Academic Commons

Data (Information)

Healthcare Referral Graph Database

Liao, Yumeng

Medical providers in the USA are identified by a sequence of numbers called the NPI, which is unique per provider. This data is publicly available from http://nppes.viva-it.com/NPI_Files.html. Under federal law (FOIA act), providers are required to release data of patient referrals from one provider to another. This data is publicly available from https://questions.cms.gov/faq.php?id=5005&faqId=7977 (Credit to Docgraph). This graph database was built in Neo4j, with data from the NPI registry (May 2014 distribution) as the set of nodes with properties, and with data from physician referrals (2012-2013 distribution) as the set of edges. The original node dataset source (May 2014 distribution): http://nppes.viva-it.com/NPI_Files.html. Data headers/documentation is included with the data. For the database, we chose to focus on only a few node attributes. These include NPI as the index, organization vs individual medical provider, practice city and state, and specialty (represented by taxonomy codes, see codebook). The original edge dataset source (2012-2014 distribution): https://questions.cms.gov/faq.php?id=5005&faqId=7977. The edge data has three columns: source of referral, destination of referral, and quantity of patients referred. Source and destination of referral are represented with NPI.

Files

More About This Work

Academic Units
Computer Science
Published Here
April 13, 2015
Academic Commons provides global access to research and scholarship produced at Columbia University, Barnard College, Teachers College, Union Theological Seminary and Jewish Theological Seminary. Academic Commons is managed by the Columbia University Libraries.