Table 4.

Predicted change in d2 outcomes after intervention with fish (n = 137), meat (n = 148), n-3 supplements (n = 141)

					Models adjusted for:
		Crude			Baseline scorea		Baseline, dietary complianceb
d2 test of attention outcomesc		Pre Mean ± SD	Post Mean ± SD	P-within d	IRR (95% CI)	P-value	IRR (95% CI)	P-value
E1 errors
	Fish	24.6 ± 28.3	25.3 ± 27.9	0.544	1 (ref.)		1 (ref.)
	Meat	19.0 ± 19.3	16.4 ± 15.9	0.074	0.85 (0.74, 0.98)	0.026	0.88 (0.76, 1.02)	0.084
	Supplement	25.4 ± 23.0	22.8 ± 24.1	0.161	1.01 (0.83, 1.23)	0.933	1.06 (0.88, 1.29)	0.528
E2 errors
	Fish	6.7 ± 8.8	3.5 ± 6.1	<0.001	1 (ref.)		1 (ref.)
	Meat	5.9 ± 6.2	3.5 ± 9.2	0.001	0.91 (0.59, 1.39)	0.648	0.92 (0.60, 1.40)	0.681
	Supplement	6.6 ± 9.0	3.6 ± 6.3	<0.001	0.88 (0.63, 1.24)	0.469	0.90 (0.61, 1.32)	0.586
Total errors
	Fish	31.3 ± 32.6	28.8 ± 30.1	0.093	1 (ref.)		1 (ref.)
	Meat	24.9 ± 21.7	19.9 ± 20.8	0.006	0.88 (0.75, 1.02)	0.094	0.91 (0.77, 1.07)	0.247
	Supplement	32.0 ± 27.6	26.4 ± 27.1	0.004	0.96 (0.80, 1.15)	0.671	1.03 (0.86, 1.23)	0.772

Pre and post data are presented as mean ± SD and difference between treatment groups presented as coefficients (95% CI). Abbreviations: SD Standard deviation, IRR Incidence rate ratio, CI Confidence interval, E1 Errors of omission, E2 Errors of commission; Total errors (E1 + E2 errors)

^aBetween group differences analyzed using general estimates equation model, with the negative binomial distribution, exchangeable correlation structure and robust standard errors

^bAdjusted for the equivalent outcome at baseline and for dietary dietary compliance (i.e. the total intake of study meals or supplements) during the trial

^cA decrease in the number of errors indicates improvement

^dPaired-samples T-test for comparison within treatment groups from pre to post intervention