Infrastructure for Community of Practice (CoP) Interoperability Project

Overview

This project implements a Loss to Follow Up (LTFU) workflow system for CHIS based on OpenHIE LTFU Guide.

A first version of the project can be found in the chis-interoperability repository.

Components

The components and reference information for interoperability used in this project are:

OpenHIE defines the architecture for an interoperability layer.
OpenHIM is a middleware component designed to ease interoperability between systems.
HL7 FHIR is a messaging format to allow all systems to understand the format of the message.

CHT

The structure of documents in the CHT database reflect the configuration of the system, and therefore do not map directly to a FHIR message format. To achieve interoperability we used a middleware to convert the CHT data structure into a standardized form so the other systems can read it. Below is the standard data workflow:

This project uses OpenHIM as the middleware component with Mediators to do the conversion. Outbound Push is configured to make a request to the middleware when relevant documents are created or modified in the CHT. A Mediator then creates a FHIR resource which is then routed to OpenHIM. OpenHIM routes the resource to any other configured systems.

Conversely, to bring data into the CHT, OpenHIM is configured to route the updated resource to the Mediator, which then calls the relevant CHT APIs to update the document in the CHT database. This will then be replicated to users’ devices as per usual.

See more information on the CHT interoperability page on the CHT documentation site.

Services

Services are currently available at these URLs:

OpenHIM Admin Console - https://interoperability.dev.medicmobile.org.
OpenHIM Mediator - https://interoperability.dev.medicmobile.org:5001/mediator.
CHT with LTFU configuration - https://interop-cht-test.dev.medicmobile.org/.

Credentials to the instances can be shared upon request.

GitHub repository for the kubernetes configuration.

Workflow Sequence Diagram

FHIR Resources

The following FHIR Resources are used to implement the flow above:

Test LTFU Workflow

The workflow guide explains the steps for testing the LTFU workflow using the local setup and the Medic instances. For the local setup, you are expected to have already completed the setup procedures outlined in this document in the section below.

Get Started with Local Setup

Prerequisites

docker
Postman or similar tool for API testing. This will play the role of the Requesting System from the sequence diagram above.

Troubleshooting

Users getting errors when running the following installation steps, please see the Troubleshooting guide.

Install & First Time Run

Run ./startup.sh init to start-up the docker containers on the first run or after calling ./startup.sh destroy. Use ./startup.sh up for subsequent runs after calling init without calling destroy.

OpenHIM Admin Console

Visit the OpenHIM Admin Console at http://localhost:9000 and login with the following credentials: email - interop@openhim.org and password - interop-password. The default User username for OpenHIM is interop@openhim.org and password is interop-password. The default Client username is interop-client and password is interop-password.
Once logged in, visit http://localhost:9000/#!/mediators and select the only mediator with the Name 'Loss to Follow Up Mediator'.
Select the green + button to the right of the default channel to add the mediator.
You can test the mediator by running:

curl -X GET http://localhost:5001/mediator -H "Authorization: Basic $(echo -n interop-client:interop-password | base64)"

You should get as a response:

{ "status": "success" }

If everything is successful you should see this:

CHT configuration with Docker

The following steps apply when running CHT via the Docker setup provided in this repository:

CHT can be accessed via http://localhost:5988, and the credentials are admin/password.
Create a new user in the CHT instance with the username interop-client using these instructions. For the role you can select Data entry and Analytics roles. Please note that you can use any username you prefer but you would have to update the config with the new username. You can do that by editing the cht-config/app_settings.json file and updating the username value in the outbound object e.g. on this line.
Securely save the interop-client user's password to the database using the instructions here. Change the values mykey and my pass to openhim1 and your user's password respectively. An example of the curls request is below:

curl -X PUT -H "Content-Type: text/plain" http://admin:password@localhost:5988/api/v1/credentials/openhim1 -d 'interop-password'

Local setup of CHT Configuration

The following steps apply when running CHT locally in development mode and when making configuration changes locally:

CHT Development Environment

Set up a local CHT instance using these instructions.
Create a new user in the CHT instance with the username interop-client using these instructions. For the role you can select Data entry and Analytics roles. Please note that you can use any username you prefer but you would have to update the config with the new username. You can do that by editing the cht-config/app_settings.json file and updating the username value in the outbound object e.g. on this line.
Securely save the interop-client user's password to the database using the instructions here. Change the values mykey and my pass to openhim1 and your user's password respectively. An example of the curls request is below:

curl -X PUT -H "Content-Type: text/plain" http://admin:password@localhost:5988/api/v1/credentials/openhim1 -d 'interop-password'

After updating the mediator code or cht configuration, you need to run ./startup.sh up-dev to upload the changes to docker compose.

CHT Configuration

Go into the cht-config directory by running cd cht-config.
Run npm install to install the dependencies.
Create a file named .env under /mediator folder, copy over the contents of /mediator/.env.template and update the CHT_USERNAME and CHT_PASSWORD values with the admin credentials of your CHT instance.
Set up a proxy to your local CHT instance by running using something like nginx-local-ip or ngrok and update the CHT_URL value in the .env file with the new URL.
Ensure you have cht-conf installed and run cht --local to compile and upload the app settings configuration to your local CHT instance.
To verify if the configuration is loaded correctly is to create a Patient and to access a URL like https://*****.my.local-ip.co/#/contacts/patientUUID/report/interop_follow_up. This should retrieve correctly the follow up form.
To verify if the configuration in CouchDB, access http://localhost:5984/_utils/#database/medic/settings.

Shutdown the servers

To shut-down the containers run ./startup.sh down to stop the instances.
To then restart the containers, run ./startup.sh up. You do not need to run init again like you did in the initial install above.
To shut-down and delete everything, run ./startup.sh destroy. You will have to subsequently run ./startup.sh init if you wish to start the containers.

medic/cht-interoperability