Quantitative Smoke Status Extraction

This package aims to extract discrete smoking information from clinical narrative text.

Smoking information type

• pack per day (PPD)
• pack-year (PY)
• smoking year (SY)
• quit year – e.g., quit for 10 years (QY)
• year at quit – e.g., quit at 2008 (YQ)

Input format

Individual text files:

A directory containing your individual clinical notes in .txt format. Each file will produce one output file in .ann format.

CSV file:

A .csv file with no header expected to contain the following information:

• patient ID
• note ID
• note date
• note type
• note text

The output file will be a .csv of the same format, with added fields for each category. If no value for a given category was obtained, the corresponding cell will be left blank.

Running the package:

Depending on your file input strategy, please run either of the following commands:

For individual text files

python run_engine.py directory/containing/notes/

For single .csv file

python run_engine.py path/to/your/file.csv