Development of cleaning viewpoints and evaluation of anonymity among Japanese Adverse Drug Event Report database

This paper evaluates and analyzes the data cleaning method and anonymity of the Japanese Adverse Drug Event Report database (JADER). JADER consists of multiple tables and more than three million reports of adverse events. Since there are several errors included in this database, it is hard to analyze its anonymity. There are more than two thousand type of drug names which are unique in JADER. We have cleaned the data one by one to develop data cleaning method and found error mode and cleaning procedure by utilizing some online database and general search engine, google. As a result, we found rules for data correction of drug names and there are still two hundred and forty-two records shown k=1 for its anonymity.