FHIR-Based Bulk Patient Data Export

Bahmni provides the ability to export bulk patient data in FHIR NDJSON format. This feature has options for exporting data as-is, or with anonymised data (default setting is anonymised export). To be able to use this feature a user must have requisite privileges to export data in bulk. See details below of sample formats, anonymisation options and privileges required.

1 Introduction
2 Privileges Required
3 Preconfigured Concepts
4 Screenshot of Admin Export Screen
5 Sample Export Formats
- 5.1 With Anonymisation
- 5.2 Without Anonymisation
6 Anonymisation Configuration

Introduction

The FHIR-Based Bulk Patient Data Export feature in Bahmni facilitates efficient, standardized export of large patient datasets while ensuring interoperability using the FHIR standard. It incorporates robust anonymization measures to protect sensitive data, striking a crucial balance between data accessibility and privacy compliance. This functionality empowers users to responsibly share comprehensive datasets for analytics and reporting, upholding the highest standards of patient confidentiality and regulatory requirements.

Anonymisation in Bahmni encompasses transforming or redacting sensitive information to ensure privacy protection. Organisations can conduct meaningful analytics without compromising individual privacy. Bahmni offers a privileged FHIR data export option for data sharing across systems and analytics. For instance, Snolytical leverages this option for patient data analytics.

Handle patient data confidentially! It's advised to export data in anonymised mode, ensuring careful handling by authorised parties only. Consent responsibility lies with the system user in production.

Video demo: https://www.youtube.com/watch?v=MOFFWx7_mp8

The resources available for FHIR export include:

Patient
Condition(includes Diagnosis in Bahmni)
MedicationRequest
ServiceRequest (Procedure Order)

Privileges Required

To activate the export option, two privileges are required:

Export Patient Data: Allows anonymised export, as per the anonymisation configuration.
Export Non-Anonymised Patient Data: Permits plain export without anonymisation (use carefully!).

Preconfigured Concepts

The following concepts are required for presenting export details as tabular information in the FHIR task. These concepts are predefined and are available out of the box in the snomed-default-config and snomed-clinic-config.

Download URL
FHIR Export User Name
FHIR Export Start Date
FHIR Export End Date
FHIR Export Anonymise Flag

Screenshot of Admin Export Screen

In the Admin section under FHIR Export, the patient data export feature is accessible. Only users/admins with export privileges will be able to export the data. Any facility can decide which user / role must have that privilege, allowing them to initiate exports accordingly.

Once the privileges are configured, and anonymisation settings are in place, the FHIR export is ready for use.

You can download data in NDJSON format using the 'Download' link.

Sample Export Formats

With Anonymisation

Below are examples of anonymised resources resulting from FHIR export :

{
  "resourceType": "Patient",
  "id": "6aa3fa07f7c9f4a54a1caa84689aa81c157c88a58a5931a27e1a4685be403568",
  "meta": {
    "lastUpdated": "2023-11-15T08:11:44.000+00:00"
  },
  "active": true,
  "gender": "male",
  "birthDate": "1989-01-01",
  "deceasedBoolean": false
}

{
  "resourceType": "Condition",
  "id": "113d8cd5-130b-4689-a278-dd111b2b8d3c",
  "meta": {
    "lastUpdated": "2023-11-15T08:13:11.000+00:00"
  },
  "clinicalStatus": {
    "coding": [
      {
        "system": "http://terminology.hl7.org/CodeSystem/condition-clinical",
        "code": "active"
      }
    ]
  },
  "category": [
    {
      "coding": [
        {
          "system": "http://terminology.hl7.org/CodeSystem/condition-category",
          "code": "encounter-diagnosis",
          "display": "Encounter Diagnosis"
        }
      ]
    }
  ],
  "code": {
    "coding": [
      {
        "code": "e96cc22f-2e5b-4cb4-806e-719dcd4f6943",
        "display": "Mild asthma (disorder)"
      },
      {
        "system": "http://snomed.info/sct",
        "code": "370218001"
      }
    ],
    "text": "Mild asthma (disorder)"
  },
  "subject": {
    "reference": "Patient/6aa3fa07f7c9f4a54a1caa84689aa81c157c88a58a5931a27e1a4685be403568"
  },
  "onsetDateTime": "2023-11-15T08:13:11+00:00"
}

