[Skip to Content]
[Skip to Content Landing]
Citations 0
January 21, 2009

Physical Activity and Cognitive Function in Alzheimer Disease—Reply

JAMA. 2009;301(3):272-274. doi:10.1001/jama.2008.1012

In Reply: Dr Mortimer has pointed out an error that occurred in Table 7. The correct ADAS-Cog score for the control group should be geometric mean (SD) 6.4 (1.8) for completers and 10.6 (1.4) for dropouts. The results for participants in the intervention group are correct in Table 7. The data in Table 1 are correct, so that participants in the intervention and control groups had similar ADAS-Cog scores at baseline (P = .98).

Mortimer commented that the control group mean for the variable “all moderate-intensity–plus activities” was 40% higher than in the exercise group. He may have used the quotient of the difference of scores/baseline score to reach this conclusion ([177.2-126.4]/126.4). However, the variable “all moderate-intensity–plus activities” was log-transformed because it was not normally distributed. The log-transformed means for the exercise and control groups were 2.102 and 2.2485, respectively. Therefore, the difference between them was not 40%, but 7% ([{2.2485-2.102}/2.102] × 100). As Table 1 showed, the exercise and control groups were well matched.

First Page Preview View Large
First page PDF preview
First page PDF preview