Table 2.

Primary (PHQ-9 < 10; improvement) and secondary (PHQ-9 scores reduced by at least 50% between baseline and the first and second follow-up assessments; reduction) outcomes at the 3-month and 5-month follow-ups

Improvement from and reduction in depressive symptomatology at the 3-month and 5-month follow-ups	Intervention	Control
	n of total n (%)	n of total n (%)	Risk difference in percentage points (95% CI)a	OR (95% CI)b,c	Pd
Primary outcome: improvement from depressive symptomatology at 3 monthse	109/257 (42.4)	87/270 (32.2)	10.2 (2.0–18.4)	1.57 (1.07–2.29)	0.019
Secondary outcome: improvement from depressive symptomatology at 5 monthse	127/250 (50.8)	130/260 (50.0)	0.8 (−7.9–9.5)	1.01 (0.70–1.46)	0.938
Secondary outcome: reduction in depressive symptomatology at 3 monthse	95/257 (37.0)	73/270 (27.0)	9.9 (2.0–17.9)	1.58 (1.08–2.29)	0.016
Secondary outcome: reduction in depressive symptomatology at 5 monthse	111/250 (44.4)	112/260 (43.1)	1.3 (−7.3–9.9)	1.04 (0.73–1.48)	0.841

^aThese (post hoc) analyses are based on crude (that is, unadjusted) data on a complete case basis.

^bORs and 95% CIs were calculated using fixed-effects logistic regression models, from which the P values were derived.

^cAll estimates had missing data imputed according to trial arm using the multiple imputation by chained equations (MICE) models, which included stratification (sex, age group and baseline severity of PHQ-9 score (categorical)) and predictors of missingness (extended data).

^dAll P values reported are based on two-sided Wald tests from the relevant logistic regression model, with no adjustments for multiple tests.

^eThe primary and secondary outcomes of improvement from depressive symptomatology were defined as PHQ-9 scores less than 10; the secondary outcome of reduction in depressive symptomatology was defined as PHQ-9 scores reducing by at least 50% between baseline and the first and second follow-up assessments (yes/no).