Analyzing FDA Datasets with the ELK Stack (Kibana, Machine Learning, and Graph)

Michael Heldebrant

Distinguished Solutions Architect

Elastic

In this video, we show you how to analyze two unique FDA datasets, MAUDE and FAERS, using the Elastic Stack (formerly known as the ELK Stack). MAUDE contains medical device adverse event reports submitted by mandatory reporters (manufacturers, importers, and device user facilities) and voluntary reporters (health care professionals, patients, and consumers). The FDA Adverse Event Reporting System (FAERS) contains information on adverse event and medication error reports submitted to the FDA.

Using Kibana, Graph, and the Elastic Stack’s machine learning features, we show you what you can do with both datasets. You’ll be able to find interesting words in unstructured data and guess the occupation of the person who wrote them before we reveal the answer. Using machine learning, you’ll be able to detect potential patient lawsuits hiding in the data based on unusual spikes in reports.

Highlights:

• Graph can help you find uncommonly common words — terms that appear more frequently than the background rate would suggest — in a dataset.
• Machine learning looks at thousands of data points simultaneously and spots anomalies across millions of pieces of information that have affected real people.
• Once you’ve identified an information trail to chase, you can do a full analysis.

Additional Resources:

• openFDA dataset pages
• MAUDE - Manufacturer and User Device Experience Dataset Download
• FAERS- FDA Adverse Event Reporting System
• FAERS Dataset Download