Explore and analyze FDA Drug Adverse Event (FAERS) data
FDA Drug Adverse Event Overview
FDA Adverse Event Reporting System (FAERS) Quarterly Data Extract Files
Component | Details |
---|---|
Question | Are different adverse events reported in different countries? |
Sources (FAERS) | REAC19Q3.txt, DEMO19Q3.txt -> join on PRIMARYID |
Data | for each case, record DEMO.OCCR_COUNTRY, REAC.PT pairs |
Analyses | frequency of each event-type in each country (profile) |
. | frequency profile comparisons across countries |
. | distance matrices |
. | classification/clustering |
- 2019Q3 data is only an incremental subset (<10%) of the entire FAERS dataset
2-Way Heatmap of normalized event frequencies (countries in columns):
Zoomed-in on left side with distinct sets of events for top reporting countries:
Histogram of counts of frequency values of each event in individual countries (excludes zero counts) - a frequency cutoff of 0.9 seems reasonable.
"Distinct" drug event terms above frequency cutoff here:
Possible other analyses:
Analysis | Methods |
---|---|
Clustering | K-means, biclustering, non-negative matrix factorization |