Table 2.

A statistical summary of the effect of 56-day curcuminoid supplementation on flow mediated dilation (%).

Contrast	Estimate	Confidence interval (90%)		p value	Clinical inference statistics
		Upper	Lower		Likelihood (%) benefit/trivial/harmc	Qualitativec	Benefit oddsc
Time effecta
Placebo	−0.3	1.3	−1.9	0.769
50 mg	1.4	3.1	−0.3	0.174
200 mg	2.7	4.3	1.1	0.007
Treatment ∗ time effectb
50 mg-Placebo	1.7	4.0	−0.6	0.234	68.6/28.6/2.8	↑ possible	25 : 1
200 mg-Placebo	3.0	5.3	0.7	0.032	92.8/7.0/0.2	↑ likely	546 : 1
200 mg-50 mg	1.3	3.7	−1.0	0.349	59.2/35.8/4.9	↑ possible	12 : 1

^aPost-pre intervention.

^bFor example, 50 mg post-pre–200 mg post-pre.

^cThe smallest worthwhile clinical change is 1%. The probability that a contrast was at least greater than the clinical threshold was 25–75% possible, 75–95% likely, 95–99.5% very likely, and >99.5% almost certain. The terms benefit, trivial, and harm also denoted by symbols ↑, ↔, and ↓, respectively, refer to the most likely directional outcome relative to the smallest effect threshold. Unclear refers to outcomes where the likelihood of both benefit and harm exceeded 5%. The clinical adoption threshold was expressed as a benefit: harm odds ratio >66 : 1.