e-Statistics > Medical Signal Detection

AERS Database

The present database AERS is composed of the names of medical product (DRUG), the medical terms (REAC) describing the adverse events, and the cell counts (AERS.COUNT) of the frequency table for all the combinations of medical products and adverse events reported via Adverse Event Reporting System (AERS) from January 2004 to March 2005. The dataset DRUG contains all the medical products which appeared at least in 50 individual reports from January 2004 to March 2005 ( DRUG.REPORTED$ \ge 50$ ). The dataset REAC contains all the medical terms which appears at least in 200 reports in the same period ( REAC.REPORTED$ \ge 200$ ).

The cell count (AERS.COUNT) in AERS dataset is denoted by $ N_{ij}$ for the $ i$ -th medical product and the $ j$ -th adverse event. The statistics (random variables) related to cell counts are as follows:

  1. $ N_{i \cdot} = \sum_{j} N_{ij}$ is the count of adverse events (DRUG.COUNT) involving drug $ i$ .
  2. $ N_{\cdot j} = \sum_{i} N_{ij}$ is the count of medical products (REAC.COUNT) involving adverse event $ j$ .
  3. $ N_{\cdot\cdot} = \sum_{i,j} N_{ij}$ is the total count over all the combinations of those appearing in the report involving drug $ i$ 's and adverse event $ j$ 's.
A single AERS report lists typically more than one medical product and more than one medical terms describing the event, contributing the larger total count than the total number of reports.

It displays the part of the selected dataset from to in the count ranking.