ips-ig-cl

features

IPS

Chile localization

FHIR IG

International Patient Summary

Narrative generation

languages

TypeScript

Implementation of IPS FHIR IG on Aidbox FHIR platform

This repository contains pre-configured Aidbox instance and implementation of $summary operation defined by IPS.

About IPS (International Patient Summary)

The IPS is intended to support the provision of essential healthcare information for a patient, regardless of where they are receiving care. It includes critical information such as allergies, medications, past surgeries, and other significant medical history details. See IPS Specification

Prerequisites

Docker

STEP 1: Environment and Aidbox license

Copy .env.tpl file into .env file:

cp .env.tpl .env

If you are hosting Aidbox on your local computer, obtain the self-hosted license as described in the documentation.

Add the license (AIDBOX_LICENSE) int the .env file.

STEP 2: Run aidbox and node-app in Docker

docker compose up --build

Step 3: Open and log in into Aidbox instance

Open in browser http://localhost:8888

And log in witn username: admin and password: password

Step 4: Request $summary using REST Console

In the Aidbox admin window, navigate to the APIs section and choose REST Console.

The $summary operation requires either the logical ID (Patient.id) or a business identifier (Patient.identifier) of the patient. You can use the following request to view all available patients:

GET /fhir/Patient?_elements=id,identifier

To request the IPS "document" Bundle for a specific patient using the REST Console, you can use the following request:

GET /fhir/Patient/[id]/$summary

Replace [id] with the logical ID of the patient you want to retrieve the IPS document for.

For example:

GET /fhir/Patient/2b90dd2b-2dab-4c75-9bb9-a355e07401e8/$summary

Or you can use request with identifier search parameter:

GET /fhir/Patient/$summary?identifier=<patient-identifier>

For example:

GET /fhir/Patient/$summary?identifier=574687583

Step 5: Request $summary using HTTP Client

If you're starting the Aidbox FHIR server for the first time, the initial step involves creating a Client resource with an ID and secret. Since the newly created client does not have default permissions to access the Aidbox REST API, the next step is to configure access policies. Refer to the documentation: Create and test access control

The easiest way to achieve this is by navigating to Auth > Sandbox in the Aidbox Web Admin UI and performing REST queries to create the 'basic' client and assign the AccessPolicy.

Once the client is created, you can perform the $summary operation using an HTTP tool, similar to the following example:

curl --location 'http://localhost:8888/fhir/Patient?_elements=id%2Cidentifier' \
--header 'Content-Type: application/json' \
--header 'Authorization: Basic YmFzaWM6c2VjcmV0' \
--data ''

Deploy

This application works with aidbox running under the following settings (envs):

AIDBOX_FHIR_VERSION=4.0.1
AIDBOX_FHIR_SCHEMA_VALIDATION=true
AIDBOX_FHIR_PACKAGES=hl7.fhir.r4.core#4.0.1:hl7.fhir.uv.ips#1.1.0
AIDBOX_VALIDATE_BINDING_URL="https://r4.ontoserver.csiro.au/fhir/ValueSet/\$validate-code"

It is possible to deploy Aidbox with Helm charts. To do so, follow the steps outlined below:

1. Add aidbox helm repo

helm repo add aidbox https://aidbox.github.io/helm-charts

2. Prepare database config

config: |-
  listen_addresses = '*'
  shared_buffers = '2GB'
  max_wal_size = '4GB'
  pg_stat_statements.max = 500
  pg_stat_statements.save = false
  pg_stat_statements.track = top
  pg_stat_statements.track_utility = true
  shared_preload_libraries = 'pg_stat_statements'
  track_io_timing = on
  wal_level = logical
  wal_log_hints = on
  archive_command = 'wal-g wal-push %p'
  restore_command = 'wal-g wal-fetch %f %p'

env:
  PGDATA: /data/pg
  POSTGRES_DB: postgres
  POSTGRES_PASSWORD: <your-postgres-password>

image.repository: healthsamurai/aidboxdb
image.tag: "16.1"
storage:
  size: "10Gi"
  className: <your-storage-className>

and apply it

helm upgrade --install aidboxdb aidbox/aidboxdb \
  --namespace ips --create-namespace \
  --values /path/to/db-config.yaml

3. Prepare Aidbox config

host: <your-aidbox-host>
protocol: https

config:
  PGHOST: aidboxdb.ips.svc.cluster.local
  PGDATABASE: postgres
  PGUSER: postgres
  PGPASSWORD: <your-postgres-password>
  AIDBOX_CLIENT_ID: <your-aidbox-client-id>
  AIDBOX_CLIENT_SECRET: <your-aidbox-client-password>
  AIDBOX_ADMIN_ID: <your-aidbox-admin-id>
  AIDBOX_ADMIN_PASSWORD: <your-aidbox-admin-password>
  AIDBOX_LICENSE: <aidbox-license>
  AIDBOX_DEV_MODE: true
  AIDBOX_FHIR_VERSION: 4.0.1
  AIDBOX_FHIR_SCHEMA_VALIDATION: true
  AIDBOX_FHIR_PACKAGES: hl7.fhir.r4.core#4.0.1:hl7.fhir.uv.ips#1.1.0
  AIDBOX_VALIDATE_BINDING_URL: "https://r4.ontoserver.csiro.au/fhir/ValueSet/$validate-code"
  AIDBOX_BASE_URL: <your-base-url>
  AIDBOX_PORT: 8888
  AIDBOX_COMPLIANCE: enabled

ingress:
  annotations:
    acme.cert-manager.io/http01-ingress-class: nginx
    cert-manager.io/cluster-issuer: letsencrypt
    kubernetes.io/ingress.class: nginx

and apply it

helm upgrade --install aidbox aidbox/aidbox \
  --namespace ips --create-namespace \
  --values /path/to/aidbox-config.yaml

