Table 3.

Results of proportion meta-analyses

				95% CI		Heterogeneity statistics
Analysis	k	N	Prevalence (%)	LI	UL	Q	df	p	I2 (%)
Dropout from all treatment arms excluding WL	66	2658	11.7	9.0	14.6	326.5	65	<0.001	79.0
Lower quality removed	53	2383	11.6	8.8	14.8	286.7	52	<0.001	80.7
Defined dropout	32	1386	15.9	12.0	20.2	132.0	31	<0.001	76.1
Dropout from all TFCBT arms	41	1696	10.6	7.5	14.2	206.1	40	<0.001	79.3
Lower quality removed	31	1457	10.1	6.7	14.0	166.8	30	<0.001	80.1
Defined dropout	16	778	14.7	9.4	20.9	70.1	15	<0.001	78.7
Dropout from all TFCBT and EMDR arms	48	1869	11.2	8.2	14.6	226.5	47	<0.001	77.6
Lower quality removed	36	1608	10.8	7.6	14.5	186.7	35	<0.001	79.2
Defined dropout	22	891	15.2	10.6	20.4	85.3	21	<0.001	74.9
Dropout from all EMDR arms	7	173	15.5	7.8	25.3	15.7	6	0.015	59.0
Lower quality removed	5	151	16.2	6.9	28.5	14.7	4	0.005	70.1
Defined dropout	6	160	16.7	8.0	27.8	15.1	5	0.010	63.6
Dropout from all non-trauma focussed armsa	18	789	12.8	7.6	19.1	90.1	17	<0.001	82.4
Lower quality removed	17	775	13.4	7.9	20.0	87.8	16	<0.001	83.1
Defined dropout	10	495	17.4	10.5	25.6	43.4	9	<0.001	79.2

WL, waiting list; TFCBT, trauma-focused cognitive behavioural therapies; EMDR, eye movement desensitization and reprocessing.

^aAll active control arms, non-NICE recommended psychotherapies and the arms from component studies with exposure or trauma narrative elements removed.