Study designBMJ 2009; 339 doi: https://doi.org/10.1136/bmj.b4458 (Published 05 November 2009) Cite this as: BMJ 2009;339:b4458
- Philip Sedgwick, senior lecturer in medical statistics
- 1Centre for Medical and Healthcare Education, St George’s, University of London, Tooting, London SW17 0RE
To investigate whether physical activity in preschool children (under 5 years) reduced obesity, researchers recruited 545 children from 36 preschool nurseries. Each nursery was randomly allocated to an enhanced physical activity programme plus health education or no intervention (control group). The body mass, physical activity, and sedentary behaviour of each child in each nursery was recorded at regular intervals for 12 months.1
Which one of the following study designs best describes that used in the above study?
a) Case-control study
b) Double blind randomised controlled trial
c) Cluster randomised controlled trial
d) Nested case-control study
c—In a cluster randomised trial, natural clusters or groups of individuals are randomised, rather than individuals. Each preschool nursery was randomly allocated to the enhanced physical activity programme plus health education or no intervention. Within each nursery, children undertook the intervention to which the nursery was allocated. Cluster randomised trials overcome practical and contamination problems that may arise when individuals are randomised. For example, if individuals were randomised it may prove problematic to implement a physical activity programme within a nursery for some children and not others. Children allocated the physical activity programme may influence the activity of those allocated to the control group, or vice versa.
Answer a is false. A case-control study is both observational and retrospective in design. We select two groups of subjects on the basis of disease status, one with the disease (cases) and one without (controls). The aim of a case-control study is to identify potential risk factors for the disease. Past exposure to proposed risk factors is compared between the groups to ascertain if these factors are associated with an increased or reduced risk of the disease.
Answer b is false. In a double blind randomised controlled trial, the researchers and study participants do not know the treatments allocated. The researchers measuring the outcomes in the above trial would have been unaware of the intervention group to which nurseries were allocated, thereby reducing assessment bias. However, because of the nature of informed consent, the parents and probably the children themselves would have been aware of the potential interventions. The study was therefore unlikely to be double blind.
Answer d is false. A nested case-control study is observational by design. A cohort or group of people without the disease is indentified. The cohort is followed prospectively until an individual develops the disease of interest, when they become a “case.” Subsequently, for each case a fixed number of “controls” is selected from those remaining in the cohort—that is, those who have not developed the disease when it occurred in the case. Because the cases and controls are followed prospectively, nested case-control studies can reduce “recall bias” and any temporal ambiguity about the timing of risk factors and onset of disease. Compared with following the entire cohort, time and costs are lower, and the reduction in statistical efficiency is relatively small. Compare with the traditional case-control study described above that is retrospective in design.
Cite this as: BMJ 2009;339:b4458
Competing interests: None declared.
Log in using your username and password
Log in through your institution
Register for a free trial to thebmj.com to receive unlimited access to all content on thebmj.com for 14 days.
Sign up for a free trial