Products and Platforms
TCS ADD™
TCS ADD™: A Suite of Modern and Open Technology Platforms
Highlights
- Traditional pharmacovigilance models struggle to manage rising safety case volumes, diverse data sources, and the demand for greater accuracy and efficiency.
- AI-driven pharmacovigilance platforms automate data ingestion, extraction, and analysis while integrating seamlessly with existing systems to enhance speed and consistency.
- This transformation enables faster case processing, improved regulatory compliance, cost efficiency, and stronger patient safety outcomes.
On this page
PV demand
In today's digital pharmacovigilance (PV) model, there has been a constant rise in user expectations, the need for enterprises to adopt cloud, a rise in safety case volume, constantly evolving adverse events (AEs) communication channels, and network patterns resulting in demand for enhanced case handling and cost-effective solutions. These solutions can provide operational and business simplification and benefits. However, these factors necessitated a holistic approach to new PV architecture, combining traditional PV services and processing with new intelligent solutions to address the challenges of modern PV enterprises.
The next-generation PV platforms, driven by AI, ingest and process data from various sources in different formats, extracting medically relevant information to identify safety patterns and trends. They can also generate an output report without any human intervention.
AI in PV offers many advantages, including enhanced automated data processing capabilities and improved accuracy in extracting AEs information. Traditional PV models rely on manual processing and analysis, which is time-consuming and prone to human error. These procedures can be automated by AI, resulting in more consistent and effective outcomes. By using new digital technologies, we can process and analyse large datasets from different sources, including electronic health records, social media reports, spontaneous cases for marketed products, and safety reports from clinical trials. ML algorithms can identify safety signals, patterns, and correlations that initially might not be immediately apparent to safety experts. This capability has a particular benefit in post-marketing surveillance, due to the high volume of data.
Moreover, AI can enhance the accuracy of AEs detection. Natural language processing (NLP) algorithms are now able to process unstructured information from safety reports, like patient narratives and published medical literature, and identify potential AEs. This approach complements human intelligence with scientific reasoning to improve accuracy, quality, and efficiency. AI has become an indispensable tool in the early detection of safety signals, opening for the first time the possibility of early and proactive interventions and minimising the impact of potential harms on patients.
We will elaborate on the integration and opportunities of AI in PV for the life sciences industry and present a brief comparison of marketed products.
Market analysis
During the last 10 years, the PV market has experienced the impact of the advent and development of AI platforms aimed at improving safety monitoring and AE reporting of medical products. Several commercial products have emerged, each offering unique capabilities and benefits. When comparing AI platforms available in the market, several key features stand out across different products. Here's a general comparison and analysis of these platforms:
- Designed to enhance safety monitoring by leveraging advanced analytics and ML
- Focus on user-friendly interfaces and integration capabilities with other systems to make these platforms popular choices among the pharmaceutical industry
- Enhance data analysis and reduce the burden on PV teams
- Scalability and flexibility, making them suitable for all sizes of organisations
- Case management capabilities are a common strength. They enable efficient processing of AEs reports and assist in signal detection and risk assessment
- Valuable insights for decision-making and enhancing adherence to regulatory requirements
- Ability to integrate with other systems and provide real-time analytics
Challenges of AI-based platforms
Despite advantages, the integration of AI in PV is not without challenges.
- Ethical considerations: The potential for AI to make decisions that affect patient safety raises important ethical questions. It is essential to establish clear guidelines and regulatory frameworks to ensure that AI is used responsibly and ethically.
- Quality and reliability of the data used to train AI models: The use of erroneous or biased data when developing the model can generate incorrect conclusions and potentially wrong outcomes. It is crucial that AI models are trained on high-quality, representative safety datasets.
- Transparency and interpretability of AI algorithms: This is essential to gain the trust of the industry and regulatory authorities.
AI models' data quality extraction: AI models sometimes behave differently and provide surprises during the extraction of information. Extraction is not always correct; it could extract additional information, miss critical information, or extract incorrect information.
Key factors behind the extraction issues of AI models:
- Exposure to a lack of data
- Hallucination
- Lack of examples
- Prompts lack context
- Implementation challenges: Navigating the challenges from a successfully executed pilot concept into a full-scale implementation platform is critical. This transition must address the complexity and high volume of cases in real-life PV operations. A strong implementation model requires expertise in both the domain and technology from an appropriate team.
- Change management: The scaling, delivery, deployment, and adoption of an AI-driven platform should follow standard protocols and procedures. Building consensus and cross-functional consensus within the organisation and with technology vendors is key to effective collaboration among business, IT, PV, and technology teams.
Accelerating innovation
We believe that a cloud-native, extensible, and RESTful architecture, with standard interfaces for interoperability is essential for a next-generation platform. These platforms should include:
- AI, GenAI-driven processing: Along with relevant human workflows for case processors and medical review teams.
- Automated ingestion and categorisation: Any case file format and source category to address case template and data attribute variations efficiently.
- System evaluation: A system-enabled process where the engine output is reconciled with the closed cases to report the accuracy of processing and extraction. The system should use these evaluation outputs.
- Transparency and audit trail: All actions and operations performed by the system and users should be traceable, with the provision of a rational report against every step taken in the case processing.
- Implementation of conventions and business rules: LLMs can apply business rules on case data without any need for explicit programming. This functionality reduces implementation requirements and simplifies the overall conventions and product business rule management.
- Translation of non-English source document to English: Convert the source document to English for case processing with the use of NLP and LLMs to understand the context of non-English language for data extraction.
- Automated duplicate check: Duplicate check is a critical aspect when processing safety cases. Traditionally, this step has been performed either manually by the PV team or using a rule-based approach. With the use of ML algorithms, the similarity scores can be predicted with a certain degree of accuracy.
- Increased efficiency in quality review: It is essential for the quality reviewer to know which of the attributes are extracted with a high degree of accuracy and which need to be reviewed. ML algorithms can predict the accuracy of extraction for each attribute. The end-to-end productivity of the case, from intake to reporting, can be improved by enhancing both case quality and processing efficiency.
- Individual case safety report (ICSR) auto-narrative generation: LLMs are trained in the creation of such narratives based on case data. Hence, the entire narrative generation activity can be automated without requiring any training or the use of rule-based algorithms.
AI-powered PV
AI platforms for PV offer unique strengths and capabilities. They excel in;
- Data integration and comprehensive analytics focus on efficiency and compliance, providing scalability and flexibility, and are recognized for their robust case management and regulatory compliance features.
- Choosing the right AI-driven PV platform depends on the organisation's needs and priorities, as well as the proven successful experience of the technology provider.
- The adoption of an AI-driven platform to enhance productivity, facilitate product safety analysis, contribute to business growth, support budget management, and drive innovation represents a critical decision for an organisation. This transformation decision demands a shift in mindset from both PV staff and cross-functional teams. These shifts have significant organisational impacts on different teams, which require effective change management.
In conclusion, AI has the potential to revolutionise the traditional PV model by enhancing data processing capabilities, improving the accuracy of AE data extraction, and enabling the analysis of large datasets. However, it is essential to address the challenges related to quality, transparency, and ethical considerations to realise the benefits of case processing automation fully. As AI continues to evolve, it will be crucial to strike a balance between leveraging its capabilities and ensuring patient safety.
The current commercial products in the PV market are leveraging AI and automation to enhance safety monitoring, improve efficiency, and ensure regulatory compliance. AI-driven products are leading the way in providing innovative solutions that address the challenges of traditional PV methods. As the market continues to evolve, AI-driven PV products will play a crucial role in ensuring the safety and efficacy of medical products.