FHIR CI-Build

This is the Continuous Integration Build of FHIR (will be incorrect/inconsistent at times).
See the Directory of published versions icon

6.4 Resource AuditEvent - Content

Security icon Work GroupMaturity Level: 4 Trial UseSecurity Category: Not Classified Compartments: Device, Group, Patient, Practitioner

A record of an event relevant for purposes such as operations, privacy, security, maintenance, and performance analysis.

6.4.1 Scope and Usage

The audit event is based on the IHE-ATNA Audit record definitions icon, originally from RFC 3881 icon, and now managed by DICOM (see DICOM Part 15 Annex A5 icon).

  • ASTM E2147 - Setup the concept of security audit logs for healthcare including accounting of disclosures
  • IETF RFC 3881 - Defined the Information Model (IETF rule forced this to be informative)
  • DICOM Audit Log Message - Made the information model Normative, defined Vocabulary, Transport Binding, and Schema
  • IHE ATNA - Defines the grouping with secure transport and access controls; and defined specific audit log records for specific IHE transactions.
  • NIST SP800-92 - Shows how to do audit log management and reporting - consistent with our model
  • HL7 PASS - Defined an Audit Service with responsibilities and a query interface for reporting use
  • ISO 27789 - Defined the subset of audit events that an EHR would need
  • ISO/HL7 10781 EHR System Functional Model Release 2
  • ISO 21089 Trusted End-to-End Information Flows

This resource is managed collaboratively between HL7, DICOM, and IHE.

A record of an event relevant for purposes such as operations, privacy, security, maintenance, and performance analysis.

6.4.2 Background and Context

All actors - such as applications, processes, and services - involved in an auditable event should record an AuditEvent. This will likely result in multiple AuditEvent entries that show whether privacy and security safeguards, such as access control, are properly functioning across an enterprise's system-of-systems. Thus, it is typical to get an auditable event recorded by both the application in a workflow process and the servers that support them. For this reason, duplicate entries are expected, which is helpful because it may aid in the detection of security, privacy, or other operational problems. For example, fewer than expected actors being recorded in a multi-actor process or attributes related to those records being in conflict, which is an indication of a security problem. There may be non-participating actors, such as trusted intermediary, that also detect a security, privacy, or operational relevant event and thus would record an AuditEvent.

Security relevant events are not limited to communications or RESTful events. They include:

  • software start-up and shutdown;
  • user login and logout;
  • access control decisions;
  • configuration events;
  • software installation;
  • policy rules changes; and
  • manipulation of data that exposes the data to users.

See the Audit Event Category vocabulary for guidance on some security relevant event categories.

The AuditEvent resource holds the details of an event in terms of who, what, where, when, and why. Where the identification of the who participated is the agent. An agent can be a person, an organization, software, device, or other actors that may be ascribed responsibility. What objects were used/created/updated is recorded as the entity. An entity is an identifiable physical, digital, conceptual or other kind of thing; entities may be real or imaginary.

The content of an AuditEvent is primarily intended for administrative use; used by security system administrators, security and privacy information managers, records management personnel, etc. The AuditEvent may also inform the Patient about uses of their data. This content can be accessible or used directly by other healthcare users, such as providers or patients for gaining insight into who and what has been done. The AuditEvent record includes very sensitive information so access to the AuditEvent would be highly privileged and controlled. For example, when providing AuditEvent to a patient the data feed would be limited to the Patient compartment, and the content may be subsetted or masked in order to meet privacy needs. The AuditEvent record is not intended to replace other audit logs, but rather used to enhance them, or to be used as an API to many audit logs.

Relationship of AuditEvent and Provenance resources are often (though not exclusively) created by the application responding to the create/read/query/update/delete/execute etc. event. A Provenance resource contains overlapping information, but is a record-keeping assertion that gathers information about the context in which the information in a resource "came to be" in its current state, e.g., whether it was created de novo or obtained from another entity in whole, part, or by transformation. Provenance resources are prepared by the application that initiates the create/update of the resource and may be persisted with the AuditEvent target resource.

