requested documentation, and manually reviewed documents to identify any issues or compliance violations.
While Manufacturers could send in documentation to the Government such as reasonable assumptions and methodologies, these documents where seldom responded to and largely believed to not even be reviewed. The Government’s ability to identify red flags was limited, reviews were performed periodically, and audits typically relied on sample sizes instead of comprehensive reviews. Government Agencies were already monitoring submission deadlines, but these checks were binary and did not provide additional context.
AI makes it easier to analyze submissions, policies, and supporting documentation at scale. They can cross-reference larger datasets across programs more efficiently allowing their audits to be more thorough and more often. We’re currently seeing this with CMS and their request that manufacturers submit reasonable submissions on a quarterly basis for designated topics with a character limit, making it easier for AI to analyze and identify outliers.
Get Ready and Stay Ready
AI can analyze Policies, Procedures, Reasonable Assumptions, and Methodology Grids to make sure that policy documentation lines up with how calculations are currently being performed. It can also be used to analyze that documentation against current regulations and identify any potential compliance risks. It unlocks the ability to instantly and continuously identify and mitigate compliance issues before they happen. It also can identify potential red flags where more documentation might be needed or clarified.
A Model for Success
Using AI to analyze data and report trends is an important time-saving measure. However, it becomes just as important to validate the results and have checks in place to ensure that the AI analysis is reviewing data as expected. Generative AI models that review data that’s specifically uploaded to it are usually viewed as more reliable and stable where models that are pulling sources from the internet or third parties could be referencing incorrect or unvalidated information. The analytics are only as good as the prompt. Using specific prompts that are tested and reviewed over time will yield more reliable results and create a stronger compliance program.
Traceable, Explainable, Transparent, and with a Human
Strong compliance documentation ensures that all decisions are traceable to the regulations, explainable where there’s a gray area or outlier, and transparent so the checks and balances are clearly visible. All AI models should follow the same parameters. There should be controls in place to make sure that AI output is being vetted and reviewed by a subject matter expert that can add the nuance and unwritten knowledge AI can lack. Additionally, it is critical to separate where AI is summarizing or analyzing data vs. making decisions.
Stronger controls should be utilized when AI is making decisions to ensure those decisions are compliant, something the company is aware of, and something the company agrees with. AI shouldn’t be used to replace humans, rather it should be a tool to unlock their potential and create operational efficiencies.
Recommended Manufacturer Actions
Manufacturers should consider the following:
- Setting up queries to monitor regulatory changes including new legislation, proposed and final rule-making, and regulatory guidance such as Dear Manufacturer Letters.
- Create a documentation framework that details what is documented, where it is saved, and how AI is being used for any reviews and/or updates.
- Ensure human validation of AI-decisions is happening and documented.
- Set-up continuous monitoring of audit documentation to always be audit-ready
- Utilize AI data analytics to spot data anomalies and potential issues
Conclusion
AI brings the ability to constantly monitor regulatory changes in real time, assess and update SOPs, identify potential compliance risks, analyze Government Pricing calculation results and trends, and much more. The added visibility to potential compliance risks will result in added responsibility to develop and maintain data-driven compliance.
While AI is evolving rapidly and needs to be approached with caution, refusing to embrace AI in your approach to compliance monitoring could result in a Government Agency spotting a compliance issue before your company does. It will also be critical to make sure that as AI comes in, human logic and decision making doesn’t go out. While the processing power and potential of AI is impressive, it also took AI programs about three years to properly identify that there were three Rs in strawberry and not two. Validating and documenting controls around AI-decision making will be crucial.
Reach out to explore how EVERSANA can help understand the potential use cases for AI in your company and ensure the proper controls are put in place.
Author
Megan has more than nine years of experience in the life sciences industry, specializing in Government Pricing (GP). She lead’s EVERSANA’s GP team and has a large range of experience in all areas related…