Questions & Answers of webinar (Feb 2015) on Signal detection methodology: Current and Novel approaches

This Questions & Answers Blog are from Signal Detection Webinar conducted in February 2015.

1) Which level of grouping or selection for AE term is used?
Usually the Lowest level term (LLT) that precisely reflects the reporter’s verbatim should be selected. However, the MedDRA coder should also check the hierarchy of the terms selected (i.e., PT, HLT, HLGT and SOC) as it is also important to select the LLT falling under correct hierarchy which otherwise would lead to erroneous coding.
Also, it is equally important to ensure good coding practices by establishing company procedures and having a quality check so as to ensure correct information as is reported goes in the case. In Sarjen’s PvNET, when an LLT term is selected, the system automatically enters the related PT term in the data field.

2) Is Sarjen’s PvNet having MedDRA plugin? How is the MedDRA updation mechanism built in Sarjen’s PvNet ?
Yes Sarjen’s PvNET provides the MedDRA upload/import option. MedDRA is upgraded twice in a year and hence, is loaded in Sarjen’s PvNET application accordingly. Sarjen’s PvNET also provides MedDRA upgrade impact analysis in order to identify the differences between versions. On importing the latest MedDRA version, it is possible to identify and re-code the cases with the exact LLT terms as reported by the ADR source in previous versions.

3) Do you have any example where single case is identified as signal?
There are many potential safety risks discussed in the published recommendations and minutes of meeting of PRAC (Pharmacovigilance and Risk Assessment) Committee wherein one can look for single cases that may have been categorized as Signal.
Similarly, on the FDA website, there exists a web page of Potential Signals of Serious Risks/New Safety Information identified from the FDA Adverse Event Reporting System (FAERS) that lists out the product name, potential signal and the additional information (Exemplary link:

However, there is no mention of the number of cases reviewed (whether single case or case series) for the product and potential signal.
In the previous webinar of January 2015 on Signal detection, we have discussed a single case that was identified as a potential risk/new safety information. The potential signal was of lichenoid dermatitis (drug-induced reaction) and the suspect drug was alendronic acid. One more example of a potential safety risk was discussed and can be viewed on this link-

4) Is PRR statistics calculation compared with the data available in any of the standard AER (Adverse event reporting) dictionary?*
As per the principle for PRR (Proportional reporting ratio) calculation, it infers statistical signal hypothesis based on comparison of observed frequency of reports for a drug/event combination to the expected frequency of reports for that AE in entire adverse event report database
So, usually one cannot have a standard AER dictionary for PRR calculation, but it is based on the number and nature of cases processed in database as well as type of medicinal products present in the current MAH database.
In real practice, there probably does not exist any ‘standard’ AER dictionary for such a statistical calculation. With the presence of drug(s) in market, the adverse event profile is going to be affected constantly (as number of adverse events are reported and recorded in database). It is subject to constant change with number of cases being added in the MAH database.
However, stratification of the data can be meaningfully carried out before PRR run to have specific values related to sub-population of interest. Age-wise, gender-wise, therapeutic class, MedDRA coding (LLT, PT, SMQs) etc. different stratifications can be considered for statistical data mining.
* NOTE: Question reframed for better interpretation
For regular Pharmacovigilance updates, connect with us on LinkedIn
Contact Us for more details

Book a Demo