Intended for healthcare professionals

Endgames Statistical Question

Bias in experimental study designs: randomised controlled trials with parallel groups

BMJ 2015; 351 doi: (Published 17 July 2015) Cite this as: BMJ 2015;351:h3869
  1. Philip Sedgwick, reader in medical statistics and medical education
  1. 1Institute for Medical and Biomedical Education, St George’s, University of London, London, UK
  1. Correspondence to: P Sedgwick p.sedgwick{at}

Researchers investigated the effects of antenatal lifestyle advice on health outcomes in overweight and obese pregnant women. A randomised controlled trial with a parallel groups study design was used. The intervention consisted of a comprehensive dietary and lifestyle programme during pregnancy, which incorporated a combination of dietary, exercise, and behavioural strategies delivered by a research dietitian and trained research assistants. The control treatment was standard pregnancy care, which did not include such strategies. Participants were women with a singleton pregnancy, between 10 and 20 weeks’ gestation, and who had a body mass index (BMI) of 25 or more. The setting was three public maternity hospitals across South Australia.1

In total, 5474 eligible women were approached, of whom 2212 accepted the invitation and were randomly allocated to the intervention (n=1108) or control (n=1104) treatment. To ensure allocation concealment, a computer generated allocation sequence was delivered by a telephone randomisation service not involved with recruitment or clinical care. The primary outcome was the proportion of infants born large for gestational age (birth weight ≥90th centile for gestation and sex). Secondary outcomes included birth weight greater than 4000 g, hypertension, pre-eclampsia, and gestational diabetes. All outcomes were assessed by research staff blinded to the original treatment allocation.

Although the risk of infants being born large for gestational age was lower in the intervention group than in the control group, the difference was not significant (18.9% v 21%; adjusted relative risk 0.90, 95% confidence interval 0.77 to 1.07; P=0.24). Infants born to women after lifestyle advice were significantly less likely to have a birth weight above 4000 g (15.3% v 18.8%; 0.82, 0.68 to 0.99; P=0.04). There were no differences between the treatment groups for any of the other secondary outcomes. It was concluded that, in overweight or obese women, antenatal lifestyle advice …

View Full Text

Log in

Log in through your institution


* For online subscription