Allergen Changeover on a Shared Cookie Line
A bakery produces ready-to-eat cookies on one oven line. It runs milk chocolate cookies, then changes over to dark chocolate cookies that declare 'may contain milk.' The changeover procedure is a 30-minute dry clean: scrape, vacuum, and wipe shared mixer paddles and conveyor. Allergen swabs using an ELISA with a limit of quantification of 2.5 ppm gave results of 3.0, 4.2, 2.8, and 3.6 ppm on the mixer paddle after cleaning across four changeovers. The validation record cites only 'visual clean' and a supervisor sign-off. The operator says dry cleaning is sufficient because the dark chocolate formula contains no milk. The plant makes 8,000 packs per day and has 4 lots on hold from the changeover window. The auditor must judge the allergen preventive control.
1. Evaluate the adequacy of the allergen hazard analysis and the cleaning validation for this changeover.
Read the worked answer
The hazard analysis should treat milk as a significant allergen hazard because the line runs a milk-containing product immediately before a product that does not contain milk. The preventive control is the changeover cleaning. The validation record does not validate that control: visual clean and supervisor sign-off show appearance, not allergen removal. The ELISA results above the 2.5 ppm LOQ after cleaning on four occasions show that milk residue remains on the mixer paddle. The LOQ is an analytical limit, not a health-based action level, so the plant must establish a scientifically justified action level for the surface and use it. The swab method also needs recovery and efficiency data for the surface and allergen. The operator's statement that dry cleaning is sufficient because the dark formula contains no milk misunderstands cross-contact. A 'may contain milk' label does not replace an effective preventive control if milk can be avoided. I would conclude the allergen preventive control is not validated and is not operating effectively.
2. What disposition would you recommend for the four held lots, and what finished-product evidence would you require?
Read the worked answer
I would keep the four lots on hold and not release them on the label alone. First, confirm the label and customer and regulatory requirements: a 'may contain milk' statement is a precautionary allergen statement, not a guarantee that the product is safe for all milk-allergic consumers. Then sample representative finished product from each lot and test with a validated milk ELISA; compare results with a documented action level based on risk assessment, not simply the LOQ. Review the production sequence, swab results, and whether the affected product is intended for milk-allergic consumers. If results exceed the action level, rework, reject, or divert the product; any relabel must be truthful and compliant. If results are below the action level, document the rationale and release decision. Also verify traceability so the four lots can be located if a recall becomes necessary.
3. Propose a corrective action and verification plan that addresses root cause rather than retraining alone.
Read the worked answer
The root cause is likely the changeover design and lack of validation, not operator knowledge. Corrective actions should include disassembly and cleaning of dead spots, dedicated tools or utensils where practical, a validated wet-clean or flushing step if dry cleaning cannot achieve the action level, allergen swabs at worst-case sites with defined action levels, and a hold-and-release step until swab results pass. The plant should update the hazard analysis, SSOP, validation protocol, and changeover records. Verification should include three consecutive changeovers with swabs and finished-product testing, trend review, and an independent QA check. Training can be included, but only after the procedure and verification are fixed. A 30-day follow-up should confirm the revised changeover and the monitoring results.