Food Testing, Analysis and Laboratory Quality Assurance
The legal sample chain
Food testing begins with a defensible sample. The Food Safety Officer takes a representative sample, divides it into the prescribed parts, seals and marks each part, and dispatches one part to the notified laboratory; chain-of-custody records, temperature and preservative choices must survive scrutiny. The food analyst examines the sample and issues a report; a dissatisfied FBO may seek referral analysis under the Act. An accredited laboratory works to ISO/IEC 17025:2017, with NABL scope, validated methods, calibration traceability, QC samples and documented uncertainty. A result is only as good as the sample and the method.
Quality assurance essentials
Method validation covers accuracy, precision, selectivity, linearity, LOD, LOQ, robustness and recovery. Routine QC uses blanks, duplicates, spiked samples, certified reference materials, control charts and proficiency testing. Measurement uncertainty must be considered when a result is compared with a legal limit; a result at the limit is not automatically a violation. Microbiological methods use specified protocols, such as ISO methods for Salmonella, Listeria, E. coli and aerobic plate count, and report units correctly.
Worked example
A lab spikes milk with aflatoxin M1 at 10 µg/kg and finds 9.2 µg/kg. Recovery is 9.2 ÷ 10 × 100 = 92%, inside a typical 80–110% acceptance window for trace analysis. If recovery were 55%, the analyst should suspect extraction loss or matrix suppression, check calibration and internal standard, and not report a defensible result until the method is corrected.
Traps
Never report a value below LOQ as zero. Do not ignore measurement uncertainty at the legal limit. Do not compare CFU/g with MPN/g. Do not accept a sample without seals, codes and chain-of-custody. Do not use a method outside the lab's accredited scope for a statutory sample.