Background The Disproportionate Intrauterine Growth Intervention Trial at Term (DIGITAT) compared induction with expectant management. In this secondary analysis we compared the outcomes of participants with non-participants.

Methods All patients who declined randomisation, but authorised data use were studied. Main outcomes; composite of perinatal death, 5-min Apgar score < 7, umbilical pH <7⋅05 or admission to neonatal intensive care; operative delivery. All comparisons were between participants and non participants, regardless of trial allocation or treatment received. Propensity scores were used to adjust for baseline differences between the randomised and non randomised women.

Results There were 650 randomised, and 452 non-participant women. Non-participants were older, smoked less, and had a lower BMI, and higher level of education. A total of 37 (6%) infants of participants experienced the composite adverse neonatal outcome, compared with 32 (8%) in the non-participants (difference -2% [95% CI -4%; 2%]). Caesarean sections were performed on 90 (14%) participants and 71 (16%) in the non-participants (-2% [95% CI -6%; 2%]). Among non-participants 3 (0.7%) deaths (2 stillbirths, 1 neonatal death) occurred. There were none in the randomised group.

Interpretation Despite better baseline health status there was a trend towards more adverse outcomes among non-particpants. This trial may be another example of the overall benefits of trial participation.

Funding ZonMw, the Netherlands, project number 945-04-558.