# ALCOA+ and data integrity in regulated environments
An FDA inspector opens a paper batch record at a sterile injectables plant and notices something odd: the handwriting for a "cleaning verification" step matches the operator who signed off three hours before their shift started. The audit trail in the electronic system tells a different story. The record was backdated. Within months, the site receives an FDA Form 483, the agency's formal list of observed violations, followed by a warning letter that halts shipments to the US market.
This is not rare. Data integrity failures are among the most common reasons the FDA cites manufacturers. You can browse real examples in the FDA Warning Letters database, which is public and searchable.
Let's unpack what went wrong, and the framework that would have caught it.
GxP is shorthand for "good practice" regulations: GMP (manufacturing), GLP (laboratory), GCP (clinical), and others. If a data point supports a decision about a drug or device that reaches a patient, it lives in a GxP system and is subject to inspection.
Data integrity means the data is complete, consistent, and accurate across its entire life, from the moment it is created to the day it is legally destroyed. Regulators do not just care whether the final result is correct. They care whether you can prove it was never altered, deleted, or fabricated along the way.
That "prove it" burden is why the ALCOA+ principles exist.
ALCOA started as an FDA acronym in the 1990s. The "+" was added later by regulators including the UK's MHRA. Together they define what trustworthy data looks like.
A: Attributable. You must know who did what, and when. A shared login on a lab instrument breaks this instantly, because you cannot tell which analyst ran the test.
L: Legible. Readable and permanent. A pencil entry that can be erased fails. So does a scanned PDF too blurry to read.
C: Contemporaneous. Recorded at the time the work happened. The backdated cleaning record above failed here. You write it down when you do it, not after lunch.
O: Original. The first capture of the data, or a certified true copy. If an instrument prints a raw result, that printout (or its electronic source) is the original, not a hand-transcribed summary.
A: Accurate. No errors, and no editing that hides the truth. If a result is corrected, the original value must still be visible.
The "+" adds four more:
A quick memory anchor:
ALCOA+ Attributable who + when
Legible can I read it
Contemporaneous written now, not later
Original the source, not a copy of a copy
Accurate no hidden edits
+ Complete nothing deleted
+ Consistent events in order
+ Enduring survives the retention period
+ Available I can find it on requestPaper controls are intuitive: signatures, ink, locked cabinets. But most GxP data is now electronic. 21 CFR Part 11 is the FDA regulation that says when electronic records and electronic signatures are considered trustworthy and equivalent to paper.
In plain terms, Part 11 requires:
Back to our 483. The inspector caught the backdating precisely because the electronic audit trail recorded the real timestamp. The paper said one thing; Part 11 metadata said another. Audit trails are often what expose integrity failures.
You can read the FDA's data integrity guidance for industry for the agency's own Q&A on these expectations.
Most integrity problems are not dramatic fraud. They are shortcuts under deadline pressure.
"Testing into compliance." An analyst gets a failing assay result, quietly reruns the sample, and reports only the pass. This violates Complete and Accurate. Regulators look specifically for orphan data: injections or runs with no matching report.
Shared logins. A busy QC lab uses one Windows account for a spectrophotometer because setting up individual accounts is a hassle. Now nothing is Attributable.
Audit trail turned off. A chromatography data system is configured so the audit trail can be disabled during a run. To the FDA, this is a red flag that data could have been manipulated invisibly.
"Draft" printing and re-running. An operator runs a sample in an unofficial mode to preview the result, then runs the "official" version only if it looks good. The preview data is real data and must be retained.
For non-technical readers, it is tempting to file this under "regulatory paperwork." It is much bigger.
Patient safety. A falsified sterility or potency record can put an unsafe product on the market. The controls exist because the stakes are human.
Business continuity. A warning letter can trigger an import alert, blocking a product from the US entirely. Remediation can cost a company far more than doing it right, and repeat findings can escalate to consent decrees (court-enforced agreements with heavy oversight).
Data as an asset. In biotech, your data is your submission. A regulatory filing to approve a drug is essentially a mountain of GxP data. If its integrity is questioned, the entire approval can be delayed.
Knowledge check
1. An operator uses a shared login on a lab instrument, so multiple analysts run tests under the same credentials. Which ALCOA+ principle is most directly violated?
2. In the opening scenario, the electronic audit trail exposed the backdated paper record. What does this illustrate about data integrity in regulated environments?
3. A quality manager argues that a data point on room temperature during storage is 'just internal' and not subject to inspection. Under GxP reasoning, when does a data point fall under regulatory scrutiny?
4. Select ALL correct answers about the ALCOA+ framework and its origins.
Select all the correct answers.
5. Select ALL correct answers describing what 'data integrity' requires in a regulated environment.
Select all the correct answers.
Technology alone does not fix this. The FDA and MHRA both stress that data integrity is a management responsibility.
Design out the temptation. If the system physically prevents deleting a run or editing a timestamp, you remove the shortcut. Good validated systems make the compliant path the easy path.
Review the audit trail, not just the result. Many sites approve a batch by reviewing final numbers. Mature quality units perform routine audit trail review to catch the deletions and re-runs that final numbers hide. Regulators increasingly expect this.
Blame-aware, not blame-first. If analysts fear punishment for a failed result, they hide it. If the culture treats a documented failure as normal science, honesty becomes the default. This is a leadership problem, not a software setting.
Data governance by design. MapMapUsing software to automate repetitive marketing tasks and campaigns, enabling personalisation at scale across channels like email, web, and social.View full definition → where each critical data point is created, who can touch it, and how it flows. This is the same data lineagedata lineageData lineage maps how data moves and transforms across systems, from origin to consumption, showing where it came from, what changed it, and where it goes.View full definition → discipline used in any serious analytics function, applied to regulated records.
Ask these questions of any system that holds decision-relevant data:
If any answer is uncomfortable, that is where your next 483 comes from.