Information-lifecycle analysis lens
Status and purpose
This page adapts one useful feature of the supplied reference document (SRC-169): asking where in an information lifecycle a safety control failed. It is a prospective analytical worksheet, not a coronial taxonomy, a new coding system, or evidence that a stage contributed to any particular death.
The canonical report extraction and nine sociotechnical codes remain authoritative for this review. Any case-level statement must still trace to the official report, then to a registered claim where explicit claim traceability is required. Defence requirements and assurance tests remain downstream analytical translations.
Eight questions
| Lifecycle question | What to examine | Existing Wiki fields and codes that may be relevant |
|---|---|---|
| 1. Creation | Was the identity, observation, result, prescription, risk or task created correctly and for the right person? | PROV, IDPOP; information origin; ORIGIN failure point |
| 2. Recording | Was it complete, timely, attributable, coded and distinguishable from a later amendment? | PROV, HF, GOV; system type; digital function; ORIGIN |
| 3. Transfer | Did it cross the intended technical, organisational, geographic or paper–digital boundary intact and on time? | INT, RES, LOOP; boundary type; sector boundary; TRANSFER |
| 4. Retrieval | Could the accountable user find and open the relevant source in the workflow being used? | VIS, INT, RES, HF; origin/destination; TRANSFER or DESTINATION |
| 5. Presentation | Was the important state salient, prioritised and explained without unsafe noise or ambiguity? | VIS, ALERT, HF; human-factors issue; DESTINATION |
| 6. Interpretation | Did the representation support an appropriate clinical or operational understanding, including uncertainty and missing data? | HF, ALERT, VIS; information issue; DESTINATION |
| 7. Action | Was responsibility assigned, acknowledged, completed, escalated and closed? | LOOP, ALERT, GOV; workflow issue; DESTINATION or MULTIPLE |
| 8. Assurance | Could the organisation detect drift, unsafe configuration, workarounds, overrides and recurrence? | GOV, PROV, RES; requested action; response status; candidate REQ- and AST- records |
The stages are questions, not mutually exclusive labels. A single event can involve several stages, and the same symptom can have different mechanisms. For example, an abnormal result may be correctly recorded but not actively transferred, or transferred but not made visible, or visible but never assigned to an accountable owner.
Incident-review worksheet
Use this order when applying the lens to a new report, incident or assurance scenario:
- Identify the primary source and preserve the report's own causal language.
- Build the event sequence without filling gaps from hindsight.
- Mark where the safety-relevant information or work item originated and where action was expected.
- Ask the eight lifecycle questions and record supporting source text separately from reviewer interpretation.
- Assign existing mechanism and pathway fields only where the source supports them.
- Identify counter-evidence, respondent disagreement and alternative explanations.
- Translate a recurring mechanism into a candidate requirement only after the report-level evidence is stable.
- Design an assurance test with an observable result, evidence capture and explicit pass criterion; do not describe an unexecuted test as assurance.
Minimum provenance record
An analysis produced with this lens should retain:
- source ID, official report reference and URL;
- report date and retrieval/recheck date;
- exact statement type: coroner concern, circumstance, recipient response, reviewer extraction, cross-case synthesis or Defence translation;
- supported lifecycle stage or stages and existing mechanism codes;
- causal qualification, uncertainty and known counter-position;
- downstream claim, requirement and test IDs, if created.
This preserves the project authority chain:
primary source → report → claim → mechanism → candidate requirement → assurance test
The lens helps reviewers ask better questions. It does not promote an answer into evidence.