Health Information Systems: Interoperability Standards

Healthcare data is generated by a vast array of systems — electronic health records (EHRs), laboratory information systems, pharmacy systems, imaging systems, and patient portals. For these systems to exchange data meaningfully, they must agree on common standards for message formats, data structures, and communication protocols. This article covers the three most important healthcare interoperability standards: HL7 v2, FHIR, and DICOM.

HL7 v2 — The Workhorse of Healthcare

HL7 version 2 is the most widely deployed healthcare messaging standard in the world. Developed in 1989, it defines a pipe-delimited text format for exchanging messages between healthcare systems. Despite its age, HL7 v2 remains dominant because it is simple, flexible, and well-understood by implementers. An HL7 v2 message consists of segments (each starting with a three-letter code like MSH for Message Header, PID for Patient Identification, and OBR for Observation Request), with fields separated by the pipe character (|) and sub-fields by the caret (^).

# HL7 v2 ADT (Admit/Discharge/Transfer) message example
MSH|^~\&|SENDING_APP|SENDING_FAC|RECV_APP|RECV_FAC|202607081430||ADT^A01|MSG001|P|2.5
EVN|A01|202607081430|||
PID|1||12345^^^MRN^MR||Doe^John^^||19700115|M|||123 Main St^^NYC^NY^10001||555-1234|||S
PV1|1|I|WARD^A^101^^^FAC||||ATTENDING^SMITH^J^^^DR|||||||||||VISIT12345

# Parsing HL7 v2 with Python
def parse_hl7(message):
    segments = message.strip().split('
')
    parsed = {}
    for seg in segments:
        fields = seg.split('|')
        seg_type = fields[0]
        if seg_type == 'PID':
            pid_fields = fields[3].split('^') if len(fields) > 3 else []
            parsed['mrn'] = pid_fields[0] if pid_fields else ''
            name_parts = fields[5].split('^') if len(fields) > 5 else []
            parsed['last_name'] = name_parts[0] if name_parts else ''
            parsed['first_name'] = name_parts[1] if len(name_parts) > 1 else ''
    return parsed

FHIR — Modern RESTful Healthcare APIs

Fast Healthcare Interoperability Resources (FHIR, pronounced “fire”) combines the healthcare domain knowledge of HL7 with modern web technologies. FHIR represents healthcare data as resources — JSON or XML objects with well-defined structures — accessed through a RESTful API. Each resource type (Patient, Observation, MedicationOrder, Condition, etc.) has a standard set of properties and a canonical URL. FHIR addresses many of HL7 v2’s shortcomings: it uses JSON (familiar to web developers), supports modern authentication (OAuth 2.0), and provides built-in versioning, search, and extensibility.

// FHIR Patient resource (JSON)
{
  "resourceType": "Patient",
  "id": "example",
  "identifier": [{
    "system": "urn:oid:1.2.3.4.5.6.7",
    "value": "12345"
  }],
  "name": [{
    "family": "Doe",
    "given": ["Jane"]
  }],
  "gender": "female",
  "birthDate": "1985-03-22",
  "address": [{
    "line": ["123 Main St"],
    "city": "Boston",
    "state": "MA",
    "postalCode": "02114"
  }]
}

// FHIR RESTful interactions
GET /fhir/Patient/example                    // read patient
GET /fhir/Patient?birthdate=gt1980-01-01     // search patients
POST /fhir/Observation                        // create observation
PUT /fhir/Patient/example                     // update patient

DICOM — Medical Imaging

Digital Imaging and Communications in Medicine (DICOM) is the international standard for medical imaging. It defines both the file format for storing images (with embedded metadata) and the network protocol for transmitting them. Each DICOM file contains a header with hundreds of standardized tags covering patient demographics, study information, equipment parameters, and image acquisition details, followed by the pixel data. DICOM supports all major imaging modalities: CT, MRI, X-ray, ultrasound, PET, and mammography.

import pydicom

# Read and inspect a DICOM file

ds = pydicom.dcmread("scan.dcm")
# Access metadata tags

print(f"Patient: {ds.PatientName}")

print(f"Study Date: {ds.StudyDate}")

print(f"Modality: {ds.Modality}")            # CT, MR, XA, US, etc.

print(f"Image Size: {ds.Rows} x {ds.Columns}")

print(f"Slice Thickness: {ds.SliceThickness} mm")
# Extract pixel data as numpy array

pixels = ds.pixel_array

print(f"Pixel data shape: {pixels.shape}")
# Anonymize patient information

ds.PatientName = "ANONYMIZED"

ds.PatientID = "000000"

ds.save_as("scan_anonymized.dcm")
Key Challenges in Health Data Exchange
Even with standards in place, healthcare interoperability faces significant practical challenges. Semantic mapping is one of the hardest: different systems may use different terminology for the same clinical concept. For example, one system might code "heart attack" as 410.00 (ICD-9) while another uses I21.0 (ICD-10) and a third uses 22298006 (SNOMED CT). Mapping tables must translate between these coding systems, and mismatches can cause clinical decision support errors. Patient identity matching is another challenge — the same patient may have different medical record numbers across different hospitals. Probabilistic matching algorithms using name, date of birth, and address are used to link records across institutions. Privacy and security regulations (HIPAA in the US, GDPR in Europe, PDPA in India) impose strict requirements on how health data is stored, transmitted, and accessed. All health data exchange must be encrypted in transit and at rest, with audit logging and access controls to track who viewed or modified patient data.
Practical Integration Approaches
The most practical approach for new health IT projects is a FHIR-first strategy with HL7 v2 fallback. Expose all new data through FHIR APIs, use HL7 v2 adapters to communicate with legacy systems that do not yet support FHIR, and implement a terminology service for code mapping between SNOMED CT, ICD-10, LOINC, and local coding systems. Open-source tools like HAPI FHIR (Java), fhir.resources (Python), and Mirth Connect (integration engine) can accelerate implementation. For cloud-native architectures, managed FHIR services like Azure API for FHIR and Google Healthcare API provide scalable, HIPAA-compliant platforms that handle the infrastructure complexity.
Interoperability in healthcare is not just a technical challenge — it involves governance, patient consent, privacy regulations (HIPAA in the US, GDPR in Europe), and semantic mapping between different coding systems (SNOMED CT, ICD-10, LOINC). FHIR is increasingly the standard for new integrations, but HL7 v2 will remain in production for years due to the massive installed base. A practical strategy is to use FHIR as the API layer for new applications while maintaining HL7 v2 bridges to legacy systems.