Table 4.

Indirect and direct effects from causal mediation analysis on depressive symptoms at posttreatment (n=64)^a.

Mediator	Nature indirect effect		Nature direct effect			Proportion mediated
	b (95% CI)	P value	b (95% CI)	P value	Percentage, %		P value
CBTSQb	–2.29 (–3.58 to –1.01)	<.001	–1.35 (–3.38 to 0.68)	.19	62.9		.005
BADS-SFc	–1.27 (–2.39 to –.15)	.03	–2.41 (–4.09 to –0.72)	.005	34.5		.02
BNSG-Sd	–1.38 (–2.45 to –.31)	.01	–1.57 (–3.36 to 0.23)	.09	46.8		.02

^aThe outcome was depressive symptoms, as measured by the Patient Health Questionnaire-9, with a higher score indicating greater symptoms. The control group was used as the reference in the treatment model. Each mediator was evaluated individually in separate models. Each model included potential confounders, such as age, gender, education, income, living arrangement, and count of chronic physical conditions, in both the mediator and outcome models. Baseline depression was included in the outcome equation, and the baseline score of each mediator was included in the mediator equation. In addition, the outcome equation includes the treatment-mediator interaction. Analyses were conducted using complete data without imputation).

^bCBTSQ: Cognitive Behavioral Therapy Skills Questionnaire.

^cBADS-SF: Behavioral Activation for Depression Scale–Short Form.

^dBNSG-S: Basic Needs Satisfaction in General Scale.