This is a ready-to-use qualification protocol. Replace every <<FILL: ...>> placeholder with your own specifics and route it through your normal review and approval. A worked scored specimen follows the template. It is an educational aid to adapt and verify against USP <1790> and your own program, not a substitute for them.
Approval page
| Field | Entry |
|---|---|
| Protocol title | Visual Inspector Qualification (Probability-of-Detection Challenge Set) |
| Protocol number | <<FILL: protocol ID>> |
| Version | <<FILL: version>> |
| Product / container in scope | <<FILL: product and container format>> |
| Author | <<FILL: name, role>> |
| QA approver | <<FILL: name, role>> |
1. Objective
Qualify a named inspector to perform 100 percent visual inspection of the product-and-container combination in scope, by demonstrating documented visual capability and a controlled detection performance on a blinded challenge set that spans the gray zone, where probability of detection is partial and where the challenge actually discriminates.
2. Scope
This protocol qualifies manual inspection for the specified product and container format. A separate qualification applies for other formats, because probability of detection is a property of the product-and-container combination, not the container alone.
3. Prerequisites
- The inspector holds current near-vision acuity and color-vision screening (see section 5.1).
- The inspector has completed documented training on the current defect library
<<FILL: defect library document number>>. - The qualified challenge set and its characterization data are available and within their validity period.
- Inspection station conditions (backgrounds, illuminance, viewing time, distance) are set and verified per
<<FILL: SOP-ID for inspection>>.
4. Roles
| Role | Responsibility |
|---|---|
| Inspector (candidate) | Performs the challenge-set inspection under qualified conditions. |
| Qualification administrator | Prepares the blinded, randomized set, records results, keeps the key. |
| QC / QA inspection owner | Reviews results against acceptance criteria and grants or withholds qualified status. |
5. Procedure and test cases
5.1 Vision screening (Test case Q1)
| Item | Detail |
|---|---|
| Step | Confirm current near-vision acuity (corrected acceptable) and color-vision screen (for example an Ishihara test) |
| Expected | Both current and passing, per <<FILL: occupational-health criterion>> |
| Frequency | On qualification and at least <<FILL: annually>> |
| Actual / Pass-Fail | <<FILL>> |
5.2 Defect-library discrimination (Test case Q2)
| Item | Detail |
|---|---|
| Step | Inspector classifies a set of known reference units against the defect library |
| Expected | Correct classification of critical defects; consistent naming |
| Actual / Pass-Fail | <<FILL>> |
5.3 Blinded challenge set (Test case Q3)
Build a blinded, randomized challenge set that concentrates on the gray zone. A common construction for a 100-unit set:
| Category | Units | Purpose |
|---|---|---|
| Reject-zone particulate (high POD) | <<FILL: e.g. 30>> | Always-reject units; anyone should find these |
| Gray-zone particulate (intermediate POD) | <<FILL: e.g. 20>> | The discriminating units; results trended |
| Container / cosmetic critical | <<FILL: e.g. 20>> | Non-particulate critical defects |
| Good units (no defect) | <<FILL: e.g. 30>> | Measures false rejects |
The inspector inspects the full set under qualified conditions, in randomized blinded order, with no feedback until the run is complete.
6. Acceptance criteria
Set the criteria on two axes, effectiveness (finding defects) and efficiency (not rejecting good product), because a program that measures only detection trains inspectors to reject everything, which corrupts the AQL data.
| Metric | Acceptance criterion (set yours) |
|---|---|
| Reject-zone particulate detection | <<FILL: e.g. 100% found>> |
| Container / cosmetic critical detection | <<FILL: e.g. >= 90% found>> |
| Gray-zone particulate detection | Trended, not pass/fail; flag an inspector far outside the group POD curve |
| False reject rate on good units | <<FILL: e.g. <= 10%>> |
The inspector is qualified when the reject-zone and critical detection criteria are met and the false-reject rate is at or below the limit. Gray-zone performance is recorded and trended against the group.
7. Re-qualification
- Re-qualify at a defined frequency (annual is common), and out of cycle on triggers: a failed vision screen, a run of AQL failures traced to inspection, a long absence from inspection duty, or a change in product or container.
- Enforce fatigue controls (a defined maximum continuous inspection time with breaks), because detection degrades with fatigue.
8. Deviation handling
Any departure from this protocol during execution is recorded as a protocol deviation, assessed for impact on the qualification decision, and dispositioned before qualified status is granted.
9. Summary and conclusion
Record the result, the qualification decision, and the basis. Qualified status names the product-and-container combination and the effective period.
| Field | Entry |
|---|---|
| Overall result | Qualified / Not qualified |
| Scope of qualification | <<FILL: product and container>> |
| Valid until / next re-qualification | <<FILL: date>> |
| Approved by (name, date) | <<FILL>> |
10. References
USP General Chapter <1790>, Visual Inspection of Injections (probability of detection, qualification concepts, inspection conditions). USP General Chapter <790>, Visible Particulates in Injections. EU GMP Annex 1 (2022), on qualification and periodic re-qualification of inspectors and challenge of inspection systems.
Confirm the current version and clause numbers of each reference before issue.
Revision history
| Version | Date | Author | Summary of change |
|---|---|---|---|
<<FILL: 1.0>> | <<FILL: date>> | <<FILL: author>> | Initial issue. |
Filled specimen
A candidate inspector runs a 100-unit blinded set (30 reject-zone, 20 gray-zone, 20 container/cosmetic critical, 30 good). Scored result:
| Category | Units | Found | Criterion | Result |
|---|---|---|---|---|
| Reject-zone particulate | 30 | 30 | 100% | Pass |
| Gray-zone particulate | 20 | 14 | Trended | 70%, recorded |
| Container / cosmetic critical | 20 | 19 | >= 90% | Pass |
| Good units (false reject) | 30 | 2 over-rejected | <= 3 (10%) | Pass |
Decision: qualified for the 10 mL clear vial, monoclonal antibody product, valid 12 months. The gray-zone 70 percent is recorded and compared to the group POD curve; it is within the group range, so it is informational, not a fail. Had the inspector missed a reject-zone unit, or over-rejected five good units, the qualification would not have been granted.
Common inspection findings this protocol prevents
- Qualification against only easy, obvious defects, proving nothing about gray-zone performance.
- A single qualification reused across container formats without re-checking probability of detection.
- Measuring detection only, so inspectors over-reject and corrupt the AQL data.
- Lapsed vision or color-vision screening, undermining every inspection the person performed since.
How to adapt this protocol
- Size the challenge set and set the category counts and acceptance criteria for your products and their gray-zone size ranges.
- Build the challenge kit with characterized units and control its validity and chain of custody through the defect-standard program.
- Set your re-qualification frequency and triggers, and define fatigue limits in the parent inspection SOP.
- Confirm each reference in section 10 against its current published version before issue.