. 2021 Dec 21;326(23):2385–2394. doi: 10.1001/jama.2021.21187

Table 2. 8-Item Patient Health Questionnaire (PHQ-8) Depression Scale Mood Scores Each Year Since Randomization vs Baseline^a.

	Omega-3 group		Placebo group		Mean difference in change (95% CI)^c	P value	P value for interaction^d
	No. of participants	Adjusted mean (95% CI)^b	No. of participants	Adjusted mean (95% CI)^b	Mean difference in change (95% CI)^c	P value	P value for interaction^d
Baseline	9171	1.09 (1.06 to 1.13)	9182	1.11 (1.08 to 1.15)
Year 1 vs baseline	8471	0.05 (0.01 to 0.09)	8549	0.01 (−0.02 to 0.05)	0.03 (−0.02 to 0.09)	.21
Year 2 vs baseline	8354	0.07 (0.03 to 0.11)	8371	0.04 (0.00 to 0.08)	0.03 (−0.03 to 0.08)	.38
Year 3 vs baseline	8116	0.10 (0.06 to 0.15)	8172	0.06 (0.02 to 0.10)	0.05 (−0.01 to 0.10)	.13
Year 4 vs baseline	7676	0.08 (0.04 to 0.12)	7690	0.05 (0.01 to 0.09)	0.03 (−0.03 to 0.09)	.29
Year 5 vs baseline	5295	0.18 (0.13 to 0.23)	5252	0.19 (0.13 to 0.24)	−0.01 (−0.08 to 0.06)	.79	.52
1- to 5-y Average vs baseline	9171		9182		0.03 (−0.01 to 0.07)	.19

^{^a}

Assesses the severity of depressive symptoms (range, 0-24). Higher scores indicate higher severity of depression. The prespecified minimal clinically important difference for change in score was 0.5. Distribution of PHQ-8 scores in omega-3 fatty acid and placebo groups by study year is provided eFigure 4 in Supplement 2. Analyses were from general linear models of response profiles to estimate the means, with time modeled as indicator variables; models were controlled for age, sex, and vitamin D₃ randomization group.

^{^b}

Adjusted means (95% CI) within each treatment group are shown at baseline and adjusted mean differences in change (95% CI) within each treatment group are shown for each follow-up time point. The repeated measures model used all available responses and can handle missing outcome data, which were assumed missing at random.

^{^c}

Net difference in change comparing the omega-3 fatty acid and placebo groups from the repeated measures model. The final row shows the main outcome of the net difference between treatment groups in PHQ-8 change scores averaged across all follow-up years (years 1-5) from the repeated measures model; all participants contributed data to the repeated measures analysis at 1 or more times.

^{^d}

The P value for interaction is from the test of the 5 degree-of-freedom treatment × time interaction term in the model.

Table 2. 8-Item Patient Health Questionnaire (PHQ-8) Depression Scale Mood Scores Each Year Since Randomization vs Baselinea.

Table 2. 8-Item Patient Health Questionnaire (PHQ-8) Depression Scale Mood Scores Each Year Since Randomization vs Baseline^a.