For too long, regulatory teams have become bogged down by protracted reviews of the regulatory implications of each new product change, with potential risk to drug availability until these have been assessed and actions set in motion. Now Generative AI (GenAI) promises to change all that, say independent consultant Preeya Beczek and Agnes Cwienczek of ArisGlobal.
ogichobanov Shutterstock
Critical in ensuring ongoing product quality, safety and efficacy, assessing the regulatory impact of any change to a medicinal product or its manufacture is also essential to maintain compliance with regulations and manage risk. This applies to every market and regulatory jurisdiction.
The process is typically complex, manually painstaking and a growing headache as the life science industry strives to balance quality and safety controls with the drive to improve drug accessibility and speed to market. Harnessing AI, and specifically generative AI (GenAI), to ease the pressure is the obvious answer and one that companies themselves have identified1.
AI’s potential is already established in a pharma regulatory context, where the technology is making inroads into processes around marketing authorisation application preparation and global authorisation management across the product lifecycle. This is easing the pressure on overwrought Regulatory departments. It makes sense to extend the benefits to regulatory impact assessment, then.
Without efficient assessment of the regulatory implications, product supply could stall
A pre-cursor to any process involving a product change, reg impact assessment needs to happen swiftly and efficiently, yet the nature of the activity prevents it.
Ideally, if an urgent safety change arises, assessments of the regulatory ramifications typically should be completed within hours (in reality, they tend to take several days). Questions to be answered at speed include: “What was submitted to the authority previously?”; “What is on the labelling and how is that affected?”, “what is the reg procedure to follow and the implementation window”, and “What documentation will be needed?”
The upshot, traditionally, has been a lot of time expended on scouring and referencing of multiple disparate sources, including information contained in PDFs. The inevitable reach-out to in-country affiliates for local requirements and information (and any translations) can further add to the delay in completing the assessment and agreeing next steps.
Each country will need to determine whether and when they will need to make a change in their own market (e.g. amend product labelling), and notify the relevant health authority – and in what logical order. (Should they act first and then report, or the other way around?) And again, what are the time implications: how immediately will updates need to be made?
There is a lot to consider. A good proportion of this activity will have to happen in parallel, too, to enable ongoing demand planning, supply planning, materials availability, etc.
A growing appetite for AI in this context
Given all of the variables, applying AI technology to product change control/regulatory impact assessment might once have been deemed a stretch. Yet, as capabilities have advanced, the fit has become more pronounced. It is no coincidence that 55% of senior regulatory professionals have actively expressed interest in an advanced, AI-enabled technology support for reg impact assessment2. They see GenAI as ideally positioned to help lighten an unsustainable load and accelerate delivery.
But what does this look like in practice, and how can organisations prepare the ground? Current recommendations are to assess the individual stages of regulatory impact assessment for where intelligent automation could help most. Once specific opportunities have been identified, the next consideration will be the company’s current data and technology “readiness”. This will determine what is likely to be possible in the immediate or short term and where the biggest benefits will come.
Identifying one or two product lines, or a specific region or country, for initial deployment of AI may be an optimal way to go, rather than taking a more holistic or scattergun approach. This will make it possible to contain any unforeseen issues, and apply any learnings to an expanded implementation later.
Prompting faster and more coordinated decision-making
Once AI proves its mettle, in taking the strain of the detailed exploratory work, process stakeholders (central regulatory professionals, local regulatory representatives, as well as those in manufacturing level, demand, supply chain, Quality and Safety) can focus more of their attention to agreeing appropriate next actions.
Even simply accelerating the process of locating and searching all of the information, and determining where efforts need to be concentrated will empower teams to move more swiftly in determining and executing next moves. The ability to automatically scan the latest regulatory intelligence in different markets, and consult previous Agency exchanges, can then help further expedite next steps – or at least pinpoint where supplementary insights may be needed where the latest local requirements are less clear.
Extending the benefits laterally
The more embedded AI becomes in the end-to-end product change control/regulatory impact assessment process, the greater the benefits are likely to be. Where the technology is harnessed to pull information from multiple sources into one place, teams can be ready to review and validate the findings. GenAI tools can help with structured content authoring, meanwhile, or swiftly bring a document from version zero to a solid first draft, knowing what data to pull in - and where to find it.
Governance will be critical in all of this, as regulatory teams turn increasingly to AI to take over the administrative heavy lifting. Irrespective of how much trust there is in the AI capabilities, teams should not expect to rely on the technology to make the decisions on their behalf. Cross-functional teams will still need to agree whether and where a change is applicable, when it will need to be made and when it should be reported to the relevant regulatory agency, for example.
Last but certainly not last, pharma companies must have a solid plan to bring their people along on the journey. Strong communication of the vision, and appropriate skills reinvestment, will be needed. This is so that functional teams appreciate exactly how to make most efficient use of data and documents, as well as the need for meticulous care and attention when engaging AI-enabled automation a part of regulatory processes.