SOP: Sampling Plan for Prepackaged Foods — Codex AQL 6.5 Procedure

GIFSQ — Global Institute for Food Safety & Quality

Principle

The FAO/WHO Codex Alimentarius sampling plan covers all prepackaged fruit and vegetable products — canned and frozen fruits and vegetables, juices, jams, jellies and marmalades, tomato products — at an AQL of 6.5. A lot or production run with 6.5% defective units is accepted approximately 95% of the time. “Defective” means a sample unit that fails specified Codex requirements (total demerit points, individual defect tolerances). The plans cover lots representing substantial portions of factory production or relatively large blocks of merchandise; for small lots the required sample-size ratio is high, and when a production lot is divided into small segments, defective product is unlikely to be divided uniformly between and within the smaller lots. The plan does NOT cover factors presenting health hazards or otherwise highly objectionable to the consumer — metal contaminants, pesticide residues, rot or insect fragments, microbial spoilage — which call for an accept/reject decision by the controlling authority.

Procedure

From the lot size N (number of individual containers or units) and the container size, find the sample size n in the Codex plan table for inspection level I (normal sampling) or level II (disputes — Codex referee purposes — enforcement, or need for a better lot estimate)

↓

Draw the required number of sample units at random, giving proper consideration to code or other identifying marks in selecting the sample

↓

Examine the sample units against the requirements of the Codex Standard

↓

Accept the lot if the number of defectives is equal to or lower than the acceptance number C in the table

Worked example: a lot of 1,250 cases, each containing 24 A2½ cans of pineapple slices in syrup — N = 1,250 × 24 = 30,000 units; container size 0.85 kg; inspection level I. (Note: the can size prints as “A 2%” in the scan; carried as A2½, the standard can nomenclature.) The plan table gives sample size n = 21 and acceptance number C = 3: if the defective sample units are not more than 3 out of 21 cans, the lot is considered acceptable. If the finding is disputed and referee examination is required, take at least 29 containers at inspection level II: 4 or fewer defective cans means the lot is accepted, more than 4 means it is rejected. Sample size need not be restricted to the table minimum — a more reliable estimate comes from taking 48 or 84 samples, with acceptance numbers 6 and 9 respectively.

On the OC curves: all curves, irrespective of sample size, intersect at 6.5% defectives on the x-axis and approximately 95% probability of acceptance on the y-axis; curves H to N grow steeper as sample size rises from 13 to 126. A lot with 10% defectives is accepted 88% of the time at sample size 6 (curve E) and 65% at sample size 84 (curve M) — sample-size ratio 1:14 against an acceptance-probability ratio of 1:0.74. At 20% defectives, sample sizes 6 and 48 give 65% and 22% acceptance; at 30% defectives, sample sizes 6, 21 and 84 give 42%, 8% and 0% acceptance.

Note: the book’s plan table (Table 34-9) is mirror-garbled in the scan and is not reproduced here; use the printed Codex Alimentarius plan table (AQL 6.5) with the worked example above.

Precautions

• Health-hazard factors are outside this plan — their acceptance or rejection belongs to the controlling authority.
• Use inspection level II for disputes, referee examination, enforcement, or when a better lot estimate is needed.
• Respect code and identifying marks when drawing the sample; larger-than-minimum samples are permitted and give a more reliable estimate.

References

Ranganna, S., Handbook of Analysis and Quality Control for Fruit and Vegetable Products, 2nd edition.