Checklist: Microbiology Data Integrity Self-Inspection Checklist

Microbiology gets scrutinized hard because so much of the record is a human observation with no re-runnable raw file behind it. This checklist walks the micro lab the way an inspector does, control by control, and maps each item to the gaps that recur in 483s and warning letters for this specific area. Use it for internal self-inspection before someone external does it for you. Mark each item Pass, Fail, or N/A with a note; every Fail becomes a remediation action with an owner. Replace the <<FILL: ...>> placeholders where lab specifics belong. A filled specimen extract follows. Verify each cited regulation against the current source before you rely on it.

How to use

Walk it at the bench, not from the SOP binder. “Show me” beats “we do.” Ask an analyst to read a plate; pull a real second-person verification; open a real instrument audit trail.
Every Fail gets an owner and a target date.
Score the walk with the summary at the end so trends across analysts or shifts are visible.

Section 1: Contemporaneous recording

#	Item	Ref
1.1	Colony counts are written at the reading station at the moment of observation, not from memory later	211.194
1.2	Each plate’s actual read time is recorded individually, not a single batch time for a session	211.194
1.3	Multi-day tests (sterility, some bioburden) have a record for every scheduled observation day, not only the final read	USP <71>
1.4	No pre-filled or pre-dated forms exist anywhere in the area	211.194

Section 2: Colony counting and second-person verification

#	Item	Ref
2.1	A documented counting method exists and analysts are trained to it	USP <61>
2.2	Second-person verification occurs against the physical plate or a retained image, before disposal	211.194
2.3	Discrepant counts are resolved against physical evidence and the resolution is documented	211.194
2.4	Too-numerous-to-count results are reported per a defined convention, never as a guessed exact number	USP <61>
2.5	Any automated or imaging colony counter’s algorithm has been validated against an independent manual reference count	Annex 11
2.6	Manual overrides of an automated count carry a recorded reason and a second-person review	Part 11

Section 3: Plate retention and imaging

#	Item	Ref
3.1	Excursions, near-limit counts, sterility positives, and recovered isolates are photographed before disposal	PI 041
3.2	Retained images are named/stored so they tie unambiguously to plate, sample, date, and analyst	Part 11
3.3	The image store is access-controlled, backed up, and a restore has been proven	Annex 11
3.4	Plates are held for the defined retention window before disposal	211.194

Section 4: Incubation and reading windows

#	Item	Ref
4.1	Incubator actual temperature over the incubation period is documented by datalogger or qualified monitoring, not a setpoint label	211.68
4.2	Every plate’s read time is checked and confirmed to fall inside the defined reading window	USP <1117>
4.3	Incubation excursions (door open, power loss) are captured as deviations	211.192
4.4	Incubators are qualified and mapped for cold/hot spots	211.68

Section 5: Growth promotion and media fill

#	Item	Ref
5.1	No media lot is used before its growth promotion test result is available and passing	USP <61>/<62>
5.2	The media lot number is traceable forward to every test record it supports	211.194
5.3	Media fill filled, incubated, and rejected unit counts reconcile with no unexplained gap	Annex 1
5.4	Media fill interim reads are dated and recorded for every retained unit, not only the final read	Annex 1

Section 6: EM, bioburden, sterility, endotoxin reconciliation

#	Item	Ref
6.1	The EM sampling plan reconciles against executed records with no unexplained missed sites	PI 041
6.2	Sterility tests started reconcile against results reported, with every positive accounted for	USP <71>
6.3	A sterility or EM result invalidated as lab error carries objective, documented assignable-cause evidence	USP <71>
6.4	The BET reader’s raw electronic file is retained and reviewed, not only the transcribed value	Part 11

Section 7: Identification, chain of custody, and instrument audit trails

#	Item	Ref
7.1	Every reportable isolate carries a unique ID traceable back to its originating plate	211.194
7.2	Isolate transfers (internal or to an outside lab) are logged with date, time, and signatures on both ends	PI 041
7.3	Instrument audit trails (imaging counter, BET reader, rapid method platform) are reviewed as part of result review	Annex 11
7.4	No instrument run exists with no corresponding reported result (the orphan check)	Part 11
7.5	Any rapid or alternative microbiological method has documented comparability to its growth-based reference method	USP <1223>

Scoring summary

Section	Items	Pass	Fail	N/A	Highest-risk open gap
1 Contemporaneous recording	4	`<<FILL>>`	`<<FILL>>`	`<<FILL>>`	`<<FILL>>`
2 Counting/verification	6	`<<FILL>>`	`<<FILL>>`	`<<FILL>>`	`<<FILL>>`
3 Retention/imaging	4	`<<FILL>>`	`<<FILL>>`	`<<FILL>>`	`<<FILL>>`
4 Incubation/windows	4	`<<FILL>>`	`<<FILL>>`	`<<FILL>>`	`<<FILL>>`
5 GPT/media fill	4	`<<FILL>>`	`<<FILL>>`	`<<FILL>>`	`<<FILL>>`
6 EM/bioburden/sterility/BET	4	`<<FILL>>`	`<<FILL>>`	`<<FILL>>`	`<<FILL>>`
7 ID/custody/audit trails	5	`<<FILL>>`	`<<FILL>>`	`<<FILL>>`	`<<FILL>>`

Signoff

Role	Name	Signature	Date
Assessor	`<<FILL>>`
Lab supervisor	`<<FILL>>`
QA	`<<FILL>>`

References

21 CFR 211.68, 211.192, 211.194; 21 CFR Part 11. USP <61>, <62>, <71>, <1117>, <1223> (consult current chapters directly). EU GMP Annex 1 (contamination control) and Annex 11 (computerized systems). PIC/S PI 041, Good Practices for Data Management and Integrity.

Confirm the current version and clause numbers of each reference before issue.

Filled specimen

An extract from a walk of one QC microbiology lab, showing how Fails convert to actions.

#	Item	Result	Note / action
2.2	Second-person verification before disposal	Fail	Verifier signs from the written count sheet at end of shift, after plates are autoclaved. Action: move verification station next to plate disposal point, no plate leaves the bench until verified. Owner: Lab Supervisor. Target: 4 weeks.
3.1	Excursions imaged before disposal	Pass	Confirmed against three recent EM excursion records, all imaged same day.
5.1	No lot used before GPT result available	Fail	New TSA lot pulled into use during a stock-out, GPT result completed two days later. Action: add a physical quarantine shelf for unreleased lots. Owner: QC Micro Manager. Target: 2 weeks.
7.1	Unique isolate ID traceable to originating plate	Pass	Verified against the last five identification records; all traced cleanly.
7.3	Instrument audit trail reviewed	Fail	Imaging colony counter’s audit trail has never been opened by anyone. Action: add audit trail review to the weekly QA sampling. Owner: QA. Target: 6 weeks.

Three Fails, each with an owner and a decision, is exactly the managed posture an inspector expects to see, versus a gap nobody has named.

Common inspection findings this checklist catches early

Second-person verification that is a signature ritual rather than an independent look at the plate.
A media lot in routine use with no growth promotion result yet on file.
An instrument audit trail that exists, is enabled, and has simply never been opened.
An isolate identification result with no traceable path back to the plate it came from.

How to adapt this checklist

Add rows for site-specific test types or instruments your lab actually runs.
Walk it per analyst or per shift to see where risk concentrates, not just once for the whole lab.
Convert every Fail into an action with a named owner and a target date.
Feed results into the lab’s periodic review and the site data integrity gap assessment.
Confirm the referenced regulations and compendial chapters against their current published versions before issue.