Guides · Good documentation practice
ALCOA+ on paper: good documentation practice in batch records
ALCOA+ asks that GMP data be attributable, legible, contemporaneous, original, accurate, complete, consistent, enduring and available. On paper: signed entries in indelible ink, made at the time, corrected without hiding the original. Data copied from the record keeps those qualities and its link to the page.
ALCOA+ is shorthand for nine qualities regulators expect of GMP data. The first five (attributable, legible, contemporaneous, original, accurate) describe how an entry is made. The other four (complete, consistent, enduring, available) describe how the record holds up over its life. On a paper executed batch record each one has a visible form: initials and a date beside the entry, ink that cannot be erased, a correction that leaves the first value readable, an unused field struck through instead of left blank.
The same qualities apply once values leave the page. A process engineer who types fill weights into a spreadsheet for trending creates new data, and that data needs to say where it came from, who copied it and whether anyone checked it against the original. This guide takes the principles one by one, for the paper record and for the data taken from it, using what the regulators’ documents say rather than what is often repeated about them.
Who uses ALCOA, ALCOA+ and ALCOA++?
The regulators do not all use the same letters, and they say the expectation is the same either way.
- FDA uses five letters. Its 2018 data integrity Q&A defines data integrity as the completeness, consistency and accuracy of data, and says such data should be “attributable, legible, contemporaneously recorded, original or a true copy, and accurate (ALCOA)”. A footnote ties each letter to sections of 21 CFR parts 211 and 212, for example contemporaneous recording to 211.100(b) and 211.160(a).
- MHRA also uses ALCOA by choice. Its 2018 guidance (§3.10) explains that the “+” adds complete, consistent, enduring and available, and says: “There is no difference in expectations regardless of which acronym is used.” Its definition of data (§6.1) reads L as “legible and permanent” and O as “original record (or certified true copy)”.
- PIC/S PI 041-1 (§7.4) calls ALCOA key concepts of good documentation practice and adds the four further attributes as ALCOA+, citing EMA’s GCP inspection guidance in a footnote. Its table in §7.5 defines all nine for paper and electronic records alike.
- WHO TRS 1033 Annex 4 (§1.2 and its glossary) calls ALCOA+ a “commonly used acronym” and asks that data governance apply its principles (§4.4).
- EU GMP Chapter 4 uses no acronym, but its good documentation practices (§4.7 to §4.9) state the substance: legible, indelible, made at the time, alterations signed and dated with the original still readable.
- ALCOA++, which adds traceable, appears in EMA’s 2023 guideline on computerised systems and electronic data in clinical trials (§4.5). That is a GCP document. None of the GMP texts above uses ALCOA++.
None of these documents says who coined ALCOA. Industry articles tell different origin stories; if an SOP or a training deck needs a reference, cite the regulator’s text rather than the story. Whichever acronym your procedures use, the expectations behind it are the same.
What does each principle look like on the record and in the data?
The middle column is drawn from the guidance above. The right column applies the same principle to values copied from the record into a spreadsheet, LIMS or trending database; it is one reasonable reading, and your quality system decides the details.
| Principle | On the paper record | In data taken from it |
|---|---|---|
| Attributable | Each entry initialled or signed and dated by the person who did the task; checks by a second person identified; signature and initials log kept current | Each value carries who copied it and when, and who verified it and when |
| Legible | Clear handwriting in indelible ink; no ditto marks or private abbreviations; dates in the site’s defined format | The value as written: same digits, decimal separator and unit; nothing guessed |
| Contemporaneous | Written at the time of the action, on the controlled record, not on scrap paper first | The event’s date and time come from the record; the transcription date is kept separately |
| Original | The executed record, or a verified true copy, is the record | The copy points back to the original (batch, page, field); the original is not replaced by the copy |
| Accurate | Entries reflect what happened; calculations checked; critical corrections verified | Transcription verified by a second person or a validated system |
| Complete | No gaps; unused fields voided, signed and dated; all pages present | Every value in scope copied, including corrected values and excluded results, with a reason |
| Consistent | Same date format, units and rounding across the record; entries in time order | One unit per column, one date format, rounding rules stated |
| Enduring | Ink that will not fade or smudge; fading thermal printouts replaced by signed true copies | Files kept in a controlled location, protected from change, for the retention period |
| Available | Record retrievable for review, investigation and inspection for its retention period | Data and its link to the original retrievable by those who review it |
What makes a paper entry attributable?
PIC/S says handwritten entries should be made by the person who did the task (PI 041-1 §8.6, item 1). A scribe recording for someone else is for exceptional cases only, and then both people should be identified and the operator should countersign where possible. Records should be signed and dated with an identifier unique to the author, and PIC/S expects controlled, current signature and initials logs (item 4).
Inspectors are told to check that key entries are signed and dated as they happen, “not just signed at the end of the page”. EU GMP Chapter 4 §4.20 asks the batch processing record for the initials of the operator of each significant step and, where appropriate, the name of the person who checked it.
What do the guides say about ink, legibility and dates?
Chapter 4 §4.7 asks for handwritten entries “made in clear, legible, indelible way”. PIC/S §8.6 spells out the checks: ink that will not erase, smudge or fade, no pencil inked over later, no ditto marks or unknown symbols, and dates in “an unambiguous format defined for the site”. It also warns that some printouts, such as thermal paper, fade; indelible signed and dated true copies of these should be kept.
Legibility matters again when the value is copied. A 1 that could be a 7, or a comma that could be a decimal point, is a question to raise with the record’s owner, not one for the person copying to settle alone.
How should a correction look?
Chapter 4 §4.9 asks that any alteration be signed and dated and that it “permit the reading of the original information”, with the reason recorded where appropriate. PIC/S §8.7 adds the practical detail:
- cross out what is changed with a single line;
- initial and date the change;
- record the reason where appropriate, and verify it if the data is critical;
- use indelible ink; no correction fluid, no overwriting.
