abstraction-tool (work in progress)

React components for executing CQL snippets embedded in a FHIR Questionnaire and displaying resuts.

Components

Abstractor

This library exposes the Abstractor component to execute and render the FHIR Questionnaire with embedded CQL. The component takes the following props:

• patientData: R4.IBundle;
• library: any; (ELM JSON of the library used in the Questionnaire)
• valueSetMap: ValueSetMap;
• questionnaire: R4.IQuestionnaire;

See the types directory for the custom types defined in the library.

import Abstractor from 'abstraction-tool';

function MyComponent() {
    return (
        <Abstractor
            valueSetMap={ /* a valueSet lookup */ }
            questionnaire={ /* a FHIR Questionnaire */ }
            library={ /* ELM JSON */ }
            patientData={ /* a FHIR Bundle of patient data */ }
        />
    )
}

For the rendering of the updated questionnaire form to happen, you will need to include LHC-Forms viewer widget by adding the following to index.html

<script crossorigin="anonymous" src="https://clinicaltables.nlm.nih.gov/lforms-versions/25.1.2/lforms.min.js"></script>
<script crossorigin="anonymous" src="https://clinicaltables.nlm.nih.gov/lforms-versions/25.1.2/fhir/lformsFHIRAll.min.js"></script>   

Loaders

This library also exposes a variety of "loaders" which can be used to assemble the data necessary to provide the props to the Abstractor

• ValuesetLoader: Load FHIR ValueSets into a map provided to the CQL Execution Framework
• QuestionnaireLoader: Query for Questionnaire from FHIR server
• LibraryLoader: Parse ELM from FHIR library

import { QuestionnaireLoader, ValueSetLoader, LibraryLoader } from 'abstraction-tool';

// Questionnaire Loader
const qLoader = new QuestionnaireLoader();
const questionnaire = await qLoader.getFromUrl(/* a url */);

// ValueSet loader
const fhirLibary = /* a FHIR Library resource */;
const contextBundle = /* a FHIR Bundle of FHIR valuesets */;
const vLoader = new ValueSetLoader(fhirLibrary, contextBundle);
const vsMap = await vLoader.seedValueSets();

// Library Loader
const fhirLibary2 = /* a FHIR Library resource with CQL or ELM content */;
const lLoader = new LibraryLoader(fhirLibrary2)
const elm = await lLoader.fetchELM();

Local Development

A simple demo of some cabability of the abstractor can be seen by running it as a standalone application.

Prerequisites

• Node.js
• Yarn

$ yarn install
$ yarn start

Data Sources

With local development of the application, patient data can be provided via a hardcoded FHIR Bundle JSON file, or SMART on FHIR.

Hardcoded Bundle File

In src/config.json, ensure that dataSource.type is 'file'. In src/dataSource/FileDataSource.ts, update the getData function to return the contents of the desired file.

SMART on FHIR

In src/config.json, ensure that dataSource.type is 'smart'. From a SMART app launcher, enter the URL of the application, and upon launch it should prompt for patient selection. Once the OAuth handshake happens properly, the app will load with the selected patient.