Form: Product Complaint Investigation Report

This is a ready-to-use investigation record for a product quality complaint. It picks up where the intake and triage record leaves off and carries the file to a documented conclusion. Replace every <<FILL: ...>> placeholder, set your own document number, version, and effective date, and route it through your normal document control, review, and approval. A worked filled specimen follows so you can see how a completed record reads.

This document is educational and is written to be adapted. Confirm every cited regulation against the current published source before you rely on it, and confirm your reporting obligations with your own regulatory affairs and pharmacovigilance functions. Completing this form does not by itself create compliance; it only structures the investigation your quality system still has to perform.

Use it with the complaint intake and triage form, the complaint handling SOP, and the reportability decision matrix.

Document control header

Field table

Every field below is part of the record. “Required” means the field must carry an entry; where a field genuinely does not apply, enter “N/A” with a reason rather than leaving it blank. A blank field is indistinguishable from a step that was never performed.

Section A: Investigation identification and assignment

Section B: Product and batch context

Section C: Alleged defect and investigation plan

Alleged defect, verbatim from intake, copied without editing: <<FILL>>

Stating what would rule the allegation out, before you start, is what stops an investigation from becoming a search for reasons to close it. Write the plan before the first test, not after the result.

Section D: Sample receipt, condition, and identity

Confirm identity before any testing. Testing a unit that turns out not to be your product, or not to be the claimed lot, produces a result that will be quoted for years and means nothing. Where the returned unit is not your product, that is a finding in its own right and may raise a counterfeit or diversion question that leaves the complaint process entirely.

Section E: Testing against the release specification

The comparison a reviewer expects is: does the returned unit meet the specification the lot was released against. Where the alleged defect is not covered by a release test, define the additional testing and justify it.

Where the returned sample passes every test, say so plainly and then continue. A pass narrows the hypothesis; it does not close the investigation. See section H.

Section F: Batch record, deviation, and change review

That last row catches a recurring failure. A deviation closed as no impact on the basis of release testing alone has not been assessed against stability-indicating or in-use attributes. When a complaint arrives months later on a lot with a closed deviation, revisit the original impact assessment rather than citing it.

Section G: Cluster and trend check on the lot

Record nil results explicitly. “No prior complaints on this lot, query run on <<FILL: date>> covering <<FILL: period>>” is evidence that the check happened. A blank row is not. Rate normalisation and threshold mechanics live in the complaint trending and threshold log.

Section H: Root cause

See root cause analysis techniques for method selection.

On “user error”. Before recording user error or patient handling as the cause, answer this in the record: is the product, its labelling, its packaging, or its instructions for use easy to get wrong? A recurring use error is usually a design, labelling, or instruction problem presented as a user problem, and it is one of the most common causes an inspector will push back on. If the same “user error” appears three times, it is a signal about your product, not about your patients.

Section I: Product impact and field action assessment

“No field action required” is a decision that has to be reasoned. Record it with the same care you would record a recall. See recall management and field actions and the recall and field action plan.

Section J: Conclusion

The “could not confirm” section

Complete this section whenever the conclusion is Unconfirmed or Not determinable. It exists because “could not confirm, no further action” is the single most common closure shortcut in complaint handling and the one inspectors look for first.

An unconfirmed complaint is always trended. A pattern of unconfirmed-but-recurring complaints on one defect type is a confirmed signal about the system, even when every individual file is honestly unconfirmed.

Section K: Reportability decision at close

This is the second reportability assessment. The first was made at intake on the facts as received, inside the regulatory clocks. If the only reportability assessment in the file is this one, the clocks were assessed too late. Use the reportability decision matrix for both.

Section L: CAPA decision, sample disposition, and approval

Instructions for completion

1. Copy the verbatim allegation into section C without editing it. Rewriting it into a category is how the detail that solves the investigation gets lost.
2. Write the plan in section C before you test anything, including what would rule the allegation out. Retro-fitting a plan to a result is visible in the record and reads badly.
3. Confirm identity before testing in section D. A result on the wrong unit is worse than no result.
4. Complete every row. Where a row does not apply, write “N/A” and the reason. Blank rows read as steps that were skipped.
5. Record nil results. “Query run on this date over this period, no other complaints found” is evidence. Silence is not.
6. Never delete or overwrite an entry. Correct errors by single line strike-through with initials, date, and reason on paper, or through the audit-trailed correction function in the electronic system. Complaint investigation records are GxP records and are subject to ALCOA+ expectations.
7. Complete the “could not confirm” section in section J whenever the conclusion is not confirmed. It is not optional and it is the section a reviewer will read first.
8. Take extensions before the target passes, not after. An extension approved after the due date is a late investigation with paperwork attached.
9. Do not let the investigation gate the reporting clocks. Section K is the second reportability assessment, not the first.
10. Where the product was made under contract, record what was requested from the contract partner, when, what was received, and whether the quality agreement timelines were met.

