This is the final part of our 3 part series: human error, automation, and repeatability. For the sake of this post, all efforts will be focused on a clinical chemistry laboratory.
To begin, it is important to delve into failure rates in human systems. According to Charles P. Shelton, “human operators are one of the biggest sources of errors in any complex system”. This can be attributed to a number of factors including, emotional and physical stress, as well as laboratory location. To provide more information, below is a General Human-Error Probability Data in Various Operating Conditions: Unfortunately it is difficult to determine an established across the board error rate for automation. The reason being automation is an extremely broad term and as many people have pointed out, you can automate just about anything these days. The one statistical piece of data we were able to uncover can be found in an article titled “Automation and Expert Systems in a Core Clinical Chemistry Laboratory”. The article determined that when using an Intelligent Aliquotter an approximate error rate of 1% surfaced.
So what do these have to do with repeatability? Well looking at the Wikipedia definition for repeatability as “the variation in measurements by a single person or instrument on the same item and under the same conditions”, it becomes more obvious. In both instances, you have error rates occurring in situations where either an individual or a machine is performing processes on a repeated basis. However being that the error rate is lower in the repeated process, it can be concluded that automated processes are more repeatable.