ADVANCEMENT OF THE PRACTICE
THE PRACTITIONER’S TOOL KIT
James J. Balsamo, Jr., MS, MPH, MHA, RS, CP-FS, CSP, CHMM, DEAAS Nancy Pees Coleman, MPH, PhD, RPS, RPES, DAAS Gary P. Noonan, CAPT (Retired), MPA, RS/REHS, DEAAS
Practical Field Sampling Strategies
Robert W. Powitz, MPH, PhD, RS, CP-FS, DABFET, DLAAS Vincent J. Radke, MPH, RS, CP-FS, CPH, DLAAS Charles D. Treser, MPH, DEAAS
Editor’s Note: The National Environmental Health Association (NEHA) strives to provide relevant and useful information for environmental health practitioners. In a recent membership survey, we heard your request for information in the Journal that is more applicable to your daily work. We listened and are pleased to feature this column from a cadre of environmental health luminaries with over 300 years of combined experience in the environmental health field. This group will share their tricks of the trade to help you create a tool kit of resources for your daily work. The conclusions of this column are those of the authors and do not necessarily represent the ocial position of NEHA, nor does it imply endorsement of any products, services, or resources mentioned.
surements (remember, there is a thermocou- ple response time) diagonally across the pan and averaging the temperature readings, you are taking a systematic probability sample. This sampling strategy is defensible whereas the single, judgmental nonprobability sample is not. What you are doing by sampling in this manner is defining a gradient through a repeatable grid pattern. As an environmental health practitioner, it is important for you to understand the various types of sampling strategies available for use and to have a sense of where each type should be used. To begin, the best strategy is to prepare a sampling plan as a standard operating proce- dure that can be referenced in field notes and is easily used by more than one person. Orga- nizations and agencies, such as health depart- ments, should have a set of standard operating procedures that cover field sampling, particu- larly when used in routine inspections. Sampling plans that are similar to system- atic sampling include stratified and cluster sampling. These variations are used depend- ing on the size, configuration, and conve- nience of the things that are to be sampled but follow the same general pattern and have the same bias. Regardless of the plan used, always try to take five or more samples with each run. In this way, you are introducing sta- tistical relevance. For those of us in institutional practice or who work with commercial food preparation facilities such as dairies, bakeries, or the inner workings of any food manufacturing, the luxury of time and consistency in these opera- tions allows us to use a random sampling plan. The Military Standard for Sampling Procedures and Tables for Inspection by Attributes (MIL-
A s environmental health practitioners and sanitarians who have been there and done that, we can honestly say that of all the mistakes we made in our ca- reers, poor field sampling techniques, bias, and misinterpretations were probably the worst—and the most embarrassing. So bear with us while we share a bit of insight into this thorny topic. There are two definitions of sampling: 1.The first definition is the act, process, or technique of selecting a suitable sample. Specifically, selecting a representative part of something for the purpose of finding parameters or characteristics of the whole. Taking a single temperature reading of a pan of lasagna in no way represents the tempera- ture of the whole pan. We know that we can make anything pass or fail, depending on where we stick the thermometer. 2.The second definition goes to the heart of what we do: a small part selected as a sam- ple for inspection or analysis. Since sam- pling is a part of inspection, and inspection is checking or testing something against established standards, we are compelled
to do it correctly and defend what we do when we sample. The goal of sampling is to define objective measurements and refine subjective observations without bias. In other words, making sense of statements like “clean to sight and touch,” ensuring that temperature-sensitive foods are not held in the temperature danger zone for any longer than necessary, or measuring physical parameters for sanitation or safety such as ventilation, adequate lighting, and slip resistance, just to name a few. Always keep in mind the cardinal rule of sampling: Garbage in equals garbage out. Samples that are not representative of the source are of little use. Furthermore, poor collection procedures can yield unrepresen- tative samples, contribute to the uncertainty of the analytical results, or worse yet, result in contamination of the samples. Errors can be calculated and are easy to interpret if our sampling strategy focuses on probability. In other words, systematic or ran- dom sampling has the least bias. For exam- ple, in sampling that lasagna pan we referred to earlier, by taking several temperature mea-
40
$3091* • 91'*6
Powered by FlippingBook