1. Define the complete record and its owners
List the records needed to reconstruct and interpret the result: original reader output or justified true copy, instrument and run context, mapping, preparation and concentration data, analysis configuration, all observations, exclusions, calculation results, criteria, review, report, and downstream disposition. Assign the system and process owner for each.
The record boundary follows meaning, not file count. A PDF may be the complete customer-facing analysis record while the original instrument record and downstream release decision remain separate linked parts of the laboratory record.
2. Preserve the original measurement with attribution and time
Retain the authorized reader output unchanged or create a controlled true copy. Record the source identity, instrument and run, acquisition context, import actor or service, import time, file fingerprint, selected measurement region, and reader profile before interpretation.
Human and automated actions both need attributable identities and versions. A pasted numeric matrix without instrument, source, selection, or import context may be legible while still being incomplete and difficult to verify.
| Principle | Inspectable record |
|---|---|
| Attributable | User, service, instrument, and software identity |
| Legible | Readable values, units, settings, status, and timestamps |
| Contemporaneous | Acquisition and import recorded when performed |
| Original | Authorized source output or controlled true copy |
| Accurate | Tested import with visible validation and errors |
3. Keep mapping and every derivation linked to the source well
Preserve sample identity, role, replicate, dilution, control, reference designation, initial concentration, preparation factors, units, raw response, transformed response, inclusion state, and analysis version for each well. Derived long-form data should point directly back to the physical position and source bytes.
Use tested formulas and retain their inputs, precision, and version. If a mapping or calculation changes, preserve the prior state, actor, time, reason, and effect instead of overwriting the route that produced the earlier result.
4. Review all outcomes, exclusions, and corrections
Review complete data, including failed, invalid, aborted, repeated, and excluded observations. An excluded well retains its position, raw and derived values, reason, actor, time, and effect on the result. The report discloses material exclusions without making the underlying observation disappear.
Correct records through an authorized process that preserves the original information and the reason for change. Formatting may highlight a failure, but the stored value, rule, outcome, and history—not color or layout—carry the scientific meaning.
5. Protect, retain, and retrieve the complete record
Keep identifiers, version relationships, timestamps, units, and event order consistent across systems. Store records in durable protected media for the approved period, test backup and restoration, and verify that authorized users can retrieve both human-readable and electronic forms with content and meaning intact.
Reconcile downstream transfers and archived copies. A successful export is not sufficient if linked metadata, report identity, exclusions, signatures, or the ability to interpret the result are lost.
| Principle | Lifecycle check |
|---|---|
| Complete | All data, metadata, failed work, exclusions, history, and reports remain connected |
| Consistent | Ordered events, stable identifiers, units, and version relationships agree |
| Enduring | Protected records survive for the required retention period |
| Available | Authorized retrieval and rendering preserve content and meaning |
6. Conclude the data-integrity review against named risks
For each record and transition, document the ALCOA+ expectation, implemented technical or procedural control, evidence reviewed, gap, action, and residual risk. State whether the defined record lifecycle is acceptable for its intended use rather than merely declaring that the data are “ALCOA+.”
Reassess after changes to reader formats, mappings, calculations, roles, reports, interfaces, retention, or systems of record. The approved record map shows which controls and evidence the change can affect.
Limits and where a controlled workflow helps
Provenarium can link source identity, method, performer, time, exclusions, suitability, result, software release, and report in the controlled routine analysis record. The laboratory still retains the original source-system record and owns downstream review, release, retention, filing, recovery, permissions, audit-trail review, and reconciliation.
Frequently asked questions
Is a SHA-256 fingerprint proof that data are correct?
No. A fingerprint can help detect whether bytes changed and link related records, but it does not prove the source was accurate, complete, properly attributed, or reviewed.
Is the PDF the original record?
Not necessarily. The Provenarium PDF is the complete customer-facing analysis record for a controlled routine run. The laboratory record also includes the original source-system record and the customer’s downstream filing context.
Does ALCOA+ require one specific software feature set?
No. Translate each principle into concrete record requirements and risk-based technical and procedural controls for the defined lifecycle.