T-tests used to compare patients with different baseline clinical characteristics such as prior 6 month weight loss, performance status, and number of primary disease symptoms.

ANCOVAs used on the baseline to 12 week change scores (controlling for those baseline factors that were found to be associated with LCS & TOI scores in the prior analysis.

Distribution-based measures:

Traditionally, 1/3 to 1/2 SD corresponds to an effect size that is considered moderate. This is what we used.

We used a 1 SEM criterion (as others have used).

Very simple! The SD the HRQoL score at baseline times the square root of 1 - the score’s reliability.