Schubiger 1997.

Methods	Multicentre prospective randomised trial (from 10 centres).
Participants	A total of 602 healthy full‐term infants (> 37 weeks of gestation, birthweight 2750 g to 4200 g) of mothers who intended to stay in the hospital for 5 days postpartum and planned to breastfeed for more than 3 months. The study did not state whether twins were included.
Interventions	UNICEF group (n = 294): "bottles, teats and pacifiers were strictly forbidden"; "supplements if medically indicated were administered by cup or spoon". Standard group (n = 308): "pacifiers were offered to all infants without restriction. Supplements were conventionally offered by bottle after breastfeeding". In both groups, the fluid supplements during the first few days consisted of a 10% dextrin‐maltose solution. Fluid supplements were considered to be medically indicated in the following situations: babies agitated or screaming after breastfeeding; signs of dehydration (no urine output over 4 hours after day 1); symptoms of hypoglycaemia with blood glucose < 2 mmol/L. In the standard group fluids were more liberally offered. About 180 participants in the UNICEF group and 291 participants in the standard group completed the protocol. Almost 40% of the participants in the UNICEF group violated protocol during the first 5 days in the hospital. Upon discharge from the hospital, it was left to the mothers of both groups to decide whether to use a pacifier and/or bottle.
Outcomes	Incidence of breastfeeding at day 5, and at 2, 4, 6 months, proportion of fully or partially breastfeeding on day 5, sucking behaviour (good, mediocre, insufficient), incidence of fever, incidence of phototherapy. Questionaires administered to mothers at 2, 4, and 6 months were used to collect breastfeeding outcomes after hospital discharge.
Notes	Study conducted in Switzerland. Results were reported in 2 separate publications with slight differences in the presentation of results. This study however was not included for analysis due to high attrition bias (almost 40% loss of participants in the intervention group) due to protocol violation in the first weeks of the study.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	"Sealed protocol forms were centrally randomised." Comment: The method of random sequence generation is not described.
Allocation concealment (selection bias)	Low risk	"Sealed protocol forms were centrally randomised." Comment: Allocation concealment incompletely described but likely to have been present.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Randomising participants in the same room or ward rather than comparing routines of one ward to another ‐ not feasible to blind participants.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Blinding of clinician and outcome assessor is not described. However it is unlikely that they were blinded to the intervention.
Incomplete outcome data (attrition bias) All outcomes	High risk	The rate protocol violators is approximately 15% and 5.5% respectively and the rate of lost to follow‐up is 7.8% and 4.2%. Thus, the total dropout rate is 22% versus 9.7%, respectively after 70 protocol violaters due to pacifier use included into the analysis. The other protocol violaters were due to bottle feeding, failure to spoon/cup feed, early discharge and others.
Selective reporting (reporting bias)	Low risk	Protocol not available. All expected outcomes reported.
Other bias	High risk	Primary outcome data have to be imputed from percentages and exact denominators at 4 and 6 months follow‐up are unclear.

RCT: randomised controlled trial