PV247 offers a variety of tools to support signal detection ranging from:
- Incidence reports by product, clinical sign and time period
- Statistical reports showing proportional reporting ratios and Chi-squares
- PV-Analyzer (optional), a comprehensive statistics module that adds Bayesian statistics, graphical capability, user-defined alerts, advanced data grouping and many more functions
These can be incorporated into your pharmacovigilance process to conduct a regular review of safety profiles as required by regulatory authorities.
Volume 9 states that market authorization holders have a responsibility to
ensure that all sources of information are screened regularly to identify potential signals and that appropriate action is taken in response to new evidence
It is recommended to fully document this process, even if the review detects no changes. Signal detection is likely to be a critical issue during pharmacovigilance inspection.
Report-based Signal Detection
All PV247 users can access views that display cases that include specific clinical signs. While this enables to answer specific questions it is not a basis for efficient safety surveillance.
PV247 includes three safety monitoring reports:
- AE report rate by product
- Clinical signs – multiple periods
- PRR statistics
The first two provide incidence counts while the third converts these counts into a proportionality.
Both incidence count reports are similar – one only considers a single time period (based on date received) while the other can compare across up to four date ranges. The date ranges are entered by the user at runtime.
The PRR statistics report calculates the Proportional Reporting Ratio (PRR) for all products and signs in a time period defined by the user. PRR is a widely used and well understood frequency measure of disproportionality. It is the proportion of cases for a given product that are linked to a specific adverse outcome, divided by the corresponding proportion for all other products. PRR is one of the key methods used by EMA in safety monitoring.
PV-Analyser Data Mining and Signal Detection
PV-Analyser is Ennov’s pharmacovigilance data mining and signal detection tool for Human and Veterinary case databases (coming from PV247, or imported from authorities or third party solutions).
It includes a flexible pivot table / cube that can summarize data within categories defined by the user. PV-Analyser offers a complete statistics module based on disproportionality methods, both with and without Bayesian inference.
PV-Analyser is ideal for companies with an extensive product portfolio or ones that wish to conduct comprehensive quantitative safety surveillance of their data.
Statistical methods include:
- Proportional reporting ratio (PRR)
- Reporting odds ratio (ROR)
- Observed to expected ration (OE)
- Chi-squared significance
- Information content (IC)
- Empirical Bayesian geometric mean (EBGM)
- Multi gamma poisson shrinker (MGPS)
Features of PV-Analyser include
- Analysis by product, ingredient and batch number
- Analysis at different MedDRA / VedDRA levels
- User defined alert thresholds with notification when these are exceeded