Retention

Complaint records under 21 CFR 211.198(b) are retained at least 1 year after the expiration date of the drug product, or 1 year after the date the complaint was received, whichever is longer; the regulation provides separately for certain over-the-counter products without expiration dating. Licensed biological products, cell and gene therapy products, combination products, and non-US markets carry their own and frequently longer obligations. Set <<FILL: retention period>> to the longest applicable across your products and markets, and confirm it with regulatory affairs. Records must be readily retrievable for inspection.

Filled specimen

The following is a completed investigation for a lyophilised biologic. The company, product, people, and numbers are illustrative. It is deliberately a case where the first two complaints on the lot were honestly closed as unconfirmed and the third confirmed the defect.

Section A: Identification and assignment

Section B: Product and batch context

Section C: Alleged defect and plan

Alleged defect, verbatim: “The powder went into solution but the solution looked hazy, sort of cloudy, not clear like the last lot. We did not give it. The patient earlier that morning had a different vial from the same box and she spiked a fever about an hour after her infusion.”

Section D: Sample receipt, condition, identity

Section E: Testing against the release specification

Section F: Batch record, deviation, and change review

Section G: Cluster and trend check

Section H: Root cause

Section I: Product impact and field action

Section J: Conclusion

The “could not confirm” section is not completed here because the conclusion is confirmed. It was relevant to complaints CMP-2026-0688 and CMP-2026-0702, both closed unconfirmed with no sample returned. Reviewing those two files against this section shows the gap: both recorded that no sample was available and closed on that basis. Neither examined a reserve sample, and neither ran the cluster check that would have shown a second complaint on the same lot. Either step would have surfaced this defect six weeks earlier. That gap is carried into the CAPA below rather than being left as a comment.

Section K: Reportability at close

Section L: CAPA, disposition, approval

Common inspection findings this form prevents

• The investigation record shows a conclusion but no plan, no hypotheses, and no evidence of what was ruled out.
• A returned sample was tested before its identity was confirmed, so the result cannot be attributed to the complained-of lot.
• “Could not confirm” is recorded as the whole investigation: no reserve sample examined, no batch record reviewed, no cluster check performed, no plausibility assessment obtained.
• No sample was returned and the file shows no evidence that one was requested or that alternatives were tried.
• The returned sample passed and the file closed, with no consideration of whether the release test could detect the alleged defect at all.
• The cluster and trend check was not performed, or was performed and left unrecorded, so prior complaints on the same lot were never linked.
• A deviation previously closed as “no impact” was cited rather than reassessed against the attribute now in question.
• Root cause is stated as “user error” or “handling” with no assessment of whether the labelling, packaging, or instructions for use invite the error, and with the same cause recurring across multiple files.
• Product impact was assessed for the complained-of lot only, with no bounding of adjacent lots sharing the suspected cause.
• Field action was not assessed at all, or “no recall required” was recorded with no reasoning.
• The investigation depth is identical for a Minor cosmetic complaint and a Critical sterility complaint, showing no risk proportionality.
• Reportability appears only at closure, so the shortest regulatory clock had already expired before anyone assessed it.
• The investigation ran past its target with no extension record, or with an extension approved after the target date had passed.
• The CAPA decision is a bare Yes or No with no reasoning in either direction.

How to adapt this form

1. Set your document number, governing SOP, version, and effective date in the header.
2. Replace the test rows in section E with your own release specifications by product family, so the investigator is choosing from real tests rather than composing them.
3. Align the review areas in section F with the records your site actually holds and name the systems they live in, so the investigator knows where to go.
4. Set the cluster query periods in section G to match the trending periods in your complaint trending and threshold log, so the two documents cannot disagree.
5. Scale the form by severity if it is too heavy for Minor complaints: define which sections are mandatory for Minor, Major, and Critical in your SOP, and record the proportionality decision in section C. Do not create a separate short form with different fields, because that fragments your trending.
6. If the form lives in an electronic system, make the “N/A with reason” entry a real option rather than allowing a blank, and confirm the audit trail captures changes to the conclusion, the severity, and the reportability rows.
7. Where product is made under contract, add the specific quality agreement clause and response timeline to section B, and make the “records requested and received” row in section F mandatory.
8. Set retention to the longest period applicable across your products and markets, confirmed with regulatory affairs.
9. Confirm every regulation cited here against the current published version before issue.