Table 2 Test of EPDS score at baseline as a modifier of the intervention effect on percentage with sleep problems and mean EPDS scores by randomisation group
OutcomeEPDS baseline scoreGroupAdjusted comparative statistic (intervention relative to control)p Value for interaction
InterventionnControln
Sleep
    Sleep problem at 10 months, n (%)>934 (54)6343 (72)600.34 (0.14, 0.81)0.24
⩽958 (58)10057 (66)870.71 (0.38, 1.32)
    Sleep problem at 12 months, n (%)>930 (47)6436 (62)580.49 (0.22, 1.07)0.88
⩽934 (34)9944 (50)880.48 (0.26, 0.90)
EPDS score
    Total EPDS score at 10 months, mean (SD)>99.5 (5.6)6311.8 (5.0)60−2.1 (−3.9, −0.2)0.05
⩽95.1 (4.0)1005.1 (4.1)87−0.2 (−1.4, 0.9)
    Total EPDS score at 12 months, mean (SD)>98.9 (4.8)6410.5 (4.6)58−1.6 (−3.2, 0.1)0.56
⩽94.0 (3.8)995.0 (4.3)88−1.3 (−2.6, 0.1)
  • The comparative statistic is odds ratio for sleep problem and mean difference for EPDS score. EPDS, Edinburgh Postnatal Depression Scale.

  • All analyses adjusted for whether a mother received paid professional help for the infant’s sleep problem and level of nurse competency. The test of interaction for the sleep problem outcome was additionally adjusted for Socio-Economic Indexes For Areas (SEIFA) Index of Relative Disadvantage code and mother’s education.