Skip to content

EPR-Specific Hazards

This document catalogues hazard categories specific to an Electronic Patient Record system. It serves as a starting framework for Quill Medical's Hazard Log — not a substitute for formal hazard identification workshops.

These categories are informed by HSSIB investigations, NHS incident reports, published literature on EHR safety, and the DCB 0129 specification.


1. Patient Identification

Hazard Potential Clinical Impact Example Causes
Wrong patient record opened or displayed Clinical decisions based on wrong patient's data — wrong treatment, missed allergies, wrong diagnosis Similar names, failure to verify demographics, session state errors, cached data, browser back-button
Patient records merged incorrectly Permanent data corruption — one patient receives another's medical history Duplicate detection failures, manual merge errors
Patient not found / duplicate created Fragmented records — clinician sees incomplete picture Poor search UX, PDS integration failures, data quality issues
Patient identity not confirmed before action Data entered or viewed for wrong patient No identity banner, insufficient verification prompts

Quill Medical controls: UUID-based internal identifiers (not NHS Number for primary key), FHIR PDS integration for demographics, patient banner with identity confirmation.


2. Data Display and Presentation

Hazard Potential Clinical Impact Example Causes
Clinical data displayed against wrong patient Treatment decisions on wrong data Session management bugs, race conditions, caching, concurrent tab usage
Data displayed in wrong order Clinician acts on outdated information Sort order bugs, timezone issues, inconsistent date formatting
Clinically significant data not visible Missed critical information (allergy, key diagnosis, critical result) Responsive layout issues, information overload, poor visual hierarchy, scroll/pagination hiding data
Data displayed without context Clinical decisions based on stale or ambiguous data Lab result without date/time, observation without units, truncated values
Allergy or alert not prominently displayed Medication prescribed to allergic patient; known risk not communicated Alert display failures, poor visual prominence, alert fatigue

Quill Medical controls: Consistent patient banner on all clinical screens, responsive design testing, clear date/time display conventions, allergy/alert prominence in UI hierarchy.


3. Data Entry and Recording

Hazard Potential Clinical Impact Example Causes
Data entered against wrong patient Record contamination — may affect ongoing care of both patients No identity confirmation on entry screens, auto-complete errors, tab-switching
Free-text entry not coded or structured Downstream systems cannot use the data (e.g. free-text allergy does not trigger drug-allergy checks) Insufficient terminology binding, UX that doesn't encourage coded entry
Data loss on save Clinical information lost — may affect ongoing care Network failures, session timeouts, browser crashes, concurrent edit conflicts
Incorrect clinical coding Wrong diagnosis/problem recorded, affects decision-making Poor SNOMED CT search, ambiguous code descriptions, clinician unfamiliarity with coding
Copy-paste propagation of errors Incorrect information carried forward across encounters Template/copy-paste functionality without review prompts

Quill Medical controls: OpenEHR archetypes enforce structured data capture, SNOMED CT terminology binding, save confirmation patterns, soft-delete with audit trail.


4. System Availability and Performance

Hazard Potential Clinical Impact Example Causes
System completely unavailable Clinicians cannot access patient records — delays to care, uninformed clinical decisions Server failure, network outage, database corruption, deployment error
Slow system response Clinicians skip important checks, develop unsafe workarounds, reduced thoroughness Database performance, API bottlenecks, large dataset rendering
Data not synchronised Clinician sees stale data; decisions based on outdated information Eventual consistency, replication lag, offline/PWA sync failures
Partial system failure Some functions work, others silently fail — clinician may not realise data is incomplete Microservice failures, API timeouts, degraded third-party services

Quill Medical considerations: As a PWA, specific attention is needed for offline/degraded network scenarios. What data is cached locally? What is the risk of acting on stale cached data? How is the user informed of sync status?


5. Access Control and Confidentiality

Hazard Potential Clinical Impact Example Causes
Unauthorised access to patient data Breach of confidentiality, potential for data misuse Inadequate RBAC, session hijacking, shared credentials, session not expiring
User sees data outside their role scope Inappropriate clinical decisions; breach of confidentiality RBAC misconfiguration, role inheritance bugs, permission escalation
Audit trail incomplete or missing Cannot trace who accessed or modified data — undermines clinical governance and incident investigation Logging failures, soft-delete not capturing actor/timestamp
Session persists after user leaves Subsequent user accesses previous user's patient context Inadequate session timeout, shared device without logout enforcement

Quill Medical controls: Six-role RBAC model (System Admin, Clinical Admin, Clinician, Clinical Support Staff, Patient, Patient Advocate), soft-delete with full audit trails, UUID identifiers for security.


6. Integration and Interoperability

Hazard Potential Clinical Impact Example Causes
FHIR message fails silently Data believed to be sent/received but is not — affects referrals, results, medications Integration errors, schema mismatches, network failures without user alerts
Data transformation errors Clinical meaning altered during exchange FHIR mapping errors, code system mismatches, unit conversion errors
Third-party service unavailable Dependent functionality fails (PDS lookup, terminology service, identity provider) External service outage, API rate limiting, certificate expiry
Inconsistent data across systems Different systems show different information for the same patient Sync timing, mapping discrepancies, partial update failures

Quill Medical controls: FHIR for demographics and messaging, OpenEHR for clinical data, clear integration error handling and user notification.


7. Configuration and Deployment

Hazard Potential Clinical Impact Example Causes
Incorrect local configuration System behaves differently from what was safety-assessed Admin misconfiguration, inadequate configuration validation, missing constraints
Software update introduces regression Previously safe functionality broken Insufficient regression testing, inadequate release management
Data migration errors Historical data lost or corrupted during migration from legacy systems ETL errors, schema differences, character encoding issues, mapping failures
Configuration change not safety-assessed New risk introduced without going through hazard assessment Change made outside of controlled process, emergency fix without review

8. Clinical Workflow Hazards

Hazard Potential Clinical Impact Example Causes
System workflow mismatches clinical workflow Clinicians develop workarounds that bypass safety controls Poor UX design, failure to involve end-users in design, rigid workflow enforcement
Alert fatigue Clinicians override or ignore clinically important alerts due to excessive alerting Too many low-value alerts, poor alert prioritisation, no alert escalation mechanism
Information overload Clinician misses critical information buried in excessive data Dense UI, unnecessary data display, poor information hierarchy
Task not completed due to interruption Clinical action started but not finished — patient at risk No task tracking, no "incomplete action" warnings, workflow not resumable

Notes

This catalogue is a starting point. Each hazard identified here should be formally assessed in a SWIFT workshop, with causes, existing controls, severity, likelihood, mitigations, and residual risk documented in the Hazard Log.

New hazard categories may emerge as Quill Medical's feature set grows — particularly around clinical decision support, prescribing, results management, and clinical correspondence.