Zinata News

A New & Improved Testing Laboratory is Coming Your Way

Communicating Your Testing Needs

Part 2

Do you rely on test results from analytical testing laboratories? Do you release products based on these results? In manufacturing, do you use test results to monitor and control your production process? If so, how do you communicate your testing needs to the laboratory?

A new and effective way, to both accomplish and answer these questions, is the use of Decision Rules. These are included in the new international standard, ISO/IEC 17025:2017, General Requirements for the Competence of Calibration and Testing Laboratories. The standard ensures the needs and requirements of the laboratory test results are clearly understood by all. The new systems approach enables effective communication and understanding of testing needs. Part One of this blog series outlined two of the improved systems and tools being implemented in analytical chemistry: Measurement Uncertainty and Decision Rules. The use and benefits of two additional tools, Target Measurement Uncertainty and Probability, are explored below. Combined, these changes result in increased clarity and confidence to both the customer and the laboratory.

Target Measurement Uncertainty: How often can the test results be wrong?

How many test results can you accept being wrong just due to random error? Of course, we all want 100% of our test results to be right, but that is not possible; there will be always be some amount of error. Because of this,we must define the acceptable amount of error.

Once determined, the acceptable amount of error is translates into a Target Measurement Uncertainty (TMU). This becomes the goal for the performance of the test method. As long as the measurement uncertainty resulting from the test method is less than the TMU, your test results are considered fit for your intended use.

Using Probability to Create a Clear Decision Rule

Optimal use of TMU relies on creating clearly defined Decision Rules. Decision Rules are the foundation to ensure test results are fit for their intended use. A vital component in the process of creating decision rules is the use of probability. Using an example, let’s walk through the steps required to create Decision Rules in order to better understand how probability fits into the mix.

Step 1

Define what is being tested. This sounds straightforward, but can be trickier than expected. For example, during manufacturing a salt solution is used in both a concentrated form and a diluted form. The description of the solution must clearly  state which form is being tested.

Step 2

State the decision being made with the test result. In this example, the production process requires the salt solution be between 5 mg/g and 10 mg/g.

Step 3

Incorporate probability. Those who will use the test result define an acceptable probability of being wrong. The people who use the test data know and understand the implications of a possible failure and can state the level of risk the company is willing to take. In our example, the impact of the salt solution being less than 5 mg/g or more than 10 mg/g is that the production process will fail and the lot will need to be discarded. There is no ability to re-work the lot. Because of this high risk, the acceptable probability of being wrong is set at 0.5 %.

Step 4

The laboratory uses the acceptable probability of being wrong to establish the maximum acceptable measurement uncertainty for the result, also known as the TMU. The laboratory can then ensure the chosen test method delivers results with less than this TMU.

For our example, production makes a salt solution with an expected concentration of 7.5 mg/g.  The laboratory tests the salt solution to confirm it is within specification, between 5 mg/g and 10 mg/g. As shown in Figure 1, a TMU of 0.9 mg/g yields the 0.50% probability of the test result being outside the specifications or limits. The decision rule can be worded as:

The concentrated salt solution will be considered acceptable if the test result is greater than 5 mg/g and less than 10 mg/g and the probability of being wrong is less than 0.5 %.

Decision Rules Facilitate Communication

Using the improved laboratory systems introduced in the new international standard, ISO/IEC 17025:2017, TMU and Probability are weighed on heavily to create clearly defined decision rules. This process ensures the test sample is clearly defined, the decision being made is understood, and the probability of being wrong is small enough for the company to accept.

The laboratory of the future uses decision rules to communicate with its customers clearly.

Contact Zinata for more information on how you can become the laboratory of the future. If you use test results from laboratories, contact us to find and work with a laboratory that incorporates the use of decision rules.


Jane Weitzel brings over 35 years of experience in analytical chemistry in highly regulated, fast paced companies. She is an internationally recognized expert in laboratory management systems and lab accreditation with extensive experience in method validation, use of measurement uncertainty, and improving lab system efficiency. She is an analytical decision maker with strong interpersonal skills. Jane was appointed to USP Expert Committee on Statistics for the 2015. With Zinata, Jane works as a Senior Quality Expert specializing in laboratory management systems and ISO/IEC 17025, an auditor, and an educator.