This is a ready-to-use checklist for reconciling a process historian trend, exhibit, or extracted value against what appears in the batch record, before either one is relied on for a disposition decision, a CPV report, or an investigation. Two reports of the “same” parameter that disagree are one of the most common historian-related findings, and most of the time the cause is a retrieval or aggregation setting, not data manipulation, but you have to be able to prove that rather than assert it. Replace every <<FILL: ...>> placeholder, set your document numbers, and route the checklist through your normal document control. A filled specimen follows. Verify each cited regulation against the current source before you rely on it.
Header
| Field | Entry |
|---|---|
| Checklist title | Historian-to-Batch-Record Data Reconciliation |
| Document number | <<FILL: FORM-ID>> |
| Version | <<FILL: version>> |
| Batch / lot number | <<FILL>> |
| Parameter(s) / tag(s) under reconciliation | <<FILL>> |
| Reviewer | <<FILL: name, role>> |
| Date | <<FILL>> |
Checklist
Mark each item Pass, Fail, or N/A. Any Fail requires an entry in the discrepancy log below before the reconciliation can close.
| # | Item | What it tests for | P / F / NA |
|---|---|---|---|
| 1 | The tag queried in the historian matches the equipment ID and parameter named in the batch record | Wrong tag, or a tag left pointing at a decommissioned or renumbered equipment ID | <<FILL>> |
| 2 | The time window queried matches the batch record’s start and end times for the relevant step | A window offset by a shift boundary, a time zone, or a daylight-saving change | <<FILL>> |
| 3 | The retrieval mode (raw, interpolated, sampled) used for the historian value matches what the batch record or CPV procedure specifies | An unlabeled interpolated value presented as measured | <<FILL>> |
| 4 | If aggregated, the aggregation method and interval boundaries match between the two sources being compared | Two averages computed differently and reported as if directly comparable | <<FILL>> |
| 5 | The clock source for both the historian and the batch record’s own timestamps is the same controlled reference | Timestamps that cannot actually be correlated | <<FILL>> |
| 6 | Quality flags and any data gaps in the historian window are visible, not silently filled | A sensor dropout hidden as a smooth line | <<FILL>> |
| 7 | The value in the batch record matches the historian value within the tolerance defined by the parent SOP or CPV procedure | A genuine numerical discrepancy after settings are confirmed identical | <<FILL>> |
| 8 | Where the batch record shows a trend exhibit rather than a discrete value, the exhibit is confirmed as an export from the validated historian, not a screenshot of a live trend screen | A picture of the DCS display substituted for a true copy | <<FILL>> |
Discrepancy log
Complete one row for any item marked Fail.
| # | Item | Discrepancy description | Root cause (settings difference / data integrity issue / other) | Resolution | Deviation reference (if raised) |
|---|---|---|---|---|---|
<<FILL>> | <<FILL>> | <<FILL>> | <<FILL>> | <<FILL>> | <<FILL: number or N/A>> |
Acceptance criteria
- Every checklist item is answered; no item is left blank.
- Every Fail has a discrepancy log entry with a stated root cause, not just “resolved.”
- A discrepancy attributed to a settings difference names the specific setting that differed; it is not accepted on assertion alone.
- A discrepancy that is not fully explained by settings is escalated to a deviation before the batch or report proceeds.
- The reconciliation record is signed, dated, and retained with the batch record or CPV report it supports.
Sign-off
| Role | Name | Signature | Date |
|---|---|---|---|
| Performed by | <<FILL>> | ||
| Reviewed by (QA) | <<FILL>> |
References
21 CFR 211.180-211.188 (records of manufacture and control); 21 CFR Part 11. EU GMP Annex 11 (accuracy of data, storage) and Chapter 4 (documentation). ICH Q9 (Quality Risk Management), for the risk basis of when reconciliation is warranted versus a lighter-touch check.
Confirm the current version and clause numbers of each reference before issue.
Filled specimen
A reconciliation prompted by a CPV report and a batch record disagreeing on a bioreactor pH value.
| Field | Entry |
|---|---|
| Batch / lot number | BR101-2026-0847 |
| Parameter(s) / tag(s) | BR101.PH.PV |
| Reviewer | D. Okafor, Validation |
| Date | 12 September 2026 |
| # | Item | P / F / NA |
|---|---|---|
| 1 | Tag matches equipment ID | Pass |
| 2 | Time window matches | Pass |
| 3 | Retrieval mode matches | Fail |
| 4 | Aggregation matches | Fail |
| 5 | Clock source matches | Pass |
| 6 | Quality flags visible | Pass |
| 7 | Values match within tolerance | Fail (before resolution) |
| 8 | Exhibit is a true-copy export | Pass |
| # | Discrepancy | Root cause | Resolution | Deviation |
|---|---|---|---|---|
| 3, 4, 7 | CPV report showed pH 6.82 (1-minute average); batch record review pulled pH 6.71 (raw, single point) for the same nominal time | Settings difference: CPV procedure specifies 1-minute averaged retrieval for trend purposes; batch record review specifies raw retrieval at the hold-point timestamp per the batch record SOP | Confirmed both settings are correct for their respective purposes per the governing procedures; re-ran the batch record check in raw mode against the exact hold-point timestamp and the CPV average independently against its 1-minute window; both values now consistent with their own specified method. No data integrity concern. | N/A, documented as a settings clarification, not a deviation |
The reconciliation closed without a deviation because the discrepancy fully resolved to two documented, different, and each individually correct retrieval settings applied for two different, legitimate purposes. Had item 7 still failed after items 3 and 4 were confirmed to match, this would have escalated to a deviation instead.
Common inspection findings this checklist prevents
- Two reports of the same parameter disagree with no documented investigation of why.
- A discrepancy dismissed as “just a setting” with no record of which setting or why it explains the gap.
- A trend exhibit in a batch record traced back to a screenshot rather than a validated historian export.
- Reconciliation performed only when a discrepancy is noticed by chance, with no defined trigger for when it is required.
How to adapt this checklist
- Set your document number and point the acceptance-tolerance reference to your own SOP or CPV procedure.
- Define the trigger for when this checklist is required: every CPV report, every disposition-critical trend, or risk-based on data criticality.
- If your site uses a validated reconciliation tool, reference its validation record instead of performing items 3 to 6 manually, and use this checklist as the confirmation that the tool’s output was reviewed.
- Route any discrepancy that does not resolve to a documented settings difference through your deviation process, per item 7’s acceptance criterion.