Table 2

Main outcome table. Baseline Montgomery Åsberg Depression Rating Scale (MADRS) scores and change in the depressive symptoms over 52 weeks for patients in the intention-to-treat and per-protocol analysis. Values are means unless stated otherwise. Estimated mean differences and 95% Confidence Intervals (CIs) are presented.

Intention-to-treat MADRS (n = 181)

Usual care + paroxetine (n = 85)
 Usual care (n = 96)
 Mean difference*
 SD
 95% CI
Baseline score
 23.7
 24.1


6 weeks – baseline†
 -7.6
 -6.0
 -1.6
 20.2
 -4.7; 1.4‡
13 weeks – baseline †
 -10.2
 -8.7
 -1.5
 22.5
 -5.1; 1.9
26 weeks – baseline †
 -13.0
 -10.0
 -3.0
 21.3
 -6.4; 0.3
52 weeks – baseline †
 -14.7
 -12.6
 -2.1
 24.1
 -6.1; 1.9

Per-protocol MADRS (n = 133)


Usual care + paroxetine (n = 55)
 Usual care (n = 78)
 Mean difference*
 SD
 95% CI
Baseline score
 25.1
 24.1


6 weeks-baseline†
 -7.8
 -6.7
 -1.1
 20.0
 -4.5; 2.4‡
13 weeks-baseline†
 -12.1
 -9.0
 -3.1
 20.7
 -6.8; 0.5
26 weeks-baseline†
 -13.9
 -9.3
 -4.6
 20.6
 -8.4; -0.9§
52 weeks-baseline†
 -17.6
 -12.6
 -5.0
 21.2
 -9.1; -1.0§

*A negative mean difference is a difference in favor of usual care + paroxetine.

†A negative score means a decrease in the severity of symptoms.

‡The 95% CI of the difference in symptom change lies entirely between the equivalence margins of -5 and +5 points, indicating equivalence of usual care + paroxetine and usual care.

§The 95% CI of the difference in symptom change lies entirely to the left of zero, indicating statistical significant differences in favor of usual care + paroxetine.

Hermens et al. BMC Medicine 2007 5:36   doi:10.1186/1741-7015-5-36