Intended for healthcare professionals

Endgames Statistical Question

Understanding confidence intervals

BMJ 2014; 349 doi: (Published 07 October 2014) Cite this as: BMJ 2014;349:g6051
  1. Philip Sedgwick, reader in medical statistics and medical education1
  1. 1Institute for Medical and Biomedical Education, St George’s, University of London, London, UK
  1. p.sedgwick{at}

Researchers investigated the effects of controlled cord traction in the third stage of labour on postpartum haemorrhage when carried out in a high resource setting. A randomised controlled trial study design was used. Control treatment was standard placenta expulsion (awaiting spontaneous placental separation before facilitating expulsion). The setting was university hospital maternity units in France. Participants were women aged 18 or older with a singleton fetus at 35 or more weeks’ gestation and planned vaginal delivery. Women in both treatments groups received prophylactic oxytocin just after birth.1

The primary outcome was postpartum haemorrhage, defined as blood loss of 500 mL or more. Of 2005 women allocated to intervention, 196 (9.8%) experienced postpartum haemorrhage compared with 206 of 2008 allocated to control (10.3%). The reduction in risk of postpartum haemorrhage associated with intervention was not significant (relative risk 0.95; 95% confidence interval 0.79 to 1.15). The researchers concluded that in a high resource setting, the use of controlled cord traction for the management of placenta expulsion had no significant effect on the incidence of postpartum haemorrhage.

Which of the following statements, if any, are true?

a) The 95% confidence interval represents the inaccuracy of the sample in estimating the population parameter of the relative risk of postpartum haemorrhage.

b) If the sample size of the trial was increased, the width of the 95% confidence would decrease.

c) A 99% confidence interval for the population relative risk would be narrower than the 95% confidence interval presented.

d) It can be inferred that 95% of the intervention group lost between 0.79 and 1.15 times as much blood as the control group.


Statements a and b are true, whereas c and d are false.

The aim of the trial was to investigate the effects of controlled cord traction in the third stage of …

View Full Text

Log in

Log in through your institution


* For online subscription