Participants in the TTM arms were slightly but not significantly
more likely to make positive stage changes in the TTM arms than
the control arm when those whose stage was missing were
included as having made no positive stage change (Table 3).
However, when these people were excluded, the effect of the TTM
arm was slightly greater, though this was not statistically significant
with or without adjustment for potential confounders. The effect
size was small, with an increase in odds of 20% or so. The mean
change score in the controls was 0.39 and in the TTM was 0.46,
a difference (95% CI) of 0.08 (0.00–0.16), which was not statistically
significant (p ¼0.064). On adjustment for confounders, this difference
was reduced slightly to 0.07 (0.01 to 0.15). When those with
missing data were excluded, the unadjusted and adjusted mean
differences (95% CIs) were 0.12 (0.01 to 0.24) and 0.10 (0.03 to
0.22).