Log: APR/PQR Signal and Trend Disposition Register

This is a ready-to-use, standalone register for capturing every signal a product review’s trending surfaces, and the disposition given to it. It exists because the single most common product review finding is a trend that was seen and then quietly dropped: a register with a discrete row per signal makes that impossible to do silently. Keep one register per product per review cycle, or one running register per product that a new cycle’s rows are appended to, whichever your document control model supports. Replace every <<FILL: ...>> placeholder. A filled specimen follows. This is general guidance to adapt and verify, not legal or regulatory advice.

Why a standalone register, separate from the report

The APR/PQR report includes a signal register section, but keeping the register as its own controlled, append-only record has two advantages a report section alone does not. First, it lets you trend the signals themselves across cycles: did the same CQA get flagged three years running, each time “justified acceptable,” with the underlying capability never actually improving? That pattern is invisible if each year’s register lives only inside that year’s report. Second, it is the record an inspector will ask for directly when testing whether “no adverse trends” narratives match what was actually flagged; a discrete, structured register is far faster to produce and far harder to argue with than a search through several years of report PDFs.

Field table

The disposition decision

Every signal gets exactly one of two outcomes, decided against the same test each time: is there a documented, evidence-based technical justification that the signal reflects known, bounded variability rather than an assignable or worsening cause? If yes, the disposition is justified acceptable, recorded with the supporting rationale and no CAPA. If no, or if the evidence itself points to an assignable or worsening cause, the disposition is escalated, with a CAPA opened in the existing CAPA program and its number recorded here. There is no “monitor only, no disposition” outcome; monitoring is a valid choice within either disposition (see the “carried to next cycle” field), not a substitute for making one. See the annual product review and PQR article for the worked decision logic and the reasoning behind it.

Instructions for use

1. Log a signal the same day trending identifies it, not retroactively when the report is assembled. A register built after the report is written tends to only contain the signals the report already decided to mention.
2. Never leave the disposition field blank pending “further discussion.” If a decision genuinely cannot be made yet, record “escalated, pending investigation” and open the CAPA, rather than leaving the row open with no disposition.
3. Record nil cycles explicitly. A product with zero signals in a given cycle should show a register entry (or a header note) stating the trending was performed and no signal was surfaced, with the query reference, so silence reads as evidence rather than as an omission.
4. When the same CQA or category is flagged in consecutive cycles with a “justified acceptable” disposition each time, treat the recurrence itself as a new signal: a technical justification that explains one occurrence does not automatically explain a pattern of occurrences, and that pattern belongs in this register too.
5. Do not delete or overwrite a closed entry. Corrections follow your standard GxP record correction practice: single-line strikethrough with initials, date, and reason, or the audit-trailed correction function in the electronic system.

Retention

Set retention to at least the same period as the APR/PQR reports it supports, longer where product, market, or record-type rules require it. Confirm with your document control and regulatory affairs functions.

Filled specimen

The following shows four completed entries for an illustrative solid oral dose product across one review cycle, including one recurring signal carried from the prior year. The company, product, and numbers are illustrative.

The pattern across SIG-2026-011 and SIG-2027-015 is exactly the kind of recurrence instruction 4 above calls out: dissolution capability has sat below the internal threshold for two consecutive cycles, each individually justified. The 2027 review record notes this recurrence explicitly and commits to re-deriving the capability threshold or investigating the method’s actual variability if a third consecutive cycle repeats it, rather than issuing a third identical justification.

Common inspection findings this log prevents

• A signal visible in a chart inside the report with no corresponding disposition anywhere in the quality system.
• A register built retroactively to match the report’s narrative, rather than logged as trending was performed.
• The same signal justified as acceptable, cycle after cycle, with no one asking whether the pattern of recurrence is itself the real finding.
• No record that a nil cycle (zero signals) reflects trending that was actually performed, rather than trending that was skipped.
• An escalated signal whose CAPA reference cannot be produced on request, or whose CAPA closed without the register being updated to reflect it.

How to adapt this log

1. Set your ID convention and retention period.
2. Decide whether you run one register per product per cycle or one running register per product across cycles; the recurrence check in instruction 4 works better with a running register.
3. Link the “Data source” and “Evidence reference” fields to your actual LIMS, QMS, and stability system naming conventions.
4. If your CAPA system has its own severity or classification scheme, consider adding a column here that mirrors it so the two records stay aligned.
5. Confirm the register format satisfies your electronic records requirements (audit trail, e-signature) if maintained in a validated system rather than on paper.