Resources/ALCOA for quality records

Guide

ALCOA for quality records

ALCOA is a short checklist for whether a record can be trusted as evidence of what happened. It is not a personality test for the lab, and it is not a substitute for scientific review — but broken ALCOA is often what makes an inspector stop reading the story and start reading the holes.

The five letters

A — Attributable

Who did this?

Results, entries, and changes should be traceable to a person (and role where it matters) — not “the lab,” not an anonymous table, not a correction with no initials.

L — Legible

Can a second reader still read it?

Content should remain readable and its history preserved. Obliterated values, strikeouts with no replacement, or passages marked illegible in a transcription are signals — not footnotes to ignore.

C — Contemporaneous

Recorded when the work happened?

Records made long after the fact, “from memory,” or batched in a single sitting days later undermine the story. (A sign-off dated before the work is a different failure — date order — and should not be confused with late recording.)

O — Original

Is this the real record (or a controlled true copy)?

Informal retypes, “summary instead of the raw chart,” or copies with no true-copy language are where Original breaks. Silence is not automatically a finding — only what the document exposes.

A — Accurate

Complete as recorded, free of unexplained change

Amendments without reason, selective reporting the text itself admits, or derived values that contradict the stated raw value belong here — not every figure/table mismatch (those are consistency problems of their own).

What ALCOA is not

Weak on the page
Assay results table with no analyst identified.
Correction: value overwritten, no initials or date.
“Recorded from notebook later the same week.”
The document itself exposes attributable / legible / contemporaneous problems.
What “better” looks like
Analyst and date on the result set.
Correction with strike-through, new value, initials, date, reason.
Entries tied to the time of performance (or an honest late-entry procedure).
Still requires judgment that the science is sound — ALCOA only covers the integrity of the record.

Empty signature blocks are a special case. A tool can flag that a block looks empty or undated. It should not return PASS in a way a reviewer would read as “signature verified.” Presence is not authenticity.

How this shows up in inspection prep

Under deadline, people fix the narrative (root cause language, executive summary) and leave the record mechanics untouched: who entered the value, when, what was overwritten. Those mechanics are often faster to check than to re-investigate — and they are exactly what a second pair of eyes (human or assisted) can surface early.

In Dry Run core ALCOA is an overlay on every live review — one criterion per letter, scoped so the same problem is not double-counted against document-type checks. Sources are static and human-authored.

Open ALCOA criteria →  ·  Mechanical vs judgment →

See how findings are listed per criterion on sample data.