Evaluation of technical approaches for real-time data transfer from electronic health record systems
Author: Nikola Kirilov
In this repository you will find reference implementations to evaluate the technical approaches for real-time data transfer from electronic health record systems
Warning: The provided code is not intended for production use.
To use rest-hooks and websocket make sure to activate them. For HAPI FHIR the following lines in the "application.yaml" have to be uncommented and set to true:
subscription:
resthook_enabled: true
websocket_enabled: true
To use the script you need to install the requests library:
pip install requests
By default the script does not use any authentication for the connection. If your server needs authentication we provide HTTP Basic Auth support. To use it you need to change line 5-7 of the script:
auth = 1 #Authentication needed (only basic auth supported)
fhir_user = "user"
fhir_password = "password"The url of the FHIR server and the endpoint for rest hooks is given in the variables line 8-9
fhir_url= "http://localhost:5000/fhir/Subscription/"
endpoint = "http://localhost:5001/" #For rest-hooks, we will use the rtd recipient http endpointFor endpoint you can use the http endpoint in "rtd_recipient".
The script is executed from terminal with the following two arguments: [PID] - Id of the Patient resource for which we will generate notifications when an Observation resource is created [CHANNEL] - channel type of the subscription (websocket or rest-hook)
py subscriby.py [PID] [CHANNEL]
After executing the script the Subscription resource is printed as a response in the terminal. It is important to note the Subscription resource id if you will use websockets.
To install flask use the following command:
pip install flask
The script can be started passing an argument - port number on which the endpoint will be accepting connections:
py http_endpoint.py [port number]
When RTD is received the timestamp on receipt is recorded to a csv file (receipt_[port number].csv).
pip install websocket-client
The FHIR WebSocket endpoint is set in line 34:
ws = websocket.WebSocketApp("ws://localhost:8080/websocket", ...
To run the script use:
py ws_client.py [SUBID]
Where [SUBID] is the id of the Subscription resource you generated in the previous step.
When RTD is received the timestamp on receipt is recorded to a csv file (receipt_subscription[SUBID].csv)
To generate Observation resources for the patient use the python script "generate_observations.py" in the "fhir_observation" folder. The settings and dependencies are the same as the Subscription generating script:
py generate_observations.py [PID]
where [PID] is the id of the Patient resource.
The script generates 1000 resources and records the timestamps to a csv file (generation_timestamps.csv).
Please make sure that the fhir_url is correct and where the generated observations should be sent: to the proxy or directly to the FHIR server.
Dependencies:
pip install psycopg2
The server host, port, credentials and database are set in line 15-19 in the script:
connection = psycopg2.connect(user="postgres", password="admin", host="127.0.0.1", port="5432", database="fhir")
Once the script is run it is listening for notifications.
The script takes the filename for the timestamps as an argument.
py generate_observation [filename]
When RTD is received the timestamp on receipt is recorded to a csv file ([filename].csv)
Install Golang (https://go.dev/doc/install).
To set up you FHIR server address you need to change line 72:
rev_proxy, err := NewProxy("http://localhost:8080")
The address of the reverse proxy is set in line 77:
log.Fatal(http.ListenAndServe(":5000", nil))
The address of the RTD recipient http endpoint is defined in line 56:
go f("http://localhost:5001/",jsonData)
You can use the python "http_endpoint.py" in "rtd_recipient".