Randomising groups of patientsBMJ 1998; 316 doi: https://doi.org/10.1136/bmj.316.7148.1898 (Published 20 June 1998) Cite this as: BMJ 1998;316:1898
- Chris Roberts, senior research fellow in statistics,
- Bonnie Sibbald, reader in health services research
- National Primary Care Research and Development Centre, University of Manchester, Manchester M13 9PL
- Correspondence to: Dr Roberts.
This is the sixth of an occasional series on the methods of randomised controlled trials
In a simple randomised trial each patient is assigned independently to a treatment group. Sometimes, however, randomising individual patients is either technically impossible or may compromise the evaluation.
Some interventions can be delivered only to groups. Public health initiatives, for example, may be delivered to communities rather than to individuals. When