Thus, the objective of this study was to provide code to calculate power curves and to investigate certain properties of multirule QC. Analysts could easily experiment with QC rule adjustments if such code were available. The published power curves were produced by simulation however, to our knowledge, the code was not published. Also, power curves can be generated using simulation, which is much more accessible. Power curves can be calculated, but these calculations require advanced mathematical methods (Markov analysis) that are not accessible to most laboratorians. It would be helpful if analysts had simple tools that could be used to predict the behavior of rule adjustments. Power curves have been published for a standard set of rules (Westgard rules), 2, 3 but the list of potential rules is very long, and it is impractical to publish power curves for each case. Second, it is difficult to predict the performance of the monitoring system in response to rule adjustments. First, each rule increases the probability of false rejection this probability can be substantial when many rules are applied. The process would fail if either rule were triggered.Īlthough these rules increase sensitivity, they have several disadvantages. For example, 2 consecutive results on 1 side of the 2-sigma limit might be considered a failure and be used in addition to the 3-sigma limit. Multirules are created by applying several different criteria for QC failure. For this reason, many laboratories use multirules to increase the sensitivity of QC monitoring. Although such rules have the advantage of simplicity, they are not very sensitive and often fail to detect small shifts in the mean. The simplest and most common rule is to use 3 SD (sigma) limits, classify the process as unstable, and begin troubleshooting when results exceed these limits ( 1 3S rule). Various methods are available to monitor QC processes. QC processes use statistical methods to detect a signal ( assignable cause variation) in the presence of background noise ( common cause variation). Any change in the distribution of the measured output suggests the presence of assignable cause variation that is synonymous with instability. Quality control (QC) is used to monitor processes to detect departures from normal operations or instability. Quality control, error detection, false rejection, Westgard rules, multirules, simulation
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |