Shared Health Record
Overview
According to OpenHIE standards a Shared Health Record (SHR) should fulfill the following four functions for client EMRs (like OpenMRS):
Store a useful subset of clinical data
Retrieve relevant clinical data as needed
Update the SHR with relevant clinical data as needed
Feed into reporting/analytics pipelines
Use Cases
Primary
A point-of-care system (i.e., LIS, EMR, etc.) should be able to store a normalized subset of clinical information items (that which is deemed appropriate to share) from a patient’s clinical record on that system.
We should be able to store Observations, Allergies, Care Summaries, Care Plans etc.
Store unstructured data along with associated metadata, e.g. a PDF document or digital image with attached patient demographic information
A point-of-care system should be able to retrieve relevant portions (up to the full set) of this clinical record as needed.
Retrieve a longitudinal list of patient clinical information by type, date or other query parameters
A point-of-care system should be able to update existing clinical records on the SHR while keeping the version history
The SHR should acknowledge requests from a client system and provide appropriate information in the event of errors
Secondary
Data should be available for extraction for secondary use.
Workflows
The Shared Health Record consists of a subset of workflows within the OpenHIE Framework, specifically:
Save patient-level clinical data workflow – V2.0
Query patient-level data workflow – v2.0
Query for Aggregate data from the SHR
Aggregate data exchange from patient level system to HMIS
Required FHIR Module Resources
US Core Profile:
(Resources are currently Implemented in FHIR, not necessarily the entire profile)
US Core DiagnosticReport Profile for Laboratory Results Reporting
US Core DiagnosticReport Profile for Report and Note exchange
--------------------------------------------
(Resources not currently implemented)
In addition US Core uses the Vital Signs Profile from the FHIR Specification.
--------------------------
OpenHIE SHR recommended dataset:
Items related to a patient daily care, such as:
Clinical Observations
Care summaries
Allergies
Medications that have been prescribed to the patient
Clinical Notes (Referral/provider) (not coordination of referrals)
Medical histories (we are implementing an SHR to store the history)
Quality of life indicators (these are observations)
Textual Care/Action Plans (for Asthma, Diabetes, PMTCT etc.)
Nutritional / mental health assessments (these are observations)
Problem Lists / Diagnosis / Health conditions
Radiology impression / report (not entire image, may be a pointer to the image)
Reportable items - make sure we collect enough for the next level up
Record a clinically relevant event to allow us to record details that something has happened
Lab report (without full lab sample details)
Immunizations
Required FHIR Module Functionality
Store a useful subset of clinical data
Send a bundled response with the requested/required resources
Retrieve relevant data as needed
Create/Update functionality
Maintaining Resource Identity
Support for Unstructured Document Transfer (XDS.b? CDA?)
US Core DocumentReference Profile
Update the SHR with relevant data as needed
Subscriptions?
Custom OpenMRS Scheduled Task ?
updatedSince
elements parameter support (to keep bundle size down)
Feed into reporting/analytics pipelines
Async FHIR Support
Pipeline support?
Coding conversions (terminology)
Version support
Deliverables
Specifications for an SHR in Haiti
IG and profiles
Technical specs
Selection of SHR tools (HAPI JPA server)
Contextualization to SEDISH architecture (OpenHIM etc., likely OpenCR, mediator)
Semantic decisions
architecture decisions
tool selection decisions
MVP demo
Security
Showcase a usecase in Continueum of care
Create patient in one instance (local → centralized)
Search for and get SHR of patient in another instance (centralized –> Local)
FHIR bundle stored in a complex obs and rendered, possibly using https://www.hl7.org/fhir/domainresource-definitions.html#DomainResource.text
Security Strategy
Evaluation Strategies
https://wiki.ohie.org/display/documents/Workflow+Collection+Maturity
The system should be validated against the health needs of low resource settings, e.g. HIV, TB, Maternal Care.
Relevant FHIR Docs
Relevant Talk Posts
Relevant Github Links
Open Questions
Where does the responsibility for determining patient identify / de-duplication lie? Is it with the client system, or with the SHR?
Is the second target usecase the MPI?