{
  "resourceType": "MedicationRequest",
  "id": "53d231dc-47b3-45e2-82e3-d74c6b3d9164",
  "meta": {
    "lastUpdated": "2023-11-15T08:13:12.000+00:00"
  },
  "status": "active",
  "intent": "order",
  "medicationCodeableConcept": {
    "coding": [
      {
        "system": "http://snomed.info/sct",
        "code": "318353009",
        "display": "Propranolol hydrochloride 40 mg oral tablet"
      },
      {
        "system": "https://fhir.openmrs.org",
        "code": "8e817314-a24f-4890-8a50-e24d6304de08",
        "display": "Propranolol hydrochloride 40 mg oral tablet"
      }
    ],
    "text": "Propranolol hydrochloride 40 mg oral tablet"
  },
  "subject": {
    "reference": "Patient/6aa3fa07f7c9f4a54a1caa84689aa81c157c88a58a5931a27e1a4685be403568",
    "type": "Patient"
  },
  "dispenseRequest": {
    "validityPeriod": {
      "start": "2023-11-15T08:13:12+00:00"
    },
    "numberOfRepeatsAllowed": 0,
    "quantity": {
      "value": 3,
      "unit": "Tablet(s)",
      "code": "86239663-7b04-4563-b877-d7efc4fe6c46"
    }
  }
}

Without Anonymisation

Below are examples of non-anonymised resources resulting from FHIR export :

Anonymisation Configuration

FHIR export anonymization options include:

Redact: Omit specified sensitive information during export.
Correlate: Use surrogate keys instead of sensitive data for tracking and correlation. Surrogate keys are generated through a one-way hash function, configured with 'oneWayHashSalt' to create unique identifiers.
Add Noise: Randomise data for privacy, e.g., "firstOfMonth" transforms dates to the beginning of the year.
Fixed: Mask data to a predetermined fixed value.

These methods empower users to customise anonymisation strategies for data export, ensuring privacy while enabling valuable analytics.

For instance, if we aim to anonymise the resources as follows:

Patient Attributes	Anticipated Export Results

Patient Attributes	Anticipated Export Results
resourceType	No change
id	Surrogate key for correlation
identifier	Redact all identifiers
active	No change
name	Redacted
telecom	Redacted
gender	No change
birthDate	Set to 1st Jan of the birth year
deceasedBoolean	No change
deceasedDateTime	Set to 1st Jan of the death year
address	Redacted

Condition Attributes	Anticipated Export Results

Condition Attributes	Anticipated Export Results
resourceType	No change
clinicalStatus	No change
category	No change
code	No change
subject	Surrogate key for correlating with patient resource
encounter	Redacted
onsetDateTime	No change
recordedDate	Redacted
recorder	Redacted

MedicationRequest Attributes	Anticipated Export Results

MedicationRequest Attributes	Anticipated Export Results
resourceType	No change
status	No change
intent	No change
priority	Redacted
medication	No change
subject	Surrogate key for correlating with patient resource
encounter	Redacted
authoredOn	Redacted
requester	Redacted
dosageInstruction	Redacted
dispenseRequest.validityPeriod	No change
dispenseRequest.numberOfRepeatsAllowed	No change
dispenseRequest.quantity	No change

Procedure Attributes	Anticipated Export Results

Procedure Attributes	Anticipated Export Results
resourceType	No change
status	No change
category	No change
code	No change
subject	Surrogate key for correlating with patient resource
encounter	Redacted

Here is the corresponding configuration that anonymises sensitive fields