Redefining Emergency Department patient administration using a FHIR-native model

Legacy Patient Administration Systems slow hospital operations and break down under Emergency Department conditions.

Linear workflows, heavy data requirements, and tightly coupled integrations push staff to paper workarounds and delayed reconciliation. This work defined a new PAS model designed to support uncertainty, handover, and time-critical admissions while integrating cleanly with modern hospital platforms.

At a glance

Delivery focus: Proof-of-life PAS MVP for Emergency Department admissions

System shift: FHIR-native PAS operating alongside EMR and patient flow systems

Scope: Admission workflows under emergency, time-critical conditions

Operating environment

Context

  • Organisation: Telstra Health

  • Environment: Hospital administration and patient flow systems

  • Users: Emergency Department administrative and clinical staff

  • Role: Lead UX Designer

  • Timeframe: 6-month MVP delivery

Constraints

  • Coexistence with legacy PAS

  • Complex integrations and reporting dependencies

  • Clinical safety and data integrity requirements

  • No external user testing budget

What existed and why it mattered

PART 1

The outcome

At the end of six months, the work resulted in:

  • a working PAS MVP covering a complete Emergency Department admission journey

  • a coherent interaction and system model aligned to FHIR data structures

  • a demonstrable product showing how PAS could operate alongside EMR and patient flow systems

  • a clear technical and product direction suitable for continuation and expansion

This phase was not intended to demonstrate adoption or efficiency gains. Success was defined as viability, coherence, and technical credibility.

Why PAS breaks down in Emergency

Emergency Departments operate with:

  • uncertain or partial patient identity at arrival

  • frequent interruptions and handovers

  • rapid movement across physical locations

  • high cognitive load and time pressure

Most legacy PAS platforms assume:

  • complete data capture upfront

  • linear task completion

  • single-system ownership of patient information

These assumptions do not hold in ED.

In practice, staff rely on paper, memory, or duplicate entry to stabilise patient flow. PAS interaction is deferred until workload allows, and records are corrected later. This behaviour is not accidental. It is a response to systems that slow care when used as designed.

Visual 1 - Legacy PAS reality. PAS interfaces being used in Australian hospitals . These prioritise exhaustive data entry over flow, creating cognitive load and error risk under pressure.

Duplication and multi-system fragmentation

Hospitals operate multiple systems that capture overlapping patient information:

  • Patient Administration Systems

  • Electronic Medical Records

  • Bed management and patient flow tools

  • Department-specific systems

These systems often store similar data using different identifiers, formats, and update rules. When identity or encounter data cannot be safely amended in real time, duplication becomes the default.

Multiple records are created for the same patient and reconciled later.

This duplication increases:

  • downstream error risk

  • administrative rework

  • reporting inconsistencies

  • loss of confidence in the system of record

PAS failure in ED is therefore not a UI problem. It is a system integrity problem.

Why this problem persisted

PAS sits at the centre of hospital operations and underpins:

  • funding and reporting

  • billing and compliance

  • clinical and operational workflows

  • hundreds of system integrations

Most PAS platforms are:

  • over twenty years old

  • monolithic and tightly coupled

  • heavily customised per site

  • risky to change incrementally

Retrofitting legacy PAS introduces safety and governance concerns. As a result, many organisations avoid change entirely. Innovation stalled, and workarounds became normalised.

Why a FHIR-native model changes PAS behaviour

FHIR is a modern healthcare data standard designed to support interoperability between systems.

Rather than treating patient data as something owned by a single system, FHIR enables a distributed model where multiple systems participate in maintaining a shared view of patient information.

What FHIR enables

  • Patient, Encounter, and Location as discrete resources
    Core concepts are modelled independently and linked, rather than embedded in a single record.

  • Incremental updates instead of transactional overwrites
    Information can be added or refined over time without requiring full record completion upfront.

  • Correction and versioning without losing history
    Data can be amended safely while preserving an audit trail.

  • Multiple systems reading and writing to the same data layer
    PAS, EMR, patient flow, billing, and analytics systems can interact with shared data without duplicating records.

Why this mattered for PAS

Traditional PAS platforms assume ownership of the patient record and enforce completeness early in the workflow. This clashes with Emergency Department reality, where information is often partial and evolves over time.

A FHIR-native PAS does not need to own the full record. Instead, it can:

  • support uncertainty and staged data capture

  • allow patient identity to be refined as information becomes available

  • coexist with legacy PAS during transition

  • reduce duplication by design rather than policy

This shifts PAS from a monolithic system of record to a collaborative participant in a broader hospital data ecosystem.

Visual 2 - FHIR-native PAS system model The PAS participates in a shared clinical data layer rather than acting as a siloed system of record.

Insight

A FHIR-native PAS shifts from record ownership to participation in a shared clinical data ecosystem.

Designing a viable PAS system for Emergency

PART 2

Framing the work as system validation, not feature design

This phase focused on validating whether a modern PAS model could operate safely within real Emergency Department conditions.

The objective was not to design a complete PAS. It was to demonstrate that a new system model, interaction approach, and data architecture could support a critical hospital workflow without relying on legacy assumptions.

The work treated Emergency admission as a system problem spanning people, software, data, and governance, rather than a sequence of screens.

Defining the MVP patient journey

The MVP was deliberately scoped to a single, high-risk journey:

  • ED arrival and triage

  • Identity check or creation

  • Patient movement across ED locations

  • Admission completion

  • Transfer out of ED

This journey represented the minimum path required to validate the PAS model.

