TABLE 3.

MIXED MODELS ANALYSES OF BEHAVIORAL OUTCOMES IN THE INTENT-TO-TREAT AND EVALUABLE GROUPS AMONG WOMEN WITH ILLNESS ONSET WITHIN FOUR WEEKS OF GIVING BIRTH

	ITT (n = 36)		Evaluable (n = 33)
HAM-D	F(1,104)=6.98, p=0.01	β=1.18, SE=0.45	F(1,104)=6.87, p=0.01	β=1.18, SE=0.45
HAM-A	F(1,87)=4.92, p=0.03	β=1.19, SE=0.54	F(1,87)=5.08, p=0.03	β=1.22, SE=0.54
EPDS	F(1,98)=4.16, p=0.04	β=0.91, SE=0.45	F(1,98)=4.16, p=0.04	β=0.91, SE=0.45

Legend: F-statistics and p-values, betas and standard errors (SE) from mixed models analyses. HAM-D - Hamilton Depression Rating Scale; HAM-A - Hamilton Anxiety Rating Scale; EPDS - Edinburgh Postnatal Depression Scale. Independent variables were treatment, time (linear and quadratic term), treatment by time interaction and the corresponding baseline measure as a covariate. That is, baseline HAM-D was used as a covariate for analyses of HAM-D, baseline HAM-A was used as a covariate for analyses of HAM-A and baseline EPDS was used as a covariate for analyses of EPDS) .