DCLApi

Setup

Install dependencies:

• SQLite

• Node.js & NPM

Install required node modules:

cd $PROJECT_DIRECTORY
npm install

Run tests:

npm test

Note that in order to run the application you will need to acquire the database. See the section on data below. Once acquired, the database must be called dcl.sqlite and placed at the project root. This is configured in config.js.

Once you have the database, run the application:

npm start

Examples

classbyatc:

$ curl http://localhost:3007/classbyatc/C03CA01
{"ATC":"C03CA01","class":"furosemide"}

$ curl http://localhost:3007/classbyatc/C03CA
{"ATC":"C03CA","class":"Sulfonamides, plain"}

$ curl http://localhost:3007/classbyatc/C03C
{"ATC":"C03C","class":"HIGH-CEILING DIURETICS"}

$ curl http://localhost:3007/classbyatc/C03
{"ATC":"C03","class":"DIURETICS"}

$ curl http://localhost:3007/classbyatc/C
{"ATC":"C","class":"CARDIOVASCULAR SYSTEM"}

classbydin:

$ curl http://localhost:3007/classbydin/02155907
{"DIN":"02155907","atcLevel":2,"class":"CALCIUM CHANNEL BLOCKERS"}

$ curl http://localhost:3007/classbydin/02155907?atcLevel=1
{"DIN":"02155907","atcLevel":"1","class":"CARDIOVASCULAR SYSTEM"}

$ curl http://localhost:3007/classbydin/02155907?atcLevel=3
{"DIN":"02155907","atcLevel":"3","class":"SELECTIVE CALCIUM CHANNEL BLOCKERS WITH MAINLY VASCULAR EFFECTS"}

$ curl http://localhost:3007/classbydin/02155907?atcLevel=4
{"DIN":"02155907","atcLevel":"4","class":"Dihydropyridine derivatives"}

$ curl http://localhost:3007/classbydin/02155907?atcLevel=5
{"DIN":"02155907","atcLevel":"5","class":"nifedipine"}

atcbydin:

$ curl http://localhost:3007/atcbydin/02155907
{"DIN":"02155907","ATC":"C08CA05"}

Data

The data is not included in this repository, but the scripts used to acquire it are.

Acquiring DIN to ATC Data

The Drugref database is used. The MySQL database can be dumped using the scripts/drugrefdump.sh script. The dump file can be converted to SQLite using the scripts/mysqldump2sqlite.sh script.

Acquiring ATC Description Data

The original data comes from the file rrf/RXNCONSO.RRF in http://download.nlm.nih.gov/umls/kss/rxnorm/RxNorm_full_02022015.zip at http://www.nlm.nih.gov/research/umls/rxnorm/docs/rxnormfiles.html. You need to register to access that zip file which is quite large, 217MB expanding to 1.7GB, but mostly consists of data we don't need.

The scripts/concepts_export_all.sql SQL script can be used to export the necessary data from the RxNorm file rrf/RXNCONSO.RRF. This generates a '|' separated file which can be loaded into an SQLite database using the scripts/loadAtcData.js script. To keep all of the data in one database, you should load it into the database created from the drugref dump above.

Spec

Drug API Spec

Technology — ideally a Node JS component that provides ...

REST API

############################################################################################################

— should be doable with public data sources

/classbydin/

takes an 8­ character din code — pads beginning with zeros to 8 characters — returns JSON

{“DIN”:”provided DIN code padded”, “class”:”string name of the class of the drug identified by the DIN code”}

############################################################################################################

—RXNav provides this service already

/classbyatc/

takes an 3+ character ATC code and returns JSON

{“ATC”:”provided ATC code”,”class”:”string name of the class of the drug identified by the first three characters of the provided atc code”}

http://rxnav.nlm.nih.gov/REST/rxclass/class/byId.json?classId=B01AA

############################################################################################################

— this is likely impossible without a subscription, but having the framework in place to do this is highly desirable

/classbyfddb/

takes an FDDB code and returns JSON

{“FDDB”:”provided FDDB code ­ padded if apropriate”, class:“string name of the class of the drug identified by the FDDB code”}

############################################################################################################