Weight Maintenance – JMP 19 Process Screening to the Rescue!

[#1]

After I hit a high record weight, a serious diet was undertaken to get my weight back to a healthier range.   But this wasn’t my “first rodeo” with big large diets, and I always found that I would inevitably over time gain back the weight that I just lost!  Most diets provide helpful tools and guidance to help you monitor and adjust your activities while you are trying to lose weight.  However, I wondered if methods were available to help you after the diet to maintain and keep your weight in a healthy range?

Given my early career start as a Quality Engineer, I immediately thought of using Control Charts methods (also called SPC or Statistical Process Control).  While most Control Charts for continuous measures (like weight) would require an underlying assumption of normal distributed data, Individual Moving Range type of Control Charts had been shown to be very robust and useful for any type of continuous data.  More importantly, Control Charts generate Upper and Lower Control Limits (set at + or - 3 Standard Deviations from the Mean) which would provide a range covering 99.73% of expected normal cause variation.  Seeing points beyond control limits would be indicative of special cause variation and warrant action.  We tried this for a month after the completion of the diet to see if the calculated limits could provide a good range for weight maintenance.  However, the control chart limits were very tight (LCL – Lower Control Limits at 170.1 and UCL – Upper Control Limits at 174.2) indicating a poor data type fit for the use of this tool.

(See Graph 1 - Initial Weight Loss)

(See Graph 2 – Individual & Moving Range Chart of Weight)

In many ways my daily weight performance does not behave like normally distributed data.  For example, some slight runs up in weight can be seen on weekends when larger or restaurant meals were always on my plate, especially if around a holiday party.  Then my weight often trends down as we slide back into more diet controllable home-cooked weekday eating.  So maybe this data can be handled a different way with the help of JMP 19 Process Screening!

The improved JMP 19 Process Screening platform has new features that are meant to help handle non-normal continuous data situations, such as when you have count data in environmental monitoring.  It will handle our non-normal continuous weight data and will auto select the best non-normal distribution fit type.  It also further allows us to set custom “Action and Alert Limits” that can show us when we need to investigate and potentially make changes.  You set these limits based on quantile levels of your data.   We selected 99.85% and 0.15% range for “Action” and 97.5% and 2.5% range for “Alerts” (which is a similar concept to using + and – 2 to 3 standard deviations for the mean in the setting of upper and lower control chart limits). 

(See Graph 3 – Setup and Process Screening to Establish Weight Range)

The Processing Screening chose a LogNormal distribution basis for our weight data and gave us much a better fitting range.  The upper “Alert” limit is set at 174.89 pounds, so if we get beyond this point then we need to start to be concerned and maybe cut back on our calories.  And if we get beyond the “Action” limit of 176.33 then we need to take immediate steps to address our diet.  Now we can save and use these limits on all future data points!

(See Graph 4 – Process Screening with Specified Limits)

But will this work in practice to maintain or weight well after the diet is over?  Good news is that with the help of these “Action and Alert Limits” I have been able to adjust my diet and keep my weight within a healthy range.  As you can see below, I even was able to adjust after some high calorie holiday celebrations that would have put me on the path of unhealthy sustained weight gain.

(See Graph 5 – Weight Maintenance in Action)

Also note that another big advantage of the Process Screening platform is its ability to handle many measurements at once and generate a table that will show you which ones need the most immediate attention.  Whether used for personal or business use, the JMP 19 Process Screening can help you control anywhere you have multiple and challenging measures to track!

 I can envision in the future using this tool to control other weight related measurement drivers all at the same time for things like my daily steps, workout intervals, cholesterol, oxygen and blood pressure readings, and more.  Then we can all handle our weight maintenance better and avoid the numerous large diets that Grammy Award winning country music artist Bobby Bare pays homage to in his hit “Diet Song”  where he sings “this dietin’ dietin’ dietin’ dietin’ sure is a rough way to die.”

Note: Try out the new Process Screening for yourself using the attached Practice Weight.jmp.

Practice Weight.jmp