Read Part I in this series to learn about the importance and impact of representative sampling.
Recent recalls of flour contaminated with salmonella and E. Coli have placed an increased focus on pathogen detection at milling operations.
Flour is considered a raw food product, since most flour products don’t undergo a heat treatment or kill process like ready-to-eat foods. Contaminated flour can make people sick if they eat under-cooked or uncooked flour-based foods, such as cookie dough or cake, muffin and bread mixes.
It’s essential to implement a sampling process that provides a statistically significant sample for pathogen testing so you can quickly and accurately detect any contamination issues.
Whether you’re testing for pathogens or physical properties, samples should be statistically significant and represent the characteristics of the entire lot. The bottom line is that you want to divide an entire lot into small manageable units, then take a number of units regularly from the process.
The samples should be taken from a place in the process where the material is well-mixed and won’t introduce bias (exclude different-sized particles or kill pathogens that might be present). The number of samples and sample mass collected should be related mathematically to the confidence level that is being targeted – for example, 95% confidence level requiring 60 samples of the lot.
7 Elements of Representative Sampling Systems
Our recent article for the International Association of Operative Millers offers several suggestions for implementing a representative sampling system in your milling operation.
First, it’s critical to determine these factors:
- Define the "what you want to know" parameters. Are you looking for moisture content, protein, fatty content, kernel size, pathogens, etc.?
- Define the lot in relation to process flow – sample mass vs. lot
- Define required precision
- Determine variability of “what you want to know” parameters (e.g., moisture content)
- Select time or mass sampling basis
- Determine top size
- Determine gross or composite increments
Once you know the answers to these questions, you can start planning your representative sampling system.
3 Steps to Implement a Representative Sampling System
Step One: Before implementing your representative sampling system, you should:
- Study the sample system empirically
- Use industry-recommended practices
- Consult sampling system and analytical equipment suppliers
- Run a risk assessment to define acceptable variance
Step Two: Select your representative sample point based on a combination of factors, including:
- Where it’s safe, convenient, and reduces costs
- Where you can employ automation
- Where there’s adequate space for the operator and/or equipment
- Where the product is manageable (e.g., after secondary grinding rather than raw feed)
- Where the process helps with homogeneity or reduces segregation
Step Three: Create a plan to address errors such as incorrect sampling errors, point selection error (PSE), and grouping and segregation error (GSE). General rules of thumb to reduce these errors include:
- Mix the material prior to sampling
- Sample where the material is in a dynamic state
- Collect samples in non-reactive sample containers
- Sample frequently enough to limit variance due to process cycles
Learn more about our food and beverage sampling solutions here.
Read John Powalisz’s full article in the latest print issue of the International Association of Operative Millers print publication.
