ALCOA+ explained with examples from the QC lab

Attributable — who did it, and when?

Every entry must be traceable to the person who made it. The classic failure is a shared login on an instrument PC or a generic “Analyst” account on the lab software, which makes it impossible to say who acquired, processed or deleted data. On paper, the equivalent is an unsigned entry, or initials that cannot be matched to a signature log.

• Good: individual user accounts on every system, a current signature register for paper, and electronic signatures bound to the user, the action and the time.
• Flag: “admin” accounts used for routine work, passwords written on the instrument, one analyst’s login used while they are on leave.

Legible — can it be read, now and later?

Records must be readable and understandable for their whole retention period. On paper that means permanent ink, no correction fluid and single-line corrections that leave the original visible. Electronically, it means the record can still be opened and interpreted — including the audit trail — after software upgrades.

Contemporaneous — recorded at the time

Observations must be recorded when they are made. Inspectors look for weights written on a glove or a scrap of paper then transcribed later, worksheets completed at the end of a shift, and timestamps that do not match the sequence of work. A balance printout attached to the worksheet, or a direct electronic capture, removes the opportunity for gaps.

A useful bench test: if the analyst were interrupted right now, would the record show exactly what has been done so far?

Original — the first capture, or a certified true copy

The original record is the first place the data was captured. For many lab instruments that is an electronic file with its metadata, not the printed report. MHRA guidance distinguishes static records, such as a paper printout, from dynamic records, such as a chromatogram that can be reprocessed. Keeping only the PDF of a chromatography run, and discarding the processing method and audit trail, may mean the original has been lost.

• Good: the source file is retained unaltered, with a checksum or equivalent control that would reveal any change.
• Flag: results typed into a spreadsheet from a screen; “working copies” of data on a local drive; instrument data overwritten when the next run starts.

Accurate — correct, complete and truthful

Accuracy is about more than arithmetic. It covers qualified and calibrated instruments, validated methods, verified calculations and controlled changes. Uncontrolled spreadsheets are a frequent finding: unlocked formulae, no version control, and no way to show the calculation that produced a reported result. So is manual re-integration of chromatograms without a documented, scientifically justified reason.

Any correction must keep the original value visible, show who changed it, when and why. A record where a result simply changed — with no trace of the old value — fails both Original and Accurate.

Complete — all the data, including what you would rather not report

21 CFR 211.194 requires laboratory records to include complete data from all tests performed. The best-known failure is “testing into compliance”: trial or “system suitability” injections of samples run before the official sequence, with failing results set aside. Aborted runs, invalidated results and repeat tests must all be retained and explained. If a result is excluded, the investigation that justified it must be on file.

Consistent — the sequence makes sense

Dates and times should follow the order in which the work was actually done, across paper and electronic records. Inspectors compare worksheet times with instrument audit trails; a sample “prepared” after it was injected, or a system clock that analysts can change, undermines everything else.

Enduring — it lasts the retention period

Thermal printouts fade, unmanaged USB sticks get lost and file formats become unreadable. Records must survive for their full retention period, with backups that are tested and archives that are protected against change.

Available — it can be retrieved for review

Data must be retrievable for review, investigations, product quality reviews and inspections throughout the retention period. Data that technically exists but takes days to find, or that only one person knows how to extract, will not satisfy an inspector.

Controls on their own are not enough; someone has to look. Regulators expect audit trails for GMP-relevant data to be reviewed — routinely as part of the review of each result, and periodically across systems to catch patterns.

• Result-level review: before a result is approved, a second person checks the data, calculations and relevant audit trail entries — not just the final number on the report.
• Periodic review: a risk-based look across the system for deletions, privileged or configuration changes, activity at unusual hours and changes made after approval.
• Follow-through: findings become deviations or CAPAs, with the same rigour as any other quality event.

The review is also where hybrid systems become hard. When the original is electronic but the review happens on a printout, the reviewer cannot see what was deleted, reprocessed or re-run. Reviewing in the system that holds the original data closes that gap.

1. Map your data flows: for each test, where is the first capture, where are calculations done and where is the result reported?
2. Remove shared accounts and restrict administrator rights to people independent of the testing.
3. Lock system clocks and confirm time synchronisation on instruments and servers.
4. Replace uncontrolled spreadsheets with controlled templates or system functions, and retain the source files.
5. Define what “complete data” means for each technique, including trial injections, aborted runs and reprocessing.
6. Build audit trail review into result review, then schedule periodic risk-based reviews.
7. Control blank forms: issue and reconcile numbered worksheets so missing pages are visible.

Swat iQua Suite is designed to comply with 21 CFR Part 11 and EU GMP Annex 11, following ALCOA+ data integrity principles. In the laboratory, that translates into specific controls:

• Attributable: individual accounts with multi-factor login, and electronic signatures configured per module and per site.
• Contemporaneous and Consistent: every entry is time-stamped centrally and shown in the local time of the site where the work was done.
• Original: instrument and spreadsheet results enter the record only through approved Excel/CSV import profiles or the secure REST API. The original file is kept unaltered as raw data with a SHA-256 checksum, and every value is reviewed before it counts.
• Accurate: the specification version is frozen when a sample is registered; results lock once submitted; in in-house QC testing, corrections never overwrite — a new value supersedes the old one, which is kept, with a reason and an e-signature.
• Complete: a complete audit trail records the old value, the new value, the user, the time and a mandatory reason for every change.
• Available: an auditor-facing audit trail report filters by site, module, user and date, with PDF and Excel exports — and every export is logged.

Review is built in too: results need a reviewer and an approver who are not the analyst, a second-person raw data review can raise a deviation from a finding, and periodic audit trail reviews flag deletions, privileged and configuration changes, after-hours activity and changes after approval.