A systematic step-by-step approach to vaccine signal detection

Background/Introduction: Pharmacovigilance Centers receives large numbers of reports of Adverse Events Following Immunisation (AEFIs) after childhood vaccination; mostly concerning reports of well-known AEFIs. Usually, methods for signal detection are based on either case by case analysis or disproportionality analysis. Analysing large numbers of reports in routine signal screening, carries the risk that clinical and pharmacological knowledge is not taken into account. This may limit the capability to detect potential signals. Signal detection can be optimised by taking detailed information from the reported AEFIs as well as background information of the National Vaccination Program (NIP), like the numbers of vaccinees, administered vaccines, batch numbers and changes in the NIP over time, into account. Objective/Aim: To illustrate the use of a step by step approach for vaccine signal detection, in which detailed information from the reported AEFIs and background information of the NIP is taken into account. Methods: A 6-step approach for vaccine signal detection in frequently reported well-known AEFIs in a vaccination program. 1. Collecting information: about background incidence of AEFI, their clinical characteristics, changes in NIP, media attention. 2. Selection of dataset: Select reports of the vaccination program of interest, for example The National Immunisation Program Children and defined the period of interest. 3. Selection of cases: Define the cases, for example all reports of AEFIs after administration Diphtheria, Tetanus, acellulair Pertussis, Inactivated poliovirus (DTPa-IPV) of 4 year olds in the period of interest. 4. Information of exposed: Collect information of the number of vaccinees, per administered vaccine (brand name) per year. When needed, the batch number can be taken into account. 5. Define co-variables for stratification: e.g., latency, age, gender, injection site, vaccination moment, vaccine brand name. 6. Select analyses method: Description of cases, analysis of reporting patterns/clustering, disproportionality analysis, trend-analyses, batchanalyses, evaluation of cases against the background of changes in the NIP. Results: By using the described 6-step approach we detected three new signals in well-known AEFIs. Analysis of reports of injection site abscesses at infant age showed clustering of cases of site abscesses around the 3rd or 4th injection moment at the injection site of DTPa-HBV-IPV/ Hib. In an overview of AEFIs after administration of DTPa-IPV at the age of 4 an increase was detected in the number of reports in 2015 and a decrease in 2017. The 2015 increase may have been related to a change in priming at infant age. The decrease in 2017 was linked to the replacement of different brands DTPa-IPV vaccines used. Conclusion: A systematic step-by-step approach and taking detailed information and background information into account is of added value in vaccine signal detection