SeltaSquare

SeltaSquare identifies industry shifts early and delivers meaningful insights to our clients.

How Global Pharmaceutical Companies Operate Signal Detection

At this conference, a representative from Shinra Zen's pharmacovigilance team presented an efficient Signal Detection approach grounded in real-world operational cases from global pharmaceutical companies, tailored to product characteristics and data volume.

EudraVigilance is the European pharmacovigilance database system operated by the EMA, which collects, manages, and analyzes suspected adverse reaction information for medicines authorized or under clinical investigation within the EEA. Research using EudraVigilance data has found that statistical methods alone detect only about 50% of ADRs. The remaining 50% of signals fall within the "blind spots" of statistical algorithms — areas such as fatal cases, life-threatening adverse events, and Designated Medical Events (DMEs), which require Precision Diagnosis through individual case-level review. This is why statistical methods alone cannot complete a Signal Detection system.

To overcome the limitations of relying on a single method, the presenter proposed an "Integrated Approach" — a flexible combination of three tools:

• Precision Diagnosis: Focused review of individual cases such as fatal cases, DMEs, and IMEs
• Broad Scan: Statistical techniques to capture hidden disproportionality patterns in large-scale data
• Trend Monitoring: Aggregate analysis to understand overall occurrence patterns

Operational Strategy — Designing Around Data Volume and Product Characteristics

The Shinra Zen representative outlined three core operational strategies for practical implementation.

Strategy 1. Method selection by data volume: For small datasets (under 1,000 ICSRs), the focus is on aggregate analysis combined with individual case review. For mid-sized datasets (1,000–10,000 cases), aggregate analysis is paired with manual statistical analysis (e.g., Excel) and individual case review. For large datasets (over 10,000 cases), aggregate analysis is integrated with automated statistical tools (e.g., Empirica) and individual case review. Regardless of data volume, fatal, life-threatening, and DME cases are always reviewed individually.

Strategy 2. A dual-track monitoring cycle: Monitoring is split into weekly rapid review and periodic comprehensive review. Weekly rapid review is a system for promptly addressing serious and urgent adverse events — such as fatal cases and DMEs — through Event Line Listings. Periodic comprehensive review is conducted on a quarterly, semiannual, or annual basis depending on product characteristics, integrating Safety Database and non-SDB information (literature, clinical data, class-related products) into a comprehensive aggregate and statistical review.

Strategy 3. Five variables for setting review cycles based on product characteristics: Time since launch, patient exposure, the product's inherent risk management level, PBRER submission schedule, and special situations (vaccines, campaigns, public attention, etc.) are the five variables considered together to establish the optimal monitoring cycle for each product. Products subject to additional monitoring under EU criteria in particular may require more frequent EudraVigilance monitoring.

The core conclusions of this presentation can be summarized as follows: a "tailored approach" that selects the most appropriate method based on product characteristics, data volume, and risk profile in line with domestic and international regulations; the "synergy of integration," which maximizes sensitivity by combining individual case review with aggregate and statistical analysis rather than relying on a single statistical method; a "dual-track speed" that addresses critical adverse events such as fatal cases and DMEs through weekly rapid review; and "resource efficiency," achieved by applying a quarterly, semiannual, or annual approach to periodic comprehensive review based on product characteristics. Together, these four principles define effective Signal Detection operations.

SeltaSquare Insight: In Signal Detection, Strategy Comes Before Statistics

Both speakers at this conference converged on a single shared message:

"An efficient, tailored pharmacovigilance system ultimately protects patient safety fastest."

Disproportionality analysis methods (PRR, ROR, IC, EBGM, etc.) are useful tools for identifying whether a drug–adverse event combination is being reported more frequently than expected. However, they are not methods for estimating absolute incidence or establishing causality — they are hypothesis-generating tools that produce reporting signals. What matters, therefore, is not which statistical metric is chosen, but the ability to combine methods suited to the data volume and product risk profile, and to interpret statistical signals in context.

Even an identical statistical signal requires different follow-up priorities depending on whether it reflects an expected adverse event already reflected in the label, or a newly observed, serious adverse drug reaction. Likewise, PV activities should not be conducted on the same cycle and in the same manner across all products — a risk-based approach that adjusts the operational framework according to severity and urgency is required, both from a regulatory and practical standpoint.

As access to public databases like FAERS and EudraVigilance becomes easier, disproportionality analysis research is growing rapidly — but so is the risk of over-interpreting results as if they represented actual risk or causal inference. Against this backdrop, the READUS-PV guideline, introduced in 2024, emerged as a reporting standard aimed at improving the transparency, reproducibility, and interpretive accuracy of ICSR-based disproportionality analysis research. In short, systematic and expert interpretive capability is the true competitive edge in Signal Detection.

With SeltaSquare: From ICSR Processing to Signal Strategy

Operating a Signal Detection system in practice requires more than data analysis alone. It demands upfront design decisions — which methodology to select, on what cycle to operate, and how to evaluate detected signals and translate them into regulatory action. For organizations without sufficient dedicated PV staff, or those building or upgrading a Signal Detection system from the ground up, managing this entire process alone is no easy task.

SeltaSquare is more than a PV service provider. From ICSR processing and literature monitoring (LITUS), to combined qualitative-quantitative Signal Detection, RMP design, and DSUR/PBRER authoring, we are an End-to-End Pharmacovigilance Partner supporting the full PV lifecycle under one team.

From tailored Signal Detection design based on data volume and product characteristics, through integrated qualitative-quantitative analysis and clinical interpretation, to regulatory response and label updates — our team of pharmacovigilance, pharmacoepidemiology, and medical writing experts connects every step seamlessly.

Strategically allocating limited PV resources to the safety concerns that matter most, clinically and regulatorily — SeltaSquare partners with you from the design stage onward.