Glenn Miller and I have been e-mailing each other this last week, discussing his article, "Tutorial on Monitoring Time to the Next Medication Error: A Response" (Quality Management in Health Care. 2008;17(4):349-352). There is an error on page 350 in the last paragraph of the first column that explains the constant that was used to calculate the control limits for the XmR chart. Dr Miller thought that Hover and Walsh had used the A3 factor of 2.66 for calculating control limits that are based on moving standard deviations. However, Hover andWalsh actually used E2 factor of 2.66 that is recommended for individuals and moving range charts. This value is stated in Donald Wheeler and David Chambers' book (Understanding Statistical Process Control. Knoxville, TN: SPC Press; 1992:48). Also, the following URL lists 4 tables of constants to verify the A3 factor of 2.66 (in last table): http:/www.isixsigma.com/ontrol_charts/control_chart_table_of_constants.asp#indiv. Thus, Hover and Walsh used 2.66, which was the correct constant value for the XmR chart.

I thought Dr Miller's article was extremely helpful in explaining important statistical process control concepts, except I believe readers would be confused as I was about the recommendation to use 1.88 in calculating the control limits.

Karen Homa, PhD

Leadership and Preventive Medicine Residency Program, Dartmouth-Hitchcock Medical Center, Lebanon, NH 03756

[email protected]

Dr Homa has my full agreement and support on this. She provided the clarification to me and deserves credit for it.

James G. Miller, PhD

Piedmont College, Demorest, GA 30505

[email protected]