6.4.3 References to this Resource

6.4.4 Resource Content

Structure

NameFlagsCard.TypeDescription & Constraintsdoco
.. AuditEvent TU DomainResource Record of an event

Elements defined in Ancestors: id, meta, implicitRules, language, text, contained, extension, modifierExtension
... type Σ 1..1 CodeableConcept High level categorization of audit event
Binding: Audit Event ID (Example)
... subtype Σ 0..* CodeableConcept Specific type of event
Binding: Audit Event Sub-Type (Example)

... action Σ 0..1 code Type of action performed during the event
Binding: Audit Event Action (Required)
... occurred[x] 0..1 When the activity occurred
.... occurredPeriod Period
.... occurredDateTime dateTime
... recorded Σ 1..1 instant Time when the event was recorded
... outcome Σ 0..1 BackboneElement Whether the event succeeded or failed
.... code Σ 1..1 Coding Whether the event succeeded or failed
Binding: Audit Event Outcome (Preferred)
.... detail Σ 0..* CodeableConcept Additional outcome detail
Binding: Audit Event Outcome Detail (Example)

... authorization Σ 0..* CodeableConcept Authorization related to the event
Binding: PurposeOfUse icon (Example)

... basedOn 0..* Reference(Any) Workflow authorization within which this event occurred

... patient ΣTU 0..1 Reference(Patient) The patient is the subject of the data used/created/updated/deleted during the activity
... encounter 0..1 Reference(Encounter) Encounter within which this event occurred or which the event is tightly associated
... agent Σ 1..* BackboneElement Actor involved in the event

.... type 0..1 CodeableConcept How agent participated
Binding: Participation Role Type (Preferred)
.... role 0..* CodeableConcept Agent role in the event
Binding: Security Role Type (Example)

.... who Σ 1..1 Reference(Practitioner | PractitionerRole | Organization | CareTeam | Patient | Device | DeviceDefinition | RelatedPerson | Group | HealthcareService) Identifier of who
.... requestor Σ 0..1 boolean Whether user is initiator
.... location 0..1 Reference(Location) The agent location when the event occurred
.... policy 0..* uri Policy that authorized the agent participation in the event

.... network[x] 0..1 This agent network location for the activity
..... networkReference Reference(Endpoint)
..... networkUri uri
..... networkString string
.... authorization 0..* CodeableConcept Allowable authorization for this agent
Binding: PurposeOfUse icon (Example)

... source Σ 1..1 BackboneElement Audit Event Reporter
.... site 0..1 Reference(Location) Logical source location within the enterprise
.... observer Σ 1..1 Reference(Practitioner | PractitionerRole | Organization | CareTeam | Patient | Device | RelatedPerson) The identity of source detecting the event
.... type 0..* CodeableConcept The type of source where event originated
Binding: Audit Event Source Type (Preferred)

... entity Σ 0..* BackboneElement Data or objects used

.... what Σ 0..1 Reference(Any) Specific instance of resource
.... role 0..1 CodeableConcept What role the entity played
Binding: Audit Event Entity Role (Example)
.... securityLabel 0..* CodeableConcept Security labels on the entity
Binding: Example set of Security Labels (Example)

.... description 0..1 string Descriptive text
.... query Σ 0..1 base64Binary Query parameters
.... detail 0..* BackboneElement Additional Information about the entity

..... type 1..1 CodeableConcept The name of the extra detail property
Binding: Audit Event Entity Detail Type (Example)
..... value[x] 1..1 Property value
...... valueQuantity Quantity
...... valueCodeableConcept CodeableConcept
...... valueString string
...... valueBoolean boolean
...... valueInteger integer
...... valueRange Range
...... valueRatio Ratio
...... valueTime time
...... valueDateTime dateTime
...... valuePeriod Period
...... valueBase64Binary base64Binary
.... agent 0..* see agent Entity is attributed to this agent


doco Documentation for this format icon

See the Extensions for this resource

 

Additional definitions: Master Definition XML + JSON, XML Schema/Schematron + JSON Schema, ShEx (for Turtle) + see the extensions, the spreadsheet version & the dependency analysis

6.4.4.1 Terminology Bindings

Path ValueSet Type Documentation
AuditEvent.type AuditEventID Example

Event Categories for Audit Events - defined by DICOM with some FHIR specific additions.
AuditEvent.subtype AuditEventSubType Example

More detailed code concerning the type of the audit event - defined by DICOM with some additional FHIR, HL7, and other additions.
AuditEvent.action AuditEventAction Required

Indicator for type of action performed during the event that generated the event.
AuditEvent.severity AuditEventSeverity Required

The severity of the audit entry.
AuditEvent.outcome.code AuditEventOutcome (a valid code from Issue Severity) Preferred

Indicates whether the event succeeded or failed.
AuditEvent.outcome.detail AuditEventOutcomeDetail (a valid code from Operation Outcome Codes icon) Example

Indicates more detailed reason for outcome.
AuditEvent.authorization PurposeOfUse icon Example

Supports communication of purpose of use at a general level.
AuditEvent.agent.type ParticipationRoleType Preferred

