Imprecision in vaccine adverse event reporting and a methodological analysis of reporting systems to improve pharmacovigilance and public health.

Abstract

This study documents imprecision in Japanese reports of adverse events following immunization (AEFI). In doing so, it presents methods to analyze this imprecision. These methods include use of unique Japanese data on the validity of certain AEFIs. They also include ways to estimate AEFI rates, which allow comparison of AEFI data between countries. Using US AEFI data for comparison, we show how differences in AEFI reporting systems likely influence AEFI statistics. Although our comparisons of AEFI rates are not precise, many of the difference we detected between Japanese and US statistics make sense and reflect differences in the societal and medical perspectives on various vaccines or can be explained by differences in the reporting systems including reporting sources. For example, differences in societal and medical perspective probably underly the extraordinarily high Japanese rates of anaphylaxis and other AEs following HPV immunizations from 2010 to 2016 compared to US rates and to Japanese rates for other vaccines. High US rates of reported Guillain-Barré syndrome following influenza vaccination relative to Japanese rates and to rates for other US vaccines are consistent with data suggesting that the index of suspicion for such reactions could affect AEFI rates. The findings that over half of Japanese anaphylaxis reports for every vaccine are erroneous, and that close to half of "serious" Japanese AEFI cases probably are not serious may be due in part not only to explanations unique to Japan, but also to factors that apply to the USA and other countries. Differences in reporting systems account for a much higher rate of non-serious AEFI reports in the USA compared to Japan. Japanese marketing authorization holders are probably at least as assiduous and timely in their reporting of AEFIs as health care providers, though granular level differences are apparent in reporting by various sources. The methods we used to analyze the validity of Japanese statistics can be used to analyze the validity of AEFI reports from other countries and aid the harmonization of adverse event reporting systems. Eventually, similar reporting systems might be adapted for drugs and medical devices.