If ED admission could operate safely under time pressure, uncertainty, and handover, the underlying architecture and interaction patterns could scale to the rest of the hospital.

Visual 3 - ED MVP journey map Shows how Kyra PAS supports safe progression from arrival to transfer when identity is partial, time is constrained, and patient location changes frequently.

Coordinating people, product, and data

Emergency admission involves:

  • frontstage activity by clerical staff, triage nurses, and clinicians

  • midstage coordination through the PAS

  • backstage data updates across EMR, bed management, billing, and reporting systems

Rather than treating these as separate concerns, the work mapped them together to ensure interaction design, system behaviour, and data ownership remained aligned.

Creating the service blueprint clarified:

  • where uncertainty enters the workflow

  • which system owns which decisions at each stage

  • how FHIR resources are created and updated incrementally

  • how downstream systems remain informed without blocking ED flow

This artefact makes explicit how admission work, system behaviour, and data responsibilities align across roles and systems.

Visual 4 - Service design blueprint. Shows how ED staff, Kyra PAS, the FHIR Clinical Data Repository, EMR, bed management, and downstream services coordinate during Emergency Department admission.

Designing without a built system or testing budget

There was no production system and no budget for external user testing.

Validation relied on:

  • Low-fidelity wireframes reviewed with engineering

  • SME walkthroughs with nurses, doctors, and PAS specialists

  • Architectural alignment sessions

  • Frequent internal critique and iteration

Design progressed from structural wireframes to interaction models, then to working prototypes suitable for demonstration.

Decisions were validated through coherence across the service blueprint, interaction flows, and data model.

Designing for real ED pressure

The design strategy prioritised:

  • predictable interaction patterns

  • reduced cognitive load

  • clear visibility of patient state and movement

  • minimal disruption during interruptions and handover

Key decisions included:

  • anchoring the PAS around the ED workflow

  • consolidating flow visibility into a single operational surface

  • maintaining consistent navigation across screens

  • limiting cognitive load through progressive disclosure

  • aligning terminology with real-world clinical practice

The work prioritised clarity and predictability for staff operating under sustained time pressure and frequent interruption.

Patient movement as interaction

A major departure from legacy PAS was treating patient movement as a primary interaction rather than an administrative task.

Instead of navigating multiple screens or forms:

  • Patients move visually between locations

  • State updates automatically

  • Timers and status indicators remain visible

  • Detailed actions are accessible without breaking flow

This supported rapid situational awareness during high-pressure moments.

Visual 5 – Working prototype Showing ED movement and search

Decision

Patient movement was treated as a first-class interaction to support situational awareness under pressure.

Concrete design decisions

This interaction model required a small set of explicit, repeatable design decisions.

Design decisions were explicit and repeatable:

  • a reusable patient card mapped to FHIR resources

  • a central ED movement board as the primary surface

  • consistent arrival and admission flows

  • component-based UI for scalability

  • terminology alignment to reduce ambiguity

  • status and timer indicators to support triage awareness

Validating interaction before visual refinement

Visual 6 – Interaction wireframes Interaction wireframes used to define triage logic, patient cards, and detail panels underpinning the working prototype. Focus was placed on flow, sequencing, and safety-critical information rather than visual refinement.

Working without clinical training

As the sole designer without a medical background:

  • Product experts, nurses, and doctors were consulted frequently

  • Assumptions were validated through discussion rather than preference

  • Language, sequencing, and edge cases were pressure-tested

  • Decisions were grounded in operational reality

This reinforced a workflow-first, system-led design approach rather than an interface-led one.

Results, measurement and impact

PART 3

How success was measured

Success criteria were defined for a proof-of-life phase.

Primary measures

  • a complete ED admission journey could be demonstrated end-to-end

  • identity uncertainty could be handled without blocking flow

  • data structures aligned to FHIR and supported correction over time

  • the system story held together technically and operationally

Quantifiable indicators available at this stage included:

  • six-month delivery of a working MVP

  • number of core workflows demonstrated live

  • breadth of system integrations shown in architecture artefacts

  • scale of stakeholder engagement during showcase sessions

Demonstration and reception

Kyra PAS was presented broadly across the health portfolio.

  • attendance exceeded 400 stakeholders

  • feedback focused on feasibility and clarity of direction

  • discussion centred on next steps rather than fundamental viability

Visual 7 - MVP showcase milestone Shows Kyra PAS presented as a viable platform direction to a broad health leadership audience.

Outcome

Stakeholder discussion shifted from viability to continuation and scale.

Technical credibility

The demonstration showed:

  • FHIR-based data exchange

  • coexistence with legacy PAS

  • HL7 v2 integration paths

  • SMART on FHIR extensibility

Visual 8 - Staged migration to a FHIR-native PAS  Illustrates how Kyra PAS enables staged migration rather than wholesale replacement of legacy PAS.

Reflection and next steps

What this work established

  • a viable PAS interaction model

  • a defensible technical foundation

  • a clear demonstration of PAS operation under real hospital conditions

This was foundational work rather than a polished product.

Limits of the timeframe

Six months constrained:

  • depth of functionality

  • external validation

  • breadth of scenarios

Next steps identified

  • extend the MVP to additional admission scenarios

  • deepen identity and duplicate-management tooling

  • formalise a PAS design system

  • validate workflows in live hospital environments