Response on Control Charts
March 31, 2011
Dear Eileen,
I think that the work you are doing to reduce healthcare-acquired infections in clinical settings is to be admired and is very worthwhile. I wish you many successes in your efforts. This letter is written in response to your March 15, 2011 blog entitled “Are We Captives of Control Chart.”
That blog was not written by someone who understands control charts, how they are used, or the information that is contained in this website. It is very misleading.
In addition, your blog provides links to three of the newsletters on our website. I don’t mind people linking our newsletters. In fact, I think it is great. We have a wealth of free knowledge there – not just on control charts but on many other statistical techniques. If you hadn’t brought in the website, I would probably not have responded to your blog. I wrote this month’s newsletter on the purpose of control charts. I would ask that you read it. Below I respond to some of your misleading information on control charts. I am glad your blog was short. My response is longer than your blog.
- “We spend a lot of time constructing charts that show upper and lower confidence limits.” Control limits are not confidence limits. Confidence limits deal with probability – for example a 95% confidence interval around an average. You cannot determine the probability of a point being due to special causes or common causes by its location on the control chart – whether in or out of control.
- “We have rules for what it means to be out of control: one spike above two standard deviations.” A spike above two standard deviations is not considered out of control.
- “The goal of a control chart is to keep the process average. The center line represents average performance and, in business, the goal is to keep it there.” This is very misleading. The goal of business is not “average” performance. In manufacturing of products (including many healthcare products), the goal is to provide the customer consistent and predictable product that meets their requirements. The goal of a control chart is not to keep the process average. It is to determine that average, determine the variation about that average, determine if the process is consistent and predictable (in control), to remove special causes of variation that prevent stable processes and then make a decision – is the process meeting our expectations or do we need to fundamentally improve it. It is not about keeping the process average. “Business” – including healthcare – wants to improve over time. The goal should always be continuous improvement.
- “This month I am linking you to a business consulting website that explains control charts. Being a business site, the explanations are about applying control charts to the production of widgets.” I probably would not have responded to this blog if Eileen had left this website out of it. Of course, there are some examples of the production of “widgets” on the website, but those examples don’t happen too often. Here are the examples used in our most recent newsletters starting with last month’s newsletter. It was on “g control charts” and the example used —- infections after surgery.
- Feb 2011: g control chart: infections after surgery
- Jan 2011: Xbar-R-mR control charts: reactor purity
- Dec 2010: Teaching Variation to Leadership: sales
- Nov 2010: Using Pareto Diagrams and Control Charts Together: On-time Shipments
- Oct 2010: Inferences on Proportions: Polling data
- Sept: 2010: Where Do Control Limits Come From: no examples
- Aug: 2010: Levey-Jennings Charts: Cholesterol test in a laboratory
- July 2010: Waterfall Charts: Net income
- June 2010: Attribute Gage R&R studies: WIDGETS!
- “Notably absent are recommendations to use control charts to measure complex healthcare processes that involve not just machinery, but the unpredictability of people (both patients and workers).” Yes, I agree that we don’t address healthcare too often on the website. But, I know of no process that is just machinery. Someone has to set it up, buy the material, develop the methods – the list goes on and on. People are always part of the process. Review the simple cause and effect diagram that is used for finding causes of problems. Often, the main categories are machines, methods, materials, measurement, environment and people. People are part of the process and they contribute to both common and special causes of variation. That is part of the information contained in variation.
- But there are plenty of references about people who have done this. For example, on Feb. 1, 2011, Don Goldmann gave a presentation on using continuous improvement in infection prevention. It was put on by www.webbertraining.com, which gives web-based training for infection prevention and control. Dr. Goldmann is with the Harvard Medical School as a Professor of Pediatrics and with the Harvard School of Public Health as a Professor of Immunology and Infectious Disease and Epidemiology. His presentation is very informative as it talks about some of the issues raised when trying to work on quality improvement in infection prevention. He has one slide that says: “Epidemiology can take you only so far….”
- It appears that he has a well-thought-out approach for applying continuous improvement techniques to infection prevention. He uses different tools to help him – process flow diagram, cause and effect diagrams, priority matrix – and, yes, control charts. He shows a control chart that shows the step-changes in reducing the infection rate – one of the purposes of a control chart – to show improvement and monitor for any special causes. The link to Dr. Goldman’s presentation is here: http://webbertraining.com/recordingslibraryc4.php.
- “Please let me know if there are any questions I can answer about these articles.” You provide links to three of our newsletters. Not sure what criteria you used to select them. However, I would not dream of answering questions about what you do on a daily basis. Please don’t answer questions about my articles unless you understand what you are saying
I could go with asking some questions of you – like what statistics do you use, who that “someone” was who decided that control charts had to be used in healthcare, or what kind of environment causes people to dread out of control points. But I will stop here.
Best wishes for continued success in your work. If you are interested in exploring how control charts can be used in your work, please let me know. I will be happy to work with you (no charge) plus supply our SPC for Excel software for your use (again no charge). Just contact me.
Bill