WHO Annex 4 (Appendix 1, example 2) says the same: an incorrect entry may be corrected if the reason is documented, the original entry stays readable and the correction is signed and dated.
A fictional example: an operator writes 54.2 in a temperature field, sees it should be 45.2, strikes 54.2 through once, writes 45.2 beside it with initials, date and “transposed digits”. The reviewer can read both values and see who changed what and why.
Practice varies. Some sites require a reason for every correction, others only for critical entries, and many use a short list of reason codes. Your procedures decide. When the value is copied, take the corrected value and keep the fact that it was corrected, so whoever checks the copy knows to look at that field closely.
When is an entry contemporaneous, and what about late entries?
Chapter 4 §4.8 asks for records “made or completed at the time each action is taken”. FDA’s Q&A (question 12) says it is not acceptable to record data on pieces of paper discarded after transcription. MHRA §5.1 asks for records to be available where the work happens so that informal recording and later transcription do not occur.
None of these documents sets out a late-entry procedure. Many sites have one: the entry is marked as late, carries the date it was actually written and a reason, and is never dated as if it were made at the time. WHO Annex 4 (Appendix 1, example 1) treats a date missing from a worksheet as a lapse to be risk-assessed, with its root cause still identified even when the date can be found elsewhere.
In copied data the event’s date is the one on the record. The date of transcription is a separate fact and belongs in its own column or in the audit trail.
What should happen to blank fields and N/A?
PIC/S §8.6 asks that unused, blank fields be voided (for example crossed out), dated and signed, and tells inspectors to check that records are complete and all pages present. The aim is that a reviewer can tell “not applicable” from “forgotten”.
Sites differ on the mark itself: a line through the field, “N/A” with initials and date, or a single line through a block of spare rows. The procedure decides. In copied data a voided field is neither a missing value nor a zero; give it a distinct marker so it does not turn into either in a chart.
Are scans of batch records true copies?
A scan becomes a true copy only when it has been verified. MHRA §6.11.2 defines a true copy as one verified, by a dated signature or by a validated process, to have the same information as the original, “including data that describe the context, content, and structure”. WHO Annex 4 uses the same definition. True copies may be kept in place of originals, MHRA says, “e.g. scan of a paper record”, if a documented system verifies and records the integrity of the copy; organisations should consider the risk of destroying originals. MHRA §6.17 allows paper data to be retained through a validated scanning process. PIC/S §7.7 describes issuing paper true copies: verify, then sign and date the copy as a true copy.
An unverified scan is still useful as a working copy. It is not the record.
Does ALCOA+ follow the data into a spreadsheet?
Yes. MHRA §6.1 defines data to include “all subsequent transformations and reports”, and MHRA §6.11.2 says: “Where manual transcriptions occur, these should be verified by a second person or validated system.” WHO Annex 4 §4.15 asks companies to design processes to avoid unnecessary transcription in the first place. PIC/S §9.10 lists the interfaces of hybrid systems that need procedures and records, including manual input of manually generated data and the automated transcription of printed data into computerised systems.
A spreadsheet of values is a copy, and usually not a true copy in the MHRA sense, since it drops the page’s layout and context. The batch record stays the original. Keep for each value enough to find it again (batch, page, field), who copied it, who verified it and when. Corrections made in the spreadsheet should be traceable too. Verifying transcribed GMP data covers the second-person check in detail, and decimal and date transcription errors the mistakes that most often slip through.
A checklist for reviewers and trainers
On the record:
- every key entry initialled and dated as it happened, by the person who did it;
- initials match the signature and initials log;
- indelible ink throughout, no pencil underneath, no correction fluid;
- corrections struck through once, original readable, initialled, dated, reason given where the procedure asks;
- late entries marked as late, with the real date and a reason;
- unused fields voided, signed and dated; all pages present;
- fading printouts replaced or supplemented by signed true copies.
In the data taken from it:
- each value traceable to its batch, page and field;
- who copied it and who verified it, with dates;
- values as written, with units and the record’s decimal separator;
- corrected and voided fields marked as such;
- the file kept where it is protected and can be found for the retention period.
Trainers can use the same table with operators: each row is something they can see on their own page before they hand it over.
Sources
- MHRA 'GXP' Data Integrity Guidance and Definitions, Revision 1 (March 2018), accessed 18 November 2026.
- PIC/S PI 041-1 Good Practices for Data Management and Integrity in Regulated GMP/GDP Environments (1 July 2021), accessed 18 November 2026.
- WHO TRS 1033, Annex 4: Guideline on data integrity (2021), accessed 18 November 2026.
- FDA, Data Integrity and Compliance With Drug CGMP: Questions and Answers, Guidance for Industry (December 2018), accessed 18 November 2026.
- EudraLex Volume 4, EU GMP Guide Part I, Chapter 4: Documentation, accessed 18 November 2026.
- EMA, Guideline on computerised systems and electronic data in clinical trials, EMA/INS/GCP/112288/2023, accessed 18 November 2026.
General information, not regulatory advice. Check the current text of each regulation; your quality system decides how it applies to you.
Related guides
- Verifying transcribed GMP data: second-person checks that hold up
What regulators expect when GMP values are copied from paper into a spreadsheet, LIMS or CPV database, and how to design a check that stands up.
- From paper batch records to CPV data, without breaking data integrity
How to move process parameters from paper executed batch records into a CPV, APR/PQR or investigation dataset that holds up to data integrity review.
- Decimal commas, units and dates: transcription errors in GMP data
The errors that creep in when values move from paper batch records to spreadsheets, why they distort CPV and investigations, and how to stop and find them.