Intended for healthcare professionals

Endgames Statistical Question

One way analysis of variance: post hoc testing

BMJ 2014; 349 doi: (Published 21 November 2014) Cite this as: BMJ 2014;349:g7067
  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 compared pregnancy outcomes between women who stopped smoking in early pregnancy and those who either did not smoke or continued to smoke in pregnancy. A prospective cohort study was performed. Participants were 2504 healthy nulliparous women with singleton pregnancies. The primary outcomes were spontaneous preterm birth and small for gestational age infants (birth weight <10th customised centile).1

The participants were grouped by smoking status at 15 weeks’ gestation. Of the women recruited, 80% (n=1992) were non-smokers, 10% (n=261) had stopped smoking, and 10% (n=251) were current smokers. The women’s characteristics at 15 weeks’ gestation were compared between the three smoking status groups to establish differences that might have influenced the primary outcomes. A significant difference was reported between the smoking status groups in mean age at 15 weeks’ gestation (non-smokers 29.7 years (standard deviation 5.1), those who had stopped smoking 25.2 (5.9), current smokers 23.1 (5.5); one way analysis of variance P<0.001). Tukey’s post hoc test for pairwise comparisons showed that a significant difference (P<0.05) existed between each pair of groups—that is, between non-smokers and current smokers, non-smokers and stopped smokers, as well as current smokers and stopped smokers.

It was reported that for the group of women who stopped smoking before 15 weeks’ gestation, rates of spontaneous preterm birth and small for gestational age infants did not differ significantly from those in non-smokers. The rates of spontaneous preterm birth and small for gestational age infants for the current smokers group were significantly higher than for the group that stopped smoking before 15 weeks’ gestation. It was concluded that the …

View Full Text

Log in

Log in through your institution


* For online subscription