Analysis of outcome measures within treatment groupsBMJ 2012; 345 doi: https://doi.org/10.1136/bmj.e7201 (Published 26 October 2012) Cite this as: BMJ 2012;345:e7201
- Philip Sedgwick, reader in medical statistics and medical education
- 1Centre for Medical and Healthcare Education, St George’s, University of London, Tooting, London, UK
Researchers investigated the effects of a school based educational programme aimed at reducing consumption of carbonated drinks to prevent weight gain in children. The programme was delivered over one school year and focused on promoting a healthy diet. A cluster randomised controlled trial study design was used. Six primary schools in southwest England were recruited. The control group received no intervention. In total, 325 children aged 7-11 years were randomised to intervention and 319 to control.1
The main outcome measures included consumption of carbonated drinks. Consumption was measured as the number of glasses (average size 250 ml) drunk during a three day period, recorded both at baseline and at the end of the trial. The number of drinks consumed was normally distributed. Over one school year, the consumption of carbonated drinks decreased by a mean of 0.6 glasses in the intervention group (95% confidence interval −1.0 to −0.1; P=0.02) but increased by 0.2 glasses (−0.2 to 0.5; P=0.4) in the control group.
Which of the following statements, if any, are true?
a) Within each group, the change in consumption of carbonated drinks was investigated using the independent samples t test
b) Because the mean change in consumption of carbonated drinks for intervention was a significant reduction, it can be concluded that the education programme was effective
c) Because the 95% confidence interval for the mean change in drinks consumed for the two groups overlapped, the difference between intervention and control was not significant at the 5% level
Statements a, b …