Patient Identification and Matching based on Blue Button data
Blocking phase - filter all patient population using blocker filter
Candidates comparison phase - compare candidates with patient data one by one
Scoring phase - Weight two records for potential match
Match evaluation phase - Evaluate score result
PIM relies on Data Matching Algorithm use by the Oklahoma Department of Mental Health and Substance Abuse Services (ODHMSAS) which in turn was influenced by the article by Matthew A. Jaro published in the Statistics in Medicine Journal, Vol. 14, 491-198 (1995) titled ["Probabilistic Linkage of Large Public Health Data Files."] (http://www3.interscience.wiley.com/journal/114131327/abstract)
##Quick up and running quide
Node.js (v0.10+) and NPM
# you need Node.js and Grunt.js installed#install dependencies and build
This function returns an object with the blocking traits for a given patient's demographic information.
The purpose of this object is for targeted filtering in MongoDB queries. For exampe, we could use the following
code to return a set of candidates for comparison:
This function will take demographic data and a list of candidates and return a list of matches and flagged candidates.
Each object in the result array has a pat_key and a flag match. An ideal implementation of this module would result in
only one automatic result at a time. In the case that there is a manual result, a user should be presented with
all results flagged as manual. An optional shim can be passed for translating demographic data within a database
to match the schema detailed in blocker.js.