Sid Talks Testing: Producing Valid Results
Quality measurements are not made by accident. They are not the result of a single action, occurrence, or event. They are a collection of activities that are planned, interrelated and cohesive; they should be considered alongside the development of manufacturing systems. These activities collectively are referred to as Measurement Assurance. Measurement assurance is good for product quality and good for business, and is even legislated, as it has been in the Dietary Supplement Industry, since 2007. Since then, this fact has been creating more confusion and/or raising more issues for many companies that are just learning all the details that it takes to produce valid results for the materials they need to have tested.
A comprehensive Measurement Assurance program designed to mitigate risk and ensure quality has many components. It starts with the product ingredients and identification of the required measurements needed to prove ‘that what is on the label is actually in the bottle’. While there are many components, there are some that play a more important role than others. Obviously, the equipment is important and must be suitable for the measurement tasks and it also must be capable of producing valid results. Even good equipment can produce in-valid results if not handled, maintained, used and/or stored properly.
Ideally it is the process or product specifications that should drive the test and measurement process. Once it is determined what measurements are needed, we have to select a measurement process that has appropriate capability, and is practical and affordable from a business perspective or, in other words, Fit-for-Purpose. There can be many different measurement techniques or methods to get the needed measurement information but without reliable or adequate reference materials or standards to be able to generate data of the highest integrity, the method or technique used will not be very useful and the data will be nugatory.
Precision is universally considered to be the mutual agreement of individual measurements about some mean/average value (not necessarily the true value), while accuracy refers to the degree of agreement of individual measurements with some true or accepted reference value of that which is being measured. Accuracy has to do with the difference between individual measurements and some reference value or material. At best, every measurement is an estimate of the true value and it is an integral part of every laboratory’s purpose; to be competent at the chosen measurement process to ensure that the confidence level in the data generated is of the highest integrity. High integrity data is not solely a function of having a qualified reference standard or material, but without it, it would be impossible to achieve accurate measurements, no matter what method is used. All other parameters that go into a measurement can vary from lab to lab, operator to operator and instrument to instrument, but will only be as good as the reference standard or material being used to make those measurements.
The most important single attribute of a measurement process is whether it can be made to run in a state of ‘statistical control’. We know that repetition of measurement, as noted above, is subject to variability; the achievement of statistical control implies that the statistical properties of this variability are uniform over time, so that it becomes meaningful to use measurements over a limited time span to predict limits of variation for both those and future measurements.
All that said, there are different procedures that apply to qualitative vs. quantitative measurements. With quantitative analyses, we must demonstrate statistical control by evaluating the measurements, to be confident that they are consistent with every repetition of the analysis. With qualitative analyses the measurements, not being quantitative, cannot be evaluated statistically and are dependent on the quality of the reference material used for comparison to minimize variability of the qualitative measurements. This means that the more representative a botanical reference material is, the more capacity it has to measure with greater accuracy the identity of a test sample, thus increasing the confidence in the botanical identity measurement process.
Now we can see how with a qualified botanical reference material combined with suitable measurement methods, such as High Performance Thin-Layer Chromatography (HPTLC) and/or Microscopy, we can analyze an unknown sample based on its unique chemical signature or fingerprint and/or its unique anatomical characteristics to make a confident measurement of the correct genus and species of the test sample.
Each month Sidney Sudberg, Alkemist Founder & CSO, helps to demystify the science behind testing by discussing a testing-related topic.
Categories: News & Events