FHIR mapping in hospitals: From heterogeneous data to clear structures
How files, interface messages, and system data become usable FHIR resources for clinical workflows and interoperability platforms.
FHIR mapping is more than format conversion
FHIR stands for Fast Healthcare Interoperability Resources and is an HL7 standard for exchanging healthcare information electronically. Its basic building block is the resource: a clearly defined data object for patients, observations, diagnostic findings, medication, documents, and other healthcare concepts.
Mapping therefore means more than converting a file into JSON or XML. The meaning of an item must be identified and represented in the correct place in the target model. Only then do machine-readable data become clinical context that different systems can understand and use.
Many sources, one common structure
Hospitals generate information in many forms: structured system exports, interface messages, spreadsheet rows, PDFs, text documents, and sections of diagnostic reports. Even identical content may be named, coded, or arranged differently depending on the source system.
LLMedi brings these sources into one processing flow. Content, metadata, and existing relationships are captured and classified for the intended FHIR target model while retaining the connection to the original source.
From input to a validated FHIR resource
Reliable mapping consists of several coordinated steps. Automation can handle recurring assignments, while target profiles, rules, and validation ensure that the output fits the specific integration scenario.
- Capture the source: identify structure, free text, metadata, and relationships
- Classify the semantics: normalize clinical concepts, values, units, and references
- Map to FHIR: create appropriate resources and connect them through references
- Validate the target profile: check required fields, cardinalities, terminology, and other constraints
Individual data points become patient context
A laboratory value can be represented as an Observation and connected to a DiagnosticReport and the corresponding Patient resource. A prescription can be structured as a MedicationRequest, while an existing medical letter remains traceable to its source through a DocumentReference.
The real value lies in these relationships. FHIR resources are not treated as isolated records; they form a traceable clinical context that connected systems can use for their intended workflows.
FHIR, profiles, ISiK, and the interoperability platform
The base FHIR standard intentionally leaves room for different use cases. Profiles constrain resources for a particular purpose through requirements such as mandatory fields, terminology bindings, or cardinalities. A mapping must therefore target the intended profile rather than merely producing a generic FHIR resource.
For German hospitals, ISiK specifies interoperable exchange through FHIR implementation guides and a REST-based interface. Appropriately mapped and validated resources can be passed to an interoperability platform and made available to connected clinical systems.
Automate without hiding ambiguity
Not every source is unambiguous. Abbreviations, local labels, missing units, or incomplete documents can allow several interpretations. Good mapping exposes such uncertainty instead of silently transferring it into apparently clean structures.
LLMedi combines automated extraction and semantic classification with defined mapping rules and technical validation. This can reduce repetitive work while keeping the output controllable and aligned with the target system.
Mapping within the hospital infrastructure
Processing can take place entirely within the hospital’s designated infrastructure. Source data therefore does not need to be transferred to a public cloud provider for extraction, classification, or FHIR generation.
Together with the institution, the relevant sources, target profiles, and destination interoperability platform or systems are defined. The result is not a universal one-off connector, but a controlled data flow designed for the specific system landscape.
Standards & sources
From insight to application
See how LLMedi fits your clinical workflow.
Continue reading
AI for hospitals: Where clinical support really begins
How clinical AI connects patient context, consultation, and documentation while fitting into existing hospital workflows.
On-premises AI in hospitals: What local deployment means
An introduction to local AI infrastructure, on-site data processing, and integration with existing clinical systems.