To deploy the application use prepared k8s config. Additionally, you have to add to the config resource Secret:

apiVersion: v1
kind: Secret
metadata:
  name: ips-app
  namespace: ips
stringData:
  AIDBOX_CLIENT_ID: <your_aidbox_client_id>
  AIDBOX_CLIENT_SECRET: <your_aidbox_client_secret>
  APP_SECRET: <your_app_secret> # allows aidbox safely communicate with this app

Indexes

$summary issues GET /fhir/<Type>?patient=Patient/<id> for 13 clinical resource types. To make these fast, the bundled docker-compose.yaml runs Aidbox with BOX_FHIR_SEARCH_TOKEN_OPERATOR=#>> so that ?patient= compiles to (resource #>> '{...,id}') = '<id>', which can be served by a small per-type btree.

Create the indexes (one statement per POST /$sql — CONCURRENTLY cannot run inside a transaction):

-- SP `patient` resolves to .subject (10 types)
CREATE INDEX CONCURRENTLY IF NOT EXISTS observation_subject_id_idx         ON observation         ((resource #>> '{subject,id}'));
CREATE INDEX CONCURRENTLY IF NOT EXISTS condition_subject_id_idx           ON condition           ((resource #>> '{subject,id}'));
CREATE INDEX CONCURRENTLY IF NOT EXISTS procedure_subject_id_idx           ON procedure           ((resource #>> '{subject,id}'));
CREATE INDEX CONCURRENTLY IF NOT EXISTS diagnosticreport_subject_id_idx    ON diagnosticreport    ((resource #>> '{subject,id}'));
CREATE INDEX CONCURRENTLY IF NOT EXISTS medicationrequest_subject_id_idx   ON medicationrequest   ((resource #>> '{subject,id}'));
CREATE INDEX CONCURRENTLY IF NOT EXISTS medicationstatement_subject_id_idx ON medicationstatement ((resource #>> '{subject,id}'));
CREATE INDEX CONCURRENTLY IF NOT EXISTS careplan_subject_id_idx            ON careplan            ((resource #>> '{subject,id}'));
CREATE INDEX CONCURRENTLY IF NOT EXISTS deviceusestatement_subject_id_idx  ON deviceusestatement  ((resource #>> '{subject,id}'));
CREATE INDEX CONCURRENTLY IF NOT EXISTS flag_subject_id_idx                ON flag                ((resource #>> '{subject,id}'));
CREATE INDEX CONCURRENTLY IF NOT EXISTS clinicalimpression_subject_id_idx  ON clinicalimpression  ((resource #>> '{subject,id}'));

-- SP `patient` resolves to .patient (3 types)
CREATE INDEX CONCURRENTLY IF NOT EXISTS allergyintolerance_patient_id_idx  ON allergyintolerance  ((resource #>> '{patient,id}'));
CREATE INDEX CONCURRENTLY IF NOT EXISTS immunization_patient_id_idx        ON immunization        ((resource #>> '{patient,id}'));
CREATE INDEX CONCURRENTLY IF NOT EXISTS consent_patient_id_idx             ON consent             ((resource #>> '{patient,id}'));

After building the indexes (and after any bulk load), refresh statistics so the planner picks them up:

ANALYZE observation;
ANALYZE condition;
ANALYZE procedure;
ANALYZE diagnosticreport;
ANALYZE medicationrequest;
ANALYZE medicationstatement;
ANALYZE careplan;
ANALYZE deviceusestatement;
ANALYZE flag;
ANALYZE clinicalimpression;
ANALYZE allergyintolerance;
ANALYZE immunization;
ANALYZE consent;

Name		Name	Last commit message	Last commit date
parent directory ..
src		src
.dockerfileignore		.dockerfileignore
.env.tpl		.env.tpl
Dockerfile		Dockerfile
Dockerfile.dev		Dockerfile.dev
README.md		README.md
docker-compose.yaml		docker-compose.yaml
k8s.yaml		k8s.yaml
nodemon.json		nodemon.json
package.json		package.json
pnpm-lock.yaml		pnpm-lock.yaml
tsconfig.json		tsconfig.json

Navigation Menu

Repository navigation

ips-ig-cl

ips-ig-cl

README.md

Implementation of IPS FHIR IG on Aidbox FHIR platform

About IPS (International Patient Summary)

Prerequisites

STEP 1: Environment and Aidbox license

STEP 2: Run aidbox and node-app in Docker

Step 3: Open and log in into Aidbox instance

Step 4: Request $summary using REST Console

Step 5: Request $summary using HTTP Client

Deploy

1. Add aidbox helm repo

2. Prepare database config

and apply it

3. Prepare Aidbox config

and apply it

Indexes

FilesExpand file tree

ips-ig-cl

Directory actions

More options

Directory actions

More options

Latest commit

History

ips-ig-cl

Folders and files

parent directory

README.md

Implementation of IPS FHIR IG on Aidbox FHIR platform

About IPS (International Patient Summary)

Prerequisites

STEP 1: Environment and Aidbox license

STEP 2: Run aidbox and node-app in Docker

Step 3: Open and log in into Aidbox instance

Step 4: Request $summary using REST Console

Step 5: Request $summary using HTTP Client

Deploy

1. Add aidbox helm repo

2. Prepare database config

and apply it

3. Prepare Aidbox config

and apply it

Indexes