The evolving role of AI in drug submission and approval: An expert perspective
Artificial intelligence (AI) is transforming the pharmaceutical industry, revolutionising drug discovery and approval processes, and reshaping how researchers identify potential drug candidates, streamline clinical trials, and ensure regulatory compliance.
Integrating AI into these processes enhances efficiency, reduces costs, and accelerates new medications reaching the market. As pharmaceutical companies adopt AI technologies, it is essential to explore the opportunities and challenges they present in the quest for innovative healthcare solutions.
Identifying key bottlenecks in the drug approval process
Traditional drug approval processes can be complex and time-consuming, with companies encountering obstacles at every stage, from clinical trials to regulatory reviews.
One area burdened by inefficiencies is documentation. Clinical Study Reports (CSRs), essential for regulatory submissions, require meticulous data integration and adherence to compliance standards - labour intensive tasks that are prone to human error.
In addition, managing documentation revisions and updating team members can be overwhelming. As regulations change, maintaining real-time compliance becomes a hurdle, frequently resulting in costly delays. The preparation of the Common Technical Document (CTD), a mandatory component involving extensive data compilation for regulatory submissions, exemplifies a process that could benefit from AI enhancements.
How AI can address these challenges
The integration of AI into drug submission and approval processes is recognised as a pivotal force within the pharmaceutical industry.1 This evolution brings substantial benefits and notable challenges.
Benefits of AI in drug submission and approval:
- Accelerated drug discovery and development
Traditional drug development timelines can take 12 to 15 years, with costs averaging around £2.5 billion per drug.2 However, companies leveraging AI have reduced development time by 25% to 50%.3 For instance, Insilico Medicine advanced an AI designed drug candidate through discovery and preclinical stages in just 30 months.4
- Improved accuracy and efficiency
By automating data collection from various sources, AI can minimise human error - a crucial factor, given approximately 90% of drugs tested in clinical trials fail to gain approval.5 Furthermore, AI applications can optimise lead compound identification and predict pharmacokinetics and toxicity, facilitating more informed decision-making throughout development.
- Real-time compliance monitoring
AI systems can continuously monitor regulatory changes, ensuring documentation meets the latest standards. This capability is vital, as regulatory landscapes evolve rapidly, significantly reducing the risk of compliance breaches.6
- Template-driven document assembly
Using historical data and predefined templates, AI can automate the creation of detailed CSRs and CTDs, significantly reducing the manual effort involved.
- Streamlined documentation revisions
AI can manage changes to documents while alerting team members to updates in real time. This ensures alignment and minimises the risk of inconsistencies in the document management process.
The challenges of implementing AI
Despite these advantages, integrating AI into established processes presents several challenges:
- Data quality and integration issues
High-quality data is essential for effective AI algorithms. However, obtaining comprehensive datasets from clinical trials can be complex due to standardisation issues, incomplete data, and variability/volume. Additionally, integrating data from various platforms while ensuring privacy and security remains a significant challenge.7
- Transparency and interpretability
The "black box" nature of many AI models complicates transparency in decision-making processes.8 Regulatory bodies require clear explanations for AI generated outputs, which can be difficult when models operate without interpretable logic. This lack of transparency can undermine trust among stakeholders.
- Regulatory frameworks
The absence of standardised protocols for validating AI driven processes complicates technology adoption. Regulatory agencies are still developing guidelines for overseeing AI applications in drug development, leading to uncertainty that may hinder innovation.9
Addressing the challenges of integrating AI into clinical trials is crucial for advancing drug development processes. However, overcoming these hurdles requires not only technological solutions, but also a focus on upskilling the workforce to efficiently leverage AI capabilities.
Upskilling recommendations for the evolving role of AI in drug submission and approval
The complexity of obtaining and integrating high-quality data, ensuring transparency in AI's decision-making processes, and navigating evolving regulatory frameworks demands a well-prepared team. Upskilling initiatives such as role-based simulations and the use of AI-driven tools can equip staff with the necessary skills and knowledge.
