Re: Which control chart to monitor insulin protocol in an ICU

["Fatma Pakdil" <fpakdil@baskent.edu.tr>]

First of all, please think about that: What sort of decisions are you going
to make with this control chart ? I mean, is it meaningful to do what you
thought for your department ? Control charts tries to find out if there is
any significant "variance" in the same process. But, when you put together
different data coming from different patients may not be able to give you
meaningful information. In my point of view, each patient is a different
process. But, you can do that: Let's say your USL is 110 mg/dL, and LSL is
80 mg/dL (not UCL and LCL). If your patient stays at your hospital 30 days
and you measure him/her glucose number twice or three times a day, you have
60 measures. At least, let's say one measure per day. If you use X-bar
chart, it gives you something in terms of USL, LSL and the process's own
varience which appears on UCL, and LCL like these: (if USL and LSL covers
UCL and LCL)
- If your patient's values are higher than USL, that means process is out of
control.
- If your patient's values are lower than LSL, that means process is out of
control.
- If your patient's values are higher than UCL, that means process is in
control in spec.limits , but out of control in its own control limits.
- If your patient's values are lower than LCL, that means process is in
control in spec.limits , but out of control in its own control limits.

Shortly, what I am trying to say is to assume that each patient is a unique
process. Do not put different apples and pears in the same basket and
evaluate each of them in their own variability.

Hope to be helpful.

Fatma Pakdil, MBA, Ph.D.
Baskent University/Turkey


> [Moderator's Note:  Robert is not a DEN subscriber - so please copy him on
> all responses, or he will not see the response.  Thanks. ]