This FHIR value set is comprised of Actor participation Type codes, which can be used to value FHIR agents, actors, and other role elements. The codes are intended to express how the agent participated in some activity. Sometimes refered to the agent functional-role relative to the activity.
AuditEvent.agent.role SecurityRoleType (a valid code from Sample Codes for Security Structural Role) Example

This value set contains example structural roles. In general, two types of roles can be distinguished: structural roles and functional roles. Structural Roles reflect human or organizational categories (hierarchies), and describe prerequisites, feasibilities, or competences for actions. Functional roles are bound to the realization or performance of actions.
AuditEvent.agent.authorization PurposeOfUse icon Example

Supports communication of purpose of use at a general level.
AuditEvent.source.type AuditEventSourceType Preferred

The type of process where the audit event originated from. Use of these codes is not required but is encouraged to maintain translation with DICOM AuditMessage schema.
AuditEvent.entity.role AuditEventEntityRole Example

Code representing the role the entity played in the audit event.
AuditEvent.entity.securityLabel SecurityLabelExamples Example

A sample of security labels from Healthcare Privacy and Security Classification System as the combination of data and event codes.
AuditEvent.entity.detail.type AuditEventEntityDetailType (a valid code from Sample Codes for AuditEvent.entity.detail.type) Example

The type of additional detail about an entity used in an event.

6.4.4.2 Using Coded Values

The AuditEvent resource and the ATNA Audit record are used in many contexts throughout healthcare. The coded values defined in the "extensible" bindings above are those widely used and/or defined by DICOM, IHE or ISO, who defined these codes to meet very specific use cases. These codes should be used when they are suitable. When needed, other codes can be defined.

Note: When using codes from a vocabulary, the display element for the code can be left off to keep the AuditEvent size small and minimize impact of a large audit log of similar entries.

The set of codes defined for this resource is expected to grow over time, and additional codes may be proposed / requested using the "Propose a change" link above below.

6.4.4.3 Event codes for Common Scenarios

This table summarizes common event scenarios, and the codes that should be used for each case.

Scenario category code action Other
User Login (example) 110114 icon User Authentication 110122 icon User Authentication E Execute One agent which contains the details of the logged-in user.
User Logout (example) 110114 icon User Authentication 110123 icon User Logout E Execute One agent which contains the details of the logged-out user.
REST operation logged on server (example) rest RESTful Operation [code] defined for operation * (see below) Agent for logged in user, if available.
Search operation logged on server (example) rest RESTful Operation [code] defined for operation E Execute Agent for logged in user, if available, and one object with a query element. The Execute action is used as the server must execute the search parameters to get the results, whereas a Read action identifies a specific object.
Break-Glass started (example) 110113 icon Security Alert 110127 icon Emergency Override Started E Execute Agent is the user who is authorized to break-glass and has declared an emergency override. Note there is an Emergency Override Stopped code that can be used to indicate the closing of the break-glass event, when it is known.

Audit Event Actions for RESTful operations:

Operation Action
create C
read, vread, history-instance, history-type, history-system R
update U
delete D
transaction, operation, conformance, validate, search, search-type, search-system E

6.4.4.3.1 Recording Search Events

A search event is recorded as an Execute action as the server must execute the search parameters to get the results. The category is a rest operation. The code should be search. The Server is identified in an .agent as the role Destination Role ID, and the client is identified in an .agent as the role Source Role ID. Additional .agent elements may be used to identify user, application, organization, etc.

A Search Event records one .entity element that holds the search request, and should not record the contents of the search response so as to limit duplication of sensitive health information that is already present in the system, and discoverable by replaying the search request.

The AuditEvent.entity.query shall hold the whole WHOLE http header and body encoded as base64binary. This should preserve as much of the raw http header and body as possible to best capture any attempts by clients or intermediaries to misbehave. There should be no sanitization or normalization of this value.

The FHIR specification defines a harmonized search parameter string, which is returned in the searchset bundle as the .link.url on the .link for self. This string could be recorded in the AuditEvent.entry.description as it is well behaved and represents what was actually processed as search parameters. See: conformance icon

Where there are identifiable Patient subject(s) associated with the returned Resource(s), the AuditEvent.patient should be used to record the Patient as the subject of the data or activity. When multiple patient results are returned one AuditEvent is created for every Patient identified in the resulting search set. Note this is true when the search set bundle includes any number of resources that collectively reference multiple Patients. This includes one Resource with multiple subject values, or many Resources with single subject values that are different.

6.4.4.4 Encoding a FHIR operation outcome

FHIR interactions can result in a rich description of the outcome using the OperationOutcome. The OperationOutcome Resource is a collection of error, warning or information messages that result from a system action. This describes in detail the outcome of some operation, such as when a RESTful operation fails.