1. Implement role-based simulations
- "A Day in the Life" simulations
Create immersive experiences like "A Day in the Life of a Biologist" or "A Day in the Life of a Clinical Researcher." Using virtual or augmented reality allows employees to experience the daily roles and challenges of colleagues, like biologists or clinical researchers. To implement this, a company would define crucial roles, craft accurate scenarios reflecting these roles, and employ immersive technology for a realistic experience. This method promotes understanding and teamwork across departments, enhancing organisational efficiency.
- Scenario-based training
Develop AI-driven simulations that allow teams to engage in real-time scenarios related to drug development and approval. This practical training prepares employees to navigate complex situations and appreciate the impact of their decisions on others.
2. Use AI chatbots for knowledge exchange
- Deploy document intelligence chatbots
Introduce AI chatbots focused on document intelligence to provide quick access to regulatory guidelines, project updates, and best practices. This reduces time spent searching for information, allowing teams to concentrate on their core tasks.
To effectively implement AI chatbots for document intelligence, companies should pinpoint key information needs and select a compatible technology that integrates with existing systems. Chatbots need to be customised and trained with specific data before pilot testing and full deployment, which should include comprehensive staff training. Continuous maintenance and a system for feedback will ensure the chatbot remains effective and meets evolving organisational requirements. Most pharmaceutical companies opt to work in partnership with third party vendors to develop bespoke solutions.10
- Promote continuous learning
Equip chatbots to answer questions about industry news, technology trends, and cases from events. This encourages a culture of continuous learning, ensuring employees remain informed about industry developments.
Benefits of upskilling your teams in AI
There are significant benefits to be realised by companies who train and upskill their workforce to use AI tools effectively including:
- Enhanced collaboration: improved understanding of roles leads to more effective teamwork.
- Increased engagement: immersive training methods encourage active participation and deeper learning.
- Greater efficiency: AI chatbots streamline information access, facilitating quicker decision-making.
By implementing these training tips, pharmaceutical companies can develop a workforce adept at leveraging AI technologies, while promoting better collaboration throughout drug submission and approval processes.
Successful case studies
Several companies have successfully integrated AI into their drug development processes:
- Bristol Myers Squibb used machine learning algorithms to streamline clinical trial designs, resulting in faster patient recruitment and reduced costs.11
- AstraZeneca implemented an AI-driven platform for real-time monitoring of clinical trial data, enhancing compliance tracking and decision-making efficiency.
- Asklepios Biopharmaceutical (AskBio) acquired by Bayer integrated AI into their gene therapy development processes to understand gene regulation better and identify new regulatory sequences within genomes.12
- Several biopharmaceutical companies, such as Bayer, Roche, and Pfizer, have teamed up with IT companies to develop a platform for the discovery of therapies in areas such as immuno-oncology and cardiovascular diseases.13
Reflecting on the future
The integration of AI into drug submission and approval processes signifies a shift towards more efficient pharmaceutical development. The potential benefits are substantial, including faster approvals and quicker access to essential medications while maintaining rigorous standards.
Applying AI in drug submissions and approvals promises a new era of efficiency and precision. By proactively addressing obstacles and using innovative solutions, we can improve drug development for patients and the pharmaceutical industry.
Reflecting on the upskilling tips outlined, implementing role-based simulations, deploying AI chatbots, and promoting continuous learning are essential for equipping your workforce to navigate the evolving landscape of drug submission and approval.
References
- https://www.europeanpharmaceuticalreview.com/news/186633/cphi-ai-to-revolutionise-drug-development-by-2026/
- https://phrma.org/policy-issues/Research-and-Development-Policy-Framework
- https://www.n-ix.com/ai-trends/
- Nature. (2023). Advancements in artificial intelligence for drug discovery: A case study on Insilico Medicine.
- FDA. (2023). Artificial intelligence in drug development: Current trends and future perspectives.
- https://www.certa.ai/blogs/the-future-of-ai-in-compliance-trends-to-watch
- Scribbr. (2021). How to cite a journal article | APA, MLA & Chicago examples.
- _Artificial_Intelligence_and_Transparency_Opening_the_Black_Box
- (Penn State Library Guides, 2023).
- (https://www.mckinsey.com/industries/healthcare/our-insights/generative-ai-in-healthcare-adoption-trends-and-whats-next).
- Statista. (2023). Machine learning applications in pharmaceutical research: A statistical overview
- (https://www.drugdiscoverytrends)
- (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC7577280/)