Table 4.

Effect size and study power analysis of memory performance, global cognitive function and mood in trials of memory strategy training in mild cognitive impairment (MCI)

Citation	Design	Outcome measure	Effect size and 95% confidence Interval	ES > .5 or < -.5	Power of study	> 1 whole number change in test score	Could Type II errors explain lack of significance?
Rapp et al 2002	RCT	Word List immediate T1	.33 (-.58, 1.23)	No	.16	Yes	Yes
		Word list immediate T2	-1.18 (-2.24, -.11)	Yes	.71	No
		Word list delay T1	.88 (-.07, 1.82)	Yes	.57	Yes
		Word list delay T2	-.36 (-1.36, .64)	No	.16	Yes
		Paragraph immediateT1	-.03 (-.93, .87)	No	.05	Yes
		Paragraph immediate T2	-.54 (- 1.54, .47)	Yes	.26	Yes
		Paragraph delay T1	.38 (-.53, 1.29)	No	.19	Yes
		Paragraph delay T2	-.14 (-1.13, .85)	No	.08	Yes
		Profile of mood states	No data
Belleville et al 2006	NRCT	Word list immediate	.78 (-.06,1.63)	Yes	.75	No	Yes
		Word list delay	.62 (-.22, 1.45)	Yes	.42	Yes
		Text immediate	.09 (-.73, .91)	No	.07	Yes
		Text delay	-.42 (-1.25, .41)	No	.25	No
		Scale of wellbeing	.21 (-.41, .83)	No	.13	Yes
Wenisch et al 2007	UCT	Story recall immediate	-.71 (-1.54, -.11)	Yes	Single	Yes	Yes
		Story recall delay	.09 (-.71, .89)	No	group	No
Troyer et al 2008	RCT	Word list T1	.23 (-.82, .35)	No	.18	No	No
		Word list T2	-.02 (-.58, .62)	No	.06	No

ES: Effect size. Relative ES for RCTs and NRCT, and ES for single intervention group in UCT were calculated by authors of this review as described in the methods analysis

Power analysis calculated by the authors of this review as described in the methods analysis

*Significance of effect size contrast from authors' ES analysis

NRCT: Non-randomised controlled trial

UCT: Uncontrolled trial

T1: Immediate

T2: At follow-up

C1: Control Cholinesterase inhibitor

C2: Control

ADAS-Cog: Alzheimer Disease Assessment Scale-Cognitive

NPI: Neuropsychiatry Inventory

MMSE: Mini Mental Status Examination

CVLT: California Verbal Learning Test

RBMT: Rivermead Behavioural Memory Test