When recording into an AuditEvent that some FHIR interaction has happened, the AuditEvent should include the OperationOutcome from that FHIR interaction. This is done by placing the OperationOutcome into an AuditEvent.entity. Likely as a contained resource, given that OperationOutcome resources often are not persisted.

entity.what is the OperationOutcome -- Likely contained

entity.type is code OperationOutcome

entity.description explains why this OperationOutcome was included.

See transaction failure example: When a client attempts to post (create) an Observation Resource, using a server Patient endpoint; this would result in an error with an OperationOutcome.

6.4.4.5 authorization and agent.authorization

The AuditEvent provides the element AudientEvent.authorization to convey the purpose of use for the whole event and AuditEvent.agent.authorization to convey the purpose of use that a particular actor (machine, person, software) was involved in the event.

AuditEvent.authorization is an element at the level of AuditEvent and can convey the purpose of the activity that resulted in the event. This will occur when the system that is reporting the event is aware of the purpose of the event. A specific example would be a radiology reporting system where a radiologist has created and is sending a finished report. This system likely knows the purpose, e.g., "treatment". It is multi-valued because the one event may be related to multiple purposes.

It is also commonplace that the reporting system does not have information about the purpose of the event. In these cases, the event report would not have an authorization.

It is also likely that the same event will be reported from different perspectives, e.g., by both the sender and recipient of a communication. These two different perspectives can have different knowledge regarding the purposeOfUse authorization.

AuditEvent.agent.authorization is an element at the level of agent within AuditEvent. This describes the reason that this person, machine, or software is participating in the activity that resulted in the event. For example, an individual person participating in the event may assert a purpose of use from their perspective. It is also possible that they are participating for multiple reasons and report multiple purposeOfUse.

The reporting system might not have knowledge regarding why a particular machine or person was involved and would omit this element in those cases.

When the same event is reported from multiple perspectives, the reports can have different knowledge regarding the purpose.

6.4.4.6 Patient as subject of data or activity reference in AuditEvent.patient

It is a best practice to include a reference to the Patient affected by any auditable event, in order to enable Privacy Accounting of Disclosures and Access Logs, and to enable privacy office and security office audit log analysis. Reasonable efforts should be taken to assure the Patient is recorded, but it is recognized that there are times when this is not reasonable.

Where an activity impacts more than one Patient subject; multiple AuditEvent resources should be recorded, one for each Patient subject. This best enables segmentation of the AuditEvent details so as to limit the Privacy impact. The use of multiple AuditEvent is a best-practice and should be driven by a Policy. There will be cases where the use of multiple AuditEvent resources are not necessary, such as public health reporting.

To record a REST interaction or $operation, it is often necessary to complete the transaction in order to determine the Patient subject. Inspection of the potential returned results may be necessary. Some REST and $operations include parameters limiting the results to a specific Patient, in these cases this parameter informs the inclusion of the Patient reference.

Implementation Guides may make the AuditEvent requirements more clear given the workflow or security context mandated by the Implementation Guide.

6.4.5 Search Parameters

Search parameters for this resource. See also the full list of search parameters for this resource, and check the Extensions registry for search parameters on extensions related to this resource. The common parameters also apply. See Searching for more information about searching in REST, messaging, and services.

Name Type Description Expression In Common
action token Type of action performed during the event AuditEvent.action
agent reference Identifier of who AuditEvent.agent.who
(Practitioner, Group, Organization, CareTeam, Device, DeviceDefinition, Patient, HealthcareService, PractitionerRole, RelatedPerson)
agent-role token Agent role in the event AuditEvent.agent.role
based-on reference Reference to the service request. AuditEvent.basedOn
(Any)
date date Time when the event was recorded AuditEvent.recorded 26 Resources
encounter reference Encounter related to the activity recorded in the AuditEvent AuditEvent.encounter
(Encounter)		 29 Resources
entity reference Specific instance of resource AuditEvent.entity.what
(Any)
entity-desc string Description of an entity AuditEvent.entity.description
entity-role token What role the entity played AuditEvent.entity.role
outcome token Whether the event succeeded or failed AuditEvent.outcome.code
patient reference Where the activity involved patient data AuditEvent.patient
(Patient)		 65 Resources
policy uri Policy that authorized event AuditEvent.agent.policy
purpose token The authorization (purposeOfUse) of the event AuditEvent.authorization | AuditEvent.agent.authorization
source reference The identity of source detecting the event AuditEvent.source.observer
(Practitioner, Organization, CareTeam, Device, Patient, PractitionerRole, RelatedPerson)
subtype token More specific code for the event AuditEvent.subtype
type token Type (category) of event AuditEvent.type