There is general consensus that the execution of an analytical method is, in fact, a process. As with all processes, the quality of the output (in this case a reportable value) depends on various controllable and noise factors. Over the last few years, there has been increasing regulatory emphasis on QbD for manufacturing processes (e.g., FDA¹s 21st century initiatives, FDA¹s draft guidance on process validation guidance, and ICH Q8-9). These initiatives focus on scientific knowledge building, risk assessment, good design, demonstration of performance, and life-cycle manufacturing process management. It is inevitable that these will increasingly become expectations for analytical method processes as well.
While it is important to focus on regulatory expectations, there is another good reason to take a 21st century approach to analytical methods. We, as professionals who support the health care industry have a responsibility to the public to assure the best possible healthcare systems. They should be efficient, safe and effective. In the case of analytical methods that comes down to reliable reportable values with known uncertainty. These reportable values are used to make key decisions that affect the health and safety of the public. To make those decisions optimally requires that the uncertainty in the reported value to be understood. A reportable result whose value is in the middle of the acceptable range is desirable. Even more desirable is that the uncertainty associated with that value be a small fraction of the acceptable range. Otherwise, the reportable value contributes little to the decision and, at worst, may lead to bad decisions.
So the message here is to take pride in good analytical methods. They are the foundation of knowledge building and the basis on which our pharmaceutical and in vitro diagnostics systems are based. Develop them with powerful scientific tools such as designed experiments. Incorporate prior knowledge for optimal decision making. Validate/Verify/Transfer them using statistical equivalence tests that provide genuine evidence of quality. Monitor them using control charts to keep them in control. Take the long view of analytical methods and continue to build a strong knowledge base for them; to learn from history and improve their quality at every opportunity.
This article was originally published in IVT - Read original article