Table 3 Analysis of main and interaction effects using MMRM for PHQ-9 score at week 6, with the delayed treatment as the control arm

From: Cognitive behavioral therapy skills via a smartphone app for subthreshold depression among adults in the community: the RESiLIENT randomized controlled trial

Triala

Arm

n

Change scores from baseline of the PHQ-9

Differences in change scores of the PHQ-9 at week 6

Estimated least squares mean changeb

Effect sizec

Estimated least squares mean difference

Effect sizec

P value (main effect)

P value (interaction of two components)d

Week 1

Week 2

Week 3

Week 4

Week 5

Week 6

Week 6

Trial 1

BA + CR

329

−1.40

−1.97

−2.14

−2.36

−2.32

−2.82

−1.04

(−1.35 to −0.73)

−2.17

(−2.65 to −1.69)

−0.64

(−0.78 to −0.50)

2.1 × 10−18

6.2 × 10−8

BA

327

−1.44

−1.88

−2.21

−2.40

−2.60

−2.86

−1.06

(−1.37 to −0.75)

−2.22

(−2.70 to −1.74)

−0.65

(−0.79 to −0.51)

4.4 × 10−19

 

CR

328

−1.03

−1.56

−2.04

−2.12

−2.29

−2.48

−0.92

(−1.23 to −0.61)

−1.84

(−2.32 to −1.36)

−0.54

(−0.68 to −0.40)

8.9 × 10−14

 

Delayed treatment (reference)

328

–

–

−0.42

–

–

−0.65

−0.24

(−0.55 to 0.07)

–

–

–

 

Trial 2

BA + PS

329

−1.59

−2.11

−2.70

−2.56

−3.03

−3.34

−1.24

(−1.51 to −0.97)

−2.27

(−2.74 to −1.80)

−0.67

(−0.81 to −0.53)

2.4 × 10−20

6.6 × 10−7

BA

327

−1.86

−2.31

−2.63

−2.83

−3.03

−3.29

−1.22

(−1.50 to −0.95)

−2.22

(−2.70 to −1.75)

−0.65

(−0.79 to −0.51)

2.2 × 10−19

 

PS

328

−1.24

−2.00

−1.97

−2.14

−2.35

−2.84

−1.05

(−1.33 to −0.78)

−1.77

(−2.24 to −1.29)

−0.52

(−0.66 to −0.38)

5.9 × 10−13

 

Delayed treatment (reference)

328

–

–

−0.85

–

–

−1.07

−0.40

(−0.67 to −0.12)

–

–

–

 

Trial 3

BA + AT

328

−1.42

−2.04

−2.44

−2.60

−2.77

−3.17

−1.17

(−1.49 to −0.85)

−2.08

(−2.58 to −1.59)

−0.59

(−0.73 to −0.45)

5.1 × 10−16

3.8 × 10−8

BA

327

−1.88

−2.33

−2.65

−2.85

−3.05

−3.31

−1.22

(−1.54 to −0.90)

−2.22

(−2.72 to −1.73)

−0.63

(−0.77 to −0.49)

6.4 × 10−18

 

AT

328

−1.26

−2.05

−2.20

−2.55

−2.74

−2.94

−1.08

(−1.41 to −0.76)

−1.85

(−2.35 to −1.35)

−0.52

(−0.66 to −0.38)

4.6 × 10−13

 

Delayed treatment (reference)

328

–

–

−0.87

–

–

−1.09

−0.40

(−0.72 to −0.08)

–

–

–

 

Trial 4e

BA + BI

311

−1.78

−2.23

−2.38

−2.84

−2.84

−3.22

−1.23

(−1.54 to −0.92)

−2.00

(−2.50 to −1.50)

−0.57

(−0.71 to −0.43)

7.7 × 10−15

3.8 × 10−11

BA

314

−2.06

−2.48

−2.79

−3.01

−3.21

−3.49

−1.33

(−1.64 to −1.03)

−2.27

(−2.77 to −1.77)

−0.64

(−0.79 to −0.50)

1.9 × 10−18

 

BI

313

−2.23

−2.61

−2.79

−3.00

−3.13

−3.35

−1.28

(−1.58 to −0.98)

−2.13

(−2.63 to −1.63)

−0.60

(−0.75 to −0.46)

1.3 × 10−16

 

Delayed treatment (reference)

313

–

–

−0.98

–

–

−1.22

−0.47

(−0.77 to −0.16)

–

–

–

 
  1. Least squares mean changes and differences in the PHQ-9 scores were estimated using MMRM. The regression models included fixed effects of treatment, visit (as categorical) and treatment-by-visit interactions, adjusted for baseline PHQ-9 scores, employment status, age and gender. The interaction P values are calculated from the MMRM involving interaction terms for the corresponding treatments. P values are two-sided and no adjustments are made for multiple comparisons.
  2. aSee Fig. 1 for the design of trials 1 to 4.
  3. bA minus value indicates a reduction in depression. Values in parentheses are 95% CIs.
  4. cEffect sizes (SMD) were calculated using the observed baseline s.d. for within-group change scores and the observed week 6 s.d. for group differences.
  5. dA sensitivity analysis with linear regression analyses using only week 6 data for each of the 2 × 2 trials confirmed that all interaction terms had P values <0.001.
  6. eParticipants working on a three-shift schedule were excluded in the primary and sensitivity analyses for trial 4.
Back to article page