As an acutely ambitious and overstretched Pharma industry increasingly identifies the vast potential of Generative AI to transform medical writing, across a range of use cases linked to regulatory documentation and safety report summaries, several companies are attempting to build capabilities in-house – and are coming unstuck. With reference to a new survey conducted with the Regulatory Affairs Professionals Society, Punya Abbhi, Chief Operating Officer and Co-Founder of Celegence, highlights a critical gap in capability as companies look for effective shortcuts to their most burdensome regulatory content needs.
Generative AI (GenAI) is seen as a panacea for transforming the way companies work with vast quantities of knowledge and content – sifting it and summarising it, and deriving new insights from the data it contains. And the technology certainly offers that potential, as long as it is harnessed appropriately and with the right controls. And this is where some life sciences companies are potentially opening themselves up to new risk, as they rush to embrace GenAI capabilities to make lighter work of their substantial medical writing requirements.
The Pharma industry is an ideal candidate for GenAI-based process transformation, because of its highly-regulated and painstakingly-templated repeat activities including licence application and maintenance and safety report writing. Even if GenAI-based ‘bots’ could distil the right content and craft a first draft of these typically hefty documents, for attentive checking and amendment by skilled humans, this would considerably reduce the strain on overstretched teams of Regulatory, Quality, Safety and Clinical professionals.
The trouble is that GenAI technology needs a lot of moulding and crafting to be of reliable use, even with the benefit of large language models (vast data sets) to build system ‘knowledge’. However advanced and intuitive the natural language processing and deep learning capabilities, AI models will still need to be guided in what to look for, how to repurpose information and data correctly, and what ‘good’ looks like. This requires a unique combination of AI proficiency and industry insider knowledge (hands-on experience).
Effective Shortcuts Start with Effective Instruments
In the context of Pharmaceutical medical writing, AI specialists, already in high demand in the war on talent1, need to be proficient not just in the latest combinations and applications of natural language processing and deep learning, but also need access to experts in the life science industry to understand specialist language and vocabularies, required templates, and nuanced demands of each market. They need an intricate appreciation for what will be accepted by regulatory agencies, and how that differs from region to region, and country to country, with a strong feel for specific medical writing best practices linked to each use case.
That’s because to maximise any return on investment and be trusted as a credible generator of first-round content, GenAI must be better and faster than humans at the initial collation and interpretation of what’s critical and required in technical documentation.
It also takes vast reams of successful example content to ‘teach’ a GenAI model what’s needed, and the ideal output to aim for. There needs to be a robust understanding of formal terminology (and its abbreviations and variants); of how to decipher tables, listings and figures; and of meaningful correlations between data and a drug or treatment. Only in this honed, specialist context can AI tools be trusted to interpret complex data and point to a logical outcome.
For Pharma companies to bring their own capabilities up to speed, and stay ahead, seems an impossible feat.
The Inhibitor to Advancement: Pharma’s Capability Gap
A new survey conducted with the Regulatory Affairs Professionals Society (RAPS)2 found that, in Pharma, medical writing remains a critical area for support. The main driver of this is the pressure on regulatory professionals’ time.
The appetite to intelligently automate some facet of the medical writing process is significant and growing stronger. Some 57 percent of surveyed companies specified planned investment in technology to improve medical writing over the year ahead, almost on a par with eCTD v4.0 spending (specified by 58 percent for the coming 12 months), the two categories dominating immediate Regulatory IT spending plans.
The primary medical writing needs identified by Regulatory professionals, in terms of requiring additional support, were clinical study protocol/report writing, and drafting of regulatory documents.
In the survey, more than half of respondents in the 2024 survey identified a need to harness AI in data extraction (56 percent) and information summarisation (53 percent), where just 9-10 percent are using AI for those purposes today – specifically within the context of medical writing. Twelve percent said they were actively in the process of incorporating AI into automated report generation from multiple sources, which is the ultimate opportunity on offer.
