Table 4.

Outcomes at weeks 9 and 17 for the per-protocol sample (n=117).

Outcomesa		Smartphone CBT + medication change (n=60), mean/% (95% CI)	Medication change alone (n=57), mean/% (95% CI)	Adjusted difference/OR (95% CI)b	P value
End of randomized trial (week 9)
	PHQ-9	8.92 (7.81, 10.03)	10.64 (9.52, 11.76)	–1.72 (–3.18, –0.25)c	.02
	Remission	18.2% (8.5%, 34.8%)	10.0% (4.0%, 23.2%)	1.99 (0.74, 5.38)	.17
	Response	31.6% (18.7%, 48.3%)	18.0% (9.2%, 32.3%)	2.11 (0.92, 4.85)	.08
	BDI-II	21.0 (18.6, 23.5)	24.2 (21.7, 26.8)	–3.2 (–6.3, 0.0)	.05
	FIBSER	4.11 (3.44, 4.78)	4.86 (4.18, 5.53)	–0.75 (–1.47, –0.03)	.04
Follow-up (week 17)
	PHQ-9	8.92 (7.40, 10.44)	8.85 (7.31, 10.39)	0.07 (–1.68, 1.82)	.94
	BDI-II	19.4 (15.9, 22.9)	20.0 (16.5, 23.6)	–0.6 (–4.4, 3.1)	.75
	FIBSER	4.14 (3.40, 4.88)	4.46 (3.71, 5.21)	–0.32 (–1.28, 0.63)	.50

^aBDI-II: Beck Depression Inventory-II; FIBSER: Frequency, Intensity, and Burden of Side Effects Ratings; PHQ-9: Patient Health Questionnaire 9.

^bFor each continuous outcome up to week 9, we used a linear mixed model including sites and patients as random effects and time (5 and 9 weeks), treatment, and time*treatment interaction, adjusting for its baseline score and the stratification variables, as fixed effects. For the continuous outcome at week 17 follow-up, we used the similar linear mixed model but without time*treatment interaction. For the dichotomous outcomes, the generalized linear mixed model was used along with the same random effects and the fixed effects. The summary effect measures are adjusted score differences for PHQ-9, BDI-II, and FIBSER, and are odds ratios for remission and response.

^cPrimary endpoint per protocol.