Carrick‐Sen 2014b.

Methods	Randomised trial of 162 women with a twin pregnancy; women recruited from Newcastle‐upon‐Tyne, UK, between October 2000 and March 2003.
Participants	162 women with a twin pregnancy; women booked for care prior to 20 weeks' gestation, with no known fetal anomalies.
Interventions	Women were randomised to 'standard' antenatal care (involving consultation with the woman's general practitioner, consultant obstetrician, community midwife, antenatal education sessions and breastfeeding workshop), or to a 'specialised' intervention group (consisting of the above, in addition to midwifery‐led antenatal and postnatal home visits, as well as an antenatal preparation for parenting programme).
Outcomes	The primary outcome of the trial was the incidence of depression at 6 months' postpartum.
Notes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer‐generated. Permuted block design.
Allocation concealment (selection bias)	Low risk	Web‐based allocation.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	For the study's primary outcome (postnatal depression at 6 months' postpartum), questionnaires were received from 133 (82%) of all trial participants, so less than 20% loss. 62% of women returned all questionnaires.
Selective reporting (reporting bias)	Low risk	All prespecified outcomes from the NHS R&D trial's register appear to have been reported upon in the report of the trial in the PhD thesis (postnatal depression, maternal anxiety, emotional well‐being, maternal satisfaction, parental stress) at all prespecified time points (6, 12, 26 and 52 weeks' postnatal).
Other bias	Unclear risk	"Randomisation was effective to provide two groups with the same baseline demographic and psychosocial characteristics. Acceptability and compliance with the intervention was excellent." "Women who declined participation were more likely to be multiparous, single and non‐Caucasian." Baseline characteristics were balanced, but multiparous, single non‐Caucasian women under‐represented. Women with poor command of English were excluded. Data for infant outcomes were not adjusted to take account of correlation between twins from the same pregnancy. In this version of the review we carried out a sensitivity analysis assuming high correlation of outcomes for twins.
Blinding of participants and personnel (performance bias) All outcomes	High risk	Described as an "open RCT" ‐ "Due to the visibility of the intervention the study design was based on an open RCT".
Blinding of outcome assessment (detection bias) All outcomes	High risk	Outcomes recorded in obstetric notes by staff aware of allocation and in postal questionnaires. Women in the intervention group were slightly more likely to return questionnaires.

RCT: randomised controlled trial