STD-105E, https://bit.ly/3DNQzNJ) is both easy to use and easy to understand. The random sampling plan, while it requires the most samples, is also the least biased and most accurate. It is an important tool when evaluating, auditing, or inspecting multiple suites, rooms, and living quarters. And we use it regularly to verify hazard analy- sis critical control point (HACCP) plans and evaluate sanitation, safety, and maintenance activities in correctional facilities, military bases, hospitals, schools, and hospitality ven- ues. A simple random sampling plan works extremely well in institutional food service where there is a stable population that is served by a fixed or 5-week rotational menu, along with standard sized hotel pans and chafing dishes in kitchen preparation and serving operations. By its very nature, nonprobability sampling has the most bias and is the most vulnerable for questioning and contradiction. Nonprob- ability sampling plans include judgmental, snowball, and convenience sampling: • Judgmental sampling should be reserved for verifying two conditions. The first con- dition is where there is an obvious prob- lem. These problems can include when o temperatures are encountered, obvi- ous spoilage and contamination such as mouse dropping in food are found, poor or absent ventilation is noted, no lighting is present, and obvious inadequate sanitation is noted, just to name a few. The second condition is when the sample is homoge- neous. For instance, sampling potable tap water, recreational waters, soups and other pumpable foods in larger preparation and serving containers, or where a single sam- ple (if properly documented) is acceptable. We advise caution. •Snowball sampling (i.e., a nonrandom sampling technique) is used to identify problems to trace the possibility of organ- isms such as Campylobacter , Listeria , Sal- monella , Vibrio , or E. coli in food produc- tion from raw to finished product or to identify misuse or overuse of disinfectants in living environments. Snowball sampling relies on your professional judgment to determine where a problem might exist and tracing it throughout its path to find a practical solution for remediation. The more information you provide in describ- ing your eorts with this sampling proce-
dure, the more cost-eective and cost-e- cient are the corrective actions. • Convenience sampling is used when we are interested in getting an inexpensive approx- imation of the truth. As the name implies, the samples are selected because they are convenient. This strategy is quite acceptable as a screening tool. It is used to get a gross estimate of the results and to design a more comprehensive sampling scheme. Unfor- tunately, it is often erroneously used as the final arbiter in regulatory inspections. There are two other factors we need to con- sider in selecting a sampling scheme: 1.The first factor is repeatability , which is the ability of the measurement system to provide consistent readings when used by a single inspector at a given location. It requires the following conditions to be in place: the same location, the same mea- surement procedure, the same observer, and the same measuring instrument all used under the same conditions. 2. The second factor is reproducibility , which is the ability for multiple environmental health practitioners to achieve consistent results. Reproducibility refers to the degree of agreement between the results of inspec- tions (including re-inspections) conducted by dierent individuals, at dierent loca- tions, and with dierent but similar instru- ments. Simply put, it measures our ability to replicate the findings of others. In either case, the sampling strategy needs to be concisely documented. For example, “The lasagna pan in the kitchen bain-marie was sys- tematically sampled on the diagonal, taking five readings, using the validated needle K-probe on the thermocouple. Sampling was completed at 1320 hours.” This type of documentation makes it both repeatable and reproducible. A good sampling strategy also helps us define observations. So much of what we do is subjective, where the information or obser- vation is ill-suited and based on opinion, interpretations, points of view, emotions, and judgment. Sampling, whenever possible and practical, gives us information that is fact- based, measurable, and observable. Objective data are usually suitable for decision-making and are less likely to be disputed or chal- lenged on an inspection report. By present- ing objective measurements, accurate data are collected, presented, and compared on repeated inspections.
The best sampling strategy and the most careful sampling technique are worthless if the field documentation of the strategy and the sampling is not carefully done. Recording the technique, times, data, and conditions are necessary to get an accurate interpretation of those data and to allow reproducible results. All data should be recorded in a bound field sampling book with numbered pages (water- proof is best). Included in the discussion of this topic is where are the best sampling locations? The selection of the sampling location is as much an art form as it is a science. This question is best answered using common sense and a good knowledge of the mecha- nisms of cross-contamination. What set us on a course of forensics was watching an individual “aseptically” sample food while touching everything but the sample. The act of sampling should be carefully choreo- graphed so not to contaminate other foods, critical surfaces, or put yourself at risk of injury. While we are considering location, we must also consider sample sequence, particularly if we are to validate a HACCP plan. The sampling locations should always be selected in sequential order from pro- cessed to raw, cleaned to soiled, and sani- tized to contaminated. Finally, the interpretation of sampling results brings us back to the sampling objec- tive. In presenting the collected data, con- sider five data quality indicators: 1. Precision 2. Bias 3. Representativeness
4. Completeness 5. Comparability
Each of these indicators is sensitive to the way sampling is done, and each is a reflec- tion on the thoroughness of the answers we provide to the questions posed by the sam- pling objective. The variability of monitor- ing data should also be interpreted to reflect consideration of the possible sources of sampling error. These errors include sam- pling design, sampling implementation, and data analysis. This consideration is particu- larly significant when decisions are made that result from regulatory inspections. Now you know. Make your next inspection a stel- lar professional model.
Contact: toolkit@sanitarian.com.
41
April 2023 • 3962&0 3+ 2:.6321*28&0 *&08-
Powered by FlippingBook