Year: 1999
Type: letter
Abstract: A recent article by Dr. Humble 1 cautioned readers about common problems in applying statistical process control (SPC) charts in health care, perhaps most notably the general importance of selecting an appropriate control chart and correctly calculating the center line and control limits in particular applications.Given the utility and growing use of SPC, the author's caution to avoid such errors is sound advice, and to his list of caveats I also would add the following:• Creating control charts simply for the sake of creating charts, including widespread computer-generated graphs for almost every possible piece of data, without any real plan on how resultant information will be used.Using SPC requires time and resources, and there is little point in doing any type of analysis unless it will be used to inform or improve.• Selecting charts based only on what type of data is most readily and easily available, and not on what analysis or statistical power is required.Instead, ideally, explicit questions should dictate analysis methods and chart selection, which in turn should dictate the data that should be collected, not the other way around (retrospectively matching the best chart to preexistent data).I sometimes call this "reverse chart selection" or the "recliner armchair approach."The method requiring the minimum effort often will not yield the best results.• Not checking or verifying the assumptions of the type of chart being used, such as by simple histograms and autocorrelation tests, thereby increasing the possibility of affecting false alarm rates and the ability to detect process changes.In addition, a few comments and clarifications regarding specific statements in the above article might be offered, such as with respect to when each type of chart is appropriate.For example, situations in which p, c, or u charts should be used are a little confused and in particular are not dependent on the incidence rate as
DOI: 10.1086/503140
Author: James C. Benneyan
Institution: Northeastern University
Language: English
Subfield: Statistics, Probability and Uncertainty
Field: Decision Sciences
Domain: Social Sciences
Open Access status: bronze
OpenAlex ID: W2119675188