Define significant changes in neuropsychological tests

Skip to first unread message


May 1, 2008, 12:51:20 PM5/1/08
to MedStats
We are using a series of cognitive test to see whether or not a group
of patients improve after a specific intervention. In dealing with
these type of data are susceptible to practice effects (test-retest),
the common practice is to define “improvement” as follows:
Standard deviation method: changes are defined as as a change from
preoperative test score by at least 1SD of the mean baseline score.
Fixed 20%: decline of at least 20% from baseline
We have never been happy with this standards. Reviewing new
methodologies I have came across two different options called:
Reliable Change Index (RCI), modified for practice
Standardized Regression-based (SRB) technique
These were published by
Raymond PD, Hinton-Bayre AD, Radel M, Ray MJ, Marsh NA: Assessment of
statistical change criteria used to define significant change in
neuropsychological test performance following cardiac surgery. Eur J
Cardiothorac Surg 29:82-88, 2006
I have upload the PDF to the files section.
Any of you has any suggestion about which one to use in order to
correct for the learning effect?
Thanks a lot
John B.

Reply all
Reply to author
0 new messages