Background and Challenges
The ISO IDMP (Identification of Medicinal Products) standard is an international standard established in response to the need for the unambiguous identification of medicinal products, primarily to improve information exchange and support pharmacovigilance activities. The general objectives of the standard are:
- Improve the exchange of information among regulatory authorities, companies, and within the same company
- Enhance drug safety monitoring
- Improving data integrity and reliability
- Streamline regulatory activities
- Support mechanisms for verifying the authenticity of medicines
The standard includes a data model with over 200 fields, designed to collect all product-related data.
To date, Europe is leading the way in adopting the standard: the EMA has established an ad hoc working group and defined the SPOR programme (Substance, Product, Organisation, Referential) to guide its implementation.
The standard’s adoption approach has evolved over the years, from a ‘big bang’ implementation (despite different “iterations” were envisioned) to a gradual one: recently EMA adopted an agile approach. Split into many different opportunities having “the IDMP language” in common (e.g. eAF, ePI, RPM, ESMP, etc.). This approach enables flexibility in reviewing priorities and deadlines, but also presents a significant challenge for stakeholders who must constantly keep up with the inherent changes.
What are the difficulties for Companies?
A major difficulty for companies lies in effectively managing their data, which involves defining the data source, retrieving it, normalizing it in accordance with implementation guidelines, and establishing robust processes for its management and maintenance.
Moreover, data, even when available internally, frequently resides in unstructured sources, often within documents such as the SmPC (Summary of Product Characteristics). As highlighted in the Implementation Guides, a portion of the required IDMP data can be extracted from the SmPC. Although the SmPC has a structured layout with mandatory sections, the information contained within these sections is inherently unstructured.
Finally, the various EMA systems where Companies IDMP data are reported, are often “out of sync” if the various IDMP data submissions through the various EMA services are not governed appropriately according to an effective end-to-end IDMP data management process.
Our Approach
Arithmos approach is to manage all IDMP data within the company, not relying on EMA IDMP & XEVMPD services only.
In this context, Arithmos leveraged its experience in the sector to analyze the challenges faced by
companies in collecting all IDMP data across the organization, specifically aiming to address the
difficulty of extracting data from documents. According to our benchmark, this happens in 50%-60% of cases and the (human) effort to be paid is huge.
We focused particularly on the SmPC, as it is the document containing the most IDMP relevant data, and developed a BOT that, leveraging AI logic, is capable of:
- Reading the document
- Identifying sections with the necessary information
- Extracting the data required by the IDMP standard
- Applying data normalization, such as mapping to controlled vocabulary codes (RMS list) and formatting dates as required
Importantly, AI is not the objective in itself—Arithmos’ innovation lies in using AI as a trusted accelerator to strengthen compliant, efficient, and sustainable regulatory processes.
Success Factors and Outcome
The SmPC BOT was used in a specific project to support data entry resources, with the goal of extracting the maximum possible amount of IDMP data from the company’s provided SmPCs, aligned to IDMP specification (Implementation Guidelines v2.1)
Key results emerging from the use of automation included:
- A reduction in the time needed to retrieve normalized data against the RMS list, particularly evident for substances and therapeutic indications, saving approximately 40 minutes per SmPC on average
- Improved data quality through the correction of manual retrieval errors and typos from SmPCs
- Facilitation of identifying correct values, even when the BOT provided only partial outputs
- Adherence to the project timeline for SmPC data extraction
Technical specifications
In the pilot implementation, the BOT successfully extracted up to 93 IDMP-relevant fields per SmPC document. These fields included both static (constant) values and dynamic content requiring contextual interpretation or correlation to predefined value lists (i.e. controlled vocabularies), such as full name, dosage forms, therapeutic indications, ingredients, strength, administration routes, and marketing authorization data.
Each extracted value was tagged with a confidence level (High, Medium, Low) linked to the BOT’s output. This reliability scoring mechanism supported downstream data curation efforts and enabled targeted manual reviews only where necessary and allowed to evaluate BOT’s actual performance.
Results showed that over 70% of the fields were extracted with high confidence and required minimal human correction—demonstrating the BOT’s strong baseline accuracy and its ability to accelerate data extraction.
Conclusion
This case study highlights Arithmos’ ability to strategically combine human expertise with automation in a way that elevates — not replaces — the quality of regulatory processes.
Arithmos is committed to continuously evolving the SmPC BOT — not to chase technology trends, but to responsibly support our clients in navigating the complexity of regulatory transformation.
About the Author
Educated as a Biomedical Engineer at Politecnico di Milano, Ilaria’s career spans in the Healthcare and Life Sciences industries, focusing on computerized systems, digitalization, and the successful delivery of complex projects.
She has extensive experience in managing complex technological projects, translating client needs into concrete digital solutions within regulated environments.
As a Senior Consultant, her expertise includes defining digital strategies, leading implementation projects for standards like IDMP to ensure compliance and optimize processes, and exploring the application of Artificial Intelligence.
Her background also includes a solid understanding of healthcare IT systems and processes, gained from his previous roles.
About Arithmos
With deep expertise in the Life Science Industry, Arithmos supports companies in their digital transformation journey to achieve the best value from technology-enabled solutions and excellence in business operations.
Arithmos team share extensive knowledge of the full GxP regulated environment for both deliveries of services and technologies, and our domains range from Clinical R&D, Quality Assurance, Drug Safety & Pharmacovigilance, Regulatory and Medical Affairs.
