Validation evidence stages

Use these stages to prepare a review-ready validation evidence package before live submission.

StagePurposeActionPass criteriaEvidence
1. Source preflightCatch structural and mandatory-field issues before FHIR generation.Run template self-check, confirm required columns, check row counts, and review date/identifier formats.No missing required columns, no blank mandatory identifiers, and no malformed dates in candidate rows.Template self-check export, source row count, rejected-row list, and owner notes.
2. Conversion traceabilityProve each source row creates the expected FHIR resources and identifiers.Run conversion, inspect generated Bundle identifiers, resource counts, and source-to-resource trace fields.Expected row count matches expected Bundle count, and every generated Bundle can be traced to a source row.Conversion summary, Bundle JSON, source row identifier, and traceability worksheet.
3. Local PDS requirement validationCheck generated content against the local Ontario MHA PDS requirement rules maintained in the app.Run validation and classify fatal, error, warning, and information findings.No unresolved fatal/error findings remain for the production candidate extract.Validation report, issue triage log, remediation owner, and retest date.
4. FHIR structural validationConfirm FHIR JSON structure, references, resource types, and Bundle shape are valid.Run the configured FHIR validator or HAPI/CLI-backed validation where available.Bundle structure is valid and references resolve within the Bundle or documented external context.FHIR validator output, OperationOutcome diagnostics, and validator version.
5. Authorized package validationValidate against the authorized Ontario MHA PDS package when the package is available to the tenant.Confirm package path/version/checksum, run formal package validation, and retain the exact package metadata.Validation uses the approved package version and no package-level blocker remains unresolved.Package metadata, checksum/path, validator command or endpoint, and validation output.
6. Terminology validationConfirm coded fields use accepted code systems, value sets, or documented local crosswalks.Run terminology self-check, compare local codes to mapped target codes, and classify unmapped values.Required coded values are mapped or formally accepted as gaps with remediation plans.Terminology export, crosswalk file, warning list, and clinical owner decisions.
7. Timeline and clinical consistencyCatch impossible or suspicious service dates, encounter timing, referral windows, status combinations, and discharge logic.Review validation warnings for dates/statuses and compare source-system timeline rules with generated resources.No impossible timelines remain and accepted warnings have a documented rationale.Date/status issue log, corrected source extract, and accepted-warning record.
8. Submission dry runProve the interface-engine queue can process candidate payloads without live Ontario Health submission.Queue a candidate row through Live Feed in mock or dry-run mode and process the job.Job reaches accepted/mock-accepted status with traceable Bundle ID and OperationOutcome.Feed job export, attempt history, mock/dry-run response, and Bundle identifier.
9. Go/no-go evidence packBundle validation evidence for governance, privacy, security, data, clinical, and integration approval.Export mappings, terminology, validation, conversions, audit log, readiness, and transport checklist.Every readiness gate has an owner, date, status, and retained artifact.Named evidence package, sign-off record, and go/no-go decision.