4CPS-194 The underreporting rate as a pharmacovigilance process indicator in a comprehensive cancer centre

Background and importanceUnderreporting in oncology practice is a known phenomenon linked to the predictable toxicity of these drugs. Reporting indicators are often calculated on very large catchment areas and this limits the capacity for self-assessment of the performances in each hospital.Aim and objectivesThe purpose of the study was to evaluate the feasibility of the underreporting rate index in a single cancer centre as a process indicator.Material and methodsReports of adverse drug reactions (ADRs), from 1 January 2018 to 31 January 2019 in our institute (230), were collected in a database. The nine most active ingredients reported were subjected to an indepth analysis and their reporting rates were calculated using the formula (number of drug reports X/number of patients treated with drug X)×100 and (number of drug reports X/number of drug administrations X)×100. The expected value was evaluated using the formula (expected frequency×number of patients treated with the drug X)/100, where the expected frequency was calculated by the summary of product characteristics. The rate of underreporting was calculated as a ratio (missing episodes/expected episodes)×100. Only the results of paclitaxel (the most reported drug) are reported in this abstract as an example.ResultsIn the period January 2018 to January 2019, paclitaxel related ADRs were 51 in 412 patients treated (3293 total administrations). The reporting rate for the number of patients treated was 12.4% while the reporting rate by number of administrations was 1.5%. Severe neutropenia represented the main toxicity with an expected incidence of 39%, while the reported incidence was 6.41%.The underreporting rate of ADRs related to paclitaxel were: neutropenia 82.17%, febrile neutropenia 97.22%, transaminite 94.49%, thrombocytopenia 98.46% and diarrhoea 91.02%. Some gastrointestinal and musculoskeletal system reactions were very common reactions (≥1/10) but there were no reports.Conclusion and relevanceThe indicator allowed better identification of the area of underreporting over time in a more precise way than the absolute number of reports. It is feasible, but when the expected frequency of the event decreases to below 10%, the indicator loses reliability for samples