A space to discuss all things related to pulmonary function testing! Equipment, predicted values, protocols, and challenges. I hope this will be a useful space for technicians, physicians, and others to "meet" and share information.
Which equipment does your laboratory use?
Search This Blog
Followers
Monday, July 19, 2010
Z-scores
Anyone else using them in addition to or instead of %predicteds? I am working on an interpretation schema that utilizes cutpoints of Z-scores instead of %pred. I suppose neither is perfect...
Our software (ComPAS, Morgan Scientific) already calculates and reports Z-scores for all our bigger patient spirometry. I think looking for values -1.65 < Z < +1.65 for deciding normal rather than a bunch of different cut points for %predicteds makes sense. I'll try to post my proposed idea for using Z-score cut points and would love some feedback.
Ok, here is an idea. It's based on the Intermountain Thoracic Society classification based on confidence intervals. They used:
Mild obstruction >1 and <2 CIs Moderate obstruction >2 and <4 CIs Severe obstruction >4 CIs
Assuming 1 CI = 1.65 SEs, then Normal: Z> -1.65 Mild obstruction -3.3 > Z > -1.65 Moderate obstruction -3.3 > Z > -6.6 Severe obstruction Z < -6.6
I've had some interesting discussions with Philip about this and not sure this makes sense. The incremental change of an additional standard deviation at the extremes is quite small.
Are you talking about Z scores for older patient spirometry? I think it's a great idea to develop them. Clement Ren
ReplyDeleteOur software (ComPAS, Morgan Scientific) already calculates and reports Z-scores for all our bigger patient spirometry. I think looking for values -1.65 < Z < +1.65 for deciding normal rather than a bunch of different cut points for %predicteds makes sense. I'll try to post my proposed idea for using Z-score cut points and would love some feedback.
ReplyDeleteOk, here is an idea. It's based on the Intermountain Thoracic Society classification based on confidence intervals. They used:
ReplyDeleteMild obstruction >1 and <2 CIs
Moderate obstruction >2 and <4 CIs
Severe obstruction >4 CIs
Assuming 1 CI = 1.65 SEs, then
Normal: Z> -1.65
Mild obstruction -3.3 > Z > -1.65
Moderate obstruction -3.3 > Z > -6.6
Severe obstruction Z < -6.6
I've had some interesting discussions with Philip about this and not sure this makes sense. The incremental change of an additional standard deviation at the extremes is quite small.