blue-button-pim

blue-button-pim

NPM

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

###Prerequisites

  • Node.js (v0.10+) and NPM
  • Grunt.js
# you need Node.js and Grunt.js installed

#install dependencies and build
npm install
grunt

Module components

calculateBlockers(data)

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:

var pim = calculateBlockers(data);

var query = model.find({
        $or: [{
            "pim.lnmpdob": pim.lnmpdob
        }, {
            "pim.lnmpfnpc": pim.lnmpfnpc
        }, {
            "pim.lnfn": pim.lnfn
        }, {
            "pim.smdlastname13": pim.smdlastname13
        }]
    });

query.exec(function (err, results) {
        if (err) {
            callback(err);
        } else {
            callback(null, results);
        }
    });

Within the callback we can compare candidates.

compareCandidates(data, candidates, shim)

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.

Contributing

Contributors are welcome. See issues https://github.com/amida-tech/blue-button-pim/issues

Release Notes

See release notes [here] (./RELEASENOTES.md)

License

Licensed under Apache 2.0

Pincer is a project which aims to provide best library discovery tools for developers. We're growing day by day. We have only npm platform for now but we will add the others as much as we can.