With an average drug development timeline of 10 years, stakeholders across research and development (R&D) frequently ask themselves “How can we get medicines to patients more quickly?” While that question has led to innovations in data-driven technologies in clinical trial design and clinical development, ZS’s own market research has found that only a small percentage of pharmaceutical companies make data-driven decisions throughout the regulatory life cycle.
Currently, the regulatory intelligence process is almost entirely manual, but we believe that technology-aided regulatory intelligence tools could greatly improve the drug development process. But to make progress here, pharma companies and regulatory affairs subject matter experts (SMEs) need to be aware of how existing technologies could help them and remain open to change.
The journey from study completion to regulatory approval and market authorization in all the countries a company targets for launch is a long one. But there are numerous ways to speed this up—from selecting the appropriate drug approval pathway to utilizing different accelerated trial design options. For example, the Model-Informed Drug Development (MIDD) Paired Meeting Program can be leveraged for eligible studies. This program informs clinical trial design, aids in the evaluation of critical regulatory review questions and gives early insight into the development of policies. Officials estimate that when applied properly, the MIDD Paired Meeting Program can increase the probability of regulatory success and reduce the drug development timelines. Likewise, external control arms can reduce recruitment needs and accelerate drug development.
Using artificial intelligence (AI) to curate and disseminate regulatory insights
However, pursuing nontraditional regulatory paths poses risks that can hinder adoption of novel methods. Companies often mitigate these by making informed decisions based on data using advanced analytics. But in regulatory affairs, strategic decision-making processes are heavily qualitative and human driven. Current regulatory decision-making processes rely on individual SMEs’ personal experience. While these SMEs are experienced core decision-makers with direct influence on the industry, there’s no doubt they’ll benefit from additional technological support. To optimize strategic regulatory decisions and drive more predictive and proactive insights, SME knowledge could be augmented with AI capabilities that curate and analyze both public and company-specific information in areas such as:
- Regulatory requirements, where AI-enabled systems auto summarize information and send alerts on legislation and health authority (HA) regulations.
- Internal precedence, in which AI-enabled systems guide SMEs by generating insights from data captured in the past through standard operations.
- External intelligence, or insights in the form of competitor experience, market sensing and real-world evidence.
Regulatory affairs SMEs routinely encounter challenges in implementing new regulations. In many countries, pharma companies engage with regulatory authorities when regulations are being framed to help shape them. The process of tracking these interactions and conversations—and the decisions they lead to—is very time consuming and highly manual. But an AI-enabled tool could easily help SMEs compare regulatory requirements across markets, receive regulatory notifications at the right time and evaluate the impact of any changes on regulatory submissions.
If pharma companies are willing to put thought and effort into seeking out technology solutions for these processes, we believe that digitizing these steps could accelerate drug development timelines by 10 months. They can also ensure that companies make the most optimal regulatory strategy decisions and use the most appropriate regulatory pathway. But technology by itself can’t succeed without the processes and people trained to work alongside it. Training, change management and adoption of more efficient processes are all needed to make such technology usable. The future of regulatory intelligence and decision-making is very much a human-plus-machine model.
Must-have capabilities for technology-assisted regulatory intelligence platforms
We believe that for a technology solution to be considered a success, it should both significantly shorten the drug development timeline and enable the following functions. They should:
- Be able to anticipate and pre-empt HA queries. HA queries can relate to study design, selection of endpoints or inconsistent information about data or efficacy. AI-enabled systems can scan HA documents to extract historically asked questions and responses, determine which responses are good and help predict future queries.
- Make better regulatory strategy decisions to increase probability of technical and regulatory registration success. Additionally, they should effectively collect the information senior management needs when deciding which countries to launch a drug in and when. Technology should also indicate where a drug may be eligible for an accelerated regulatory pathway.
When it comes to more technical components for regulatory solutions, they should include the following capabilities:
- Information categorization. Whether it’s historical regulations or the most recently applicable regulation for information search purposes, strong categorization capabilities are vital. Because so much regulatory intelligence data is unstructured, this capability requires strong natural language processing algorithms.
- Semantic search. It’s also essential that the end-user be able to retrieve information easily when needed. For example, users may need to constantly refer to the most updated regulatory guidelines, which require search capabilities like a Google search. Existing capabilities in this domain today include structured search with filters and drop downs. But a more intuitive semantic search capability will make accessing the right regulation much easier and user friendly.
- Knowledge graphs and recommendation engines. Regulatory intelligence must go beyond just storing and retrieving information. It must also be able to alert users about the possibility of better regulatory decisions for a given submission when a submission is being planned. This requires that a system be able to understand the context and background of the submission and make the most appropriate regulatory recommendations. Such context-sensitive recommendations can only be achieved through knowledge graphs.
While there currently are no end-to-end regulatory intelligence platforms on the market, there are tools and technologies available right now that can make a difference. In our experience, many SMEs working in this space aren’t aware of AI-assisted tools that could shorten some of their most time-consuming regulatory tasks. We encourage the professionals working in this space to consider the art of what’s possible and plan for a future where powerful data, combined with the most experienced minds, delivers novel treatments to patients with fewer delays.
Add insights to your inbox
We’ll send you content you’ll want to read – and put to use.