Healthy Route Implementation Guide
0.1.0 - ci-build
Healthy Route Implementation Guide, published by Savannah Informatics. This guide is not an authorized publication; it is the continuous build for version 0.1.0 built by the FHIR (HL7® FHIR® Standard) CI Build. This version is based on the current content of https://github.com/savannahghi/healthy-route-ig/ and changes regularly. See the Directory of published versions
The protocol as a resource, and one participant's enrolment onto it.
It is the only place they differ in what the guide models today, and that is deliberate. It is not yet established for what it does not model — see the limit of the claim below.
The participating countries run separate protocols with separate ethics approvals, separate sponsors, separate site lists and different case and control targets. None of that is a difference of structure. A country-specific profile would force every downstream consumer to branch on country, and could not be undone once anything depended on it.
So HRResearchStudy is one profile,
and each country is one instance of it. region carries the country, site
carries that country's sites, recruitment.targetNumber carries its target, and
version carries the protocol version those numbers came from.
The protocols diverge substantially in the material still in preparation. One collects anthropometry, lifestyle history and laboratory investigations the other does not. The two specify different blood pressure procedures: one times three readings around the retinal imaging, the other takes a triplicate after a rest period and uses the mean of the last two.
Whether that is instance data or a structural difference is a question for the measurement profiles. It is stated here rather than discovered later, because the architecture rests on the answer being the former.
Both protocols are investigator-sponsored. The sponsor is a local clinician or a university department. The pharmaceutical collaborator supplies the dipsticks and the retinal cameras, funds the statistical analysis, and receives the exported data — but does not hold sponsor obligations under good clinical practice.
Recording a single "sponsor" would misstate who carries those obligations, and which party a monitor or an inspector should be addressing.
associatedParty therefore requires an explicit role on every party:
| Role | Held by |
|---|---|
sponsor-investigator |
the local clinician or institution sponsoring the protocol |
funding-source, collaborator |
the pharmaceutical partner |
irb |
the approving research ethics committee |
Where a party is a named individual rather than an institution,
associatedParty.name may carry the name without a party reference, so that a
person outside the study team does not need a Practitioner resource created for
them.
version is required at 1..1.
Protocols are amended during conduct, and an amendment can change eligibility, procedures or endpoints. Which version was in force when a participant was enrolled determines which rules their record should be judged against. That is not recoverable from an enrolment date once the amendment history is gone, and it is the first thing asked when a deviation is queried.
The same reasoning applies to the consent form version on HRConsent. They are separate versions of separate documents and both have to be recorded.
ResearchStudy.status is bound in the base resource to publication status —
draft, active, retired, unknown. That describes the maturity of the
resource, not the state of the study.
Whether recruitment is open, closed or suspended is progressStatus. The two are
easy to conflate and mean entirely different things to a reader deciding whether
a site can still enrol.
HRResearchSubject requires
consent at 1..*, where the base resource makes it optional.
Enrolment without a recorded consent is precisely the finding a monitor is looking for. A reference that may legitimately be absent cannot distinguish a participant who was never consented from one whose consent was recorded somewhere this resource does not point at.
There is one reference per permission, because permissions are recorded as separate resources. A participant who joined the study but refused image retention has two references here — one permitting, one denying.
Both protocols define screen failures as participants who consent and are then not entered into the study. That is an outcome with a count attached to it, and it is reported.
progress.subjectState carries it: screening, eligible, ineligible,
on-study, off-study, withdrawn. Recording the state and the date it was
reached is what distinguishes a screen failure from a withdrawal, and both from an
enrolment still in progress. None of those three is an empty record.
actualComparisonGroup is used; assignedComparisonGroup is not.
These are observational studies with a case group and a control group defined by previously documented kidney function. A participant's group is a fact about them, established before they were approached — not an allocation made by the study. Recording it as an assignment would misdescribe the design and invite an analysis the design does not support.
Both resources are FMM 0, the lowest maturity in FHIR. This is known and accepted rather than overlooked.
There is no alternative resource for enrolment onto a protocol. Expressing it as
a bare Encounter would lose the link between a participant, the protocol version
they were enrolled under, and the consent that permitted it — which is the entire
content of the resource. The risk of an FMM 0 resource changing is smaller than
the cost of not recording the relationship at all.