Table 2.

Logistic regression model parameter estimates for dropout and remission for cognitive-behavioral therapy guided-self-help (CBTgsh) and therapist-led (CBTth).

Outcome	Model Parameter	Estimate	SE	p-value	Effect size
Dropout	Intercept	−2.41	1.31	0.0656	n/a
	EDE Weight Concern	0.42	0.21	0.0439	1.78
	Treatment length	−0.15	0.06	0.0072	1.16
	CBTgsh treatment	3.00	0.92	0.0011	19.99
	EDE Weight Concern*CBTgsh	−0.63	0.24	0.0095	2.91
	Treatment length*CBTgsh	0.26	0.08	0.0016	1.79
Binge-eating Remission	Intercept	−0.26	1.12	0.8184	n/a
	Age	0.03	0.01	0.0020	1.18
	Treatment length	0.20	0.07	0.0055	1.22
	CBTgsh treatment	−2.68	0.82	0.0011	14.63
	EDE Weight Concern*CBTgsh	0.58	0.22	0.0083	2.66
	Treatment length*CBTgsh	−0.24	0.10	0.0159	5.13

Notes: Only the significant variables are included in the logistic regression model summaries of the parameter estimates. EDE = Eating Disorder Examination interview; Binge-eating remission defined as zero binge-eating (OBE) episodes during past month assessed with EDE; SE = standard error. Effect sizes were as follows: odds ratio (OR), interpreted as the group difference in odds of being in one treatment outcome category vs. the other (holding constant the other covariates), was used to estimate effect sizes (odds ratios around 1.3 are considered “small,” 1.5 are “medium” and 2.0 are “large”). For BMI, age, and length of treatment, which were modeled as continuous, groups for effect size purposes were defined by first identifying participants who were above vs. below the covariate median, then using the covariate mean within each category to define the distance between groups.

All data presented above are for intent-to-treat analyses with all randomized participants. Parallel exploratory analyses performed with treatment completers produced essential similar model parameters suggesting absence of biases due to differential dropout and are therefore not presented.