Just Published

August 31, 2012

Krouwer JS.  Interference Testing: Why Following Standards Is Not Always the Right Thing to Do. Journal of Diabetes Science and Technology, September 2012, Volume 6, Issue 5: pages 1182–1184. The abstract is here.

Fault Trees – a neglected risk management tool

August 12, 2012

FMEA (Failure Modes Effect Analysis) is an important risk management tool to reduce the risk of assay failures that can harm patients. There are several ways to perform FMEA, in one case, one assembles a process flow of the assay and asks at each step

  1. What can go wrong
  2. How serious
  3. How probable
  4. What can be done to mitigate the problem

There can be multiple things that can go wrong for each step. This process is sometimes called a “bottoms up” process.

A fault tree starts with an adverse top level event such as “patient harmed.” The cause of this event and the causes of the causes are enumerated until all causes have been listed. This process is sometimes called a “top down” process.

FMEA and fault trees are complimentary and theoretically provide the same information but by using both techniques, it is more likely to obtain all potential risks for an assay. For example, take one of the most common and dangerous assay problems – interferences. In the above, abbreviated fault tree (click to enlarge), an interfering substance is easily thought of as a potential patient harm cause. One can then develop a list of possible interfering substances. But for a FMEA, one would have to designate an interfering substance as a failure mode for a part of the assay process. This is not as apparent. BTW in one of the CLSI standards about risk management – EP23 – neither FMEA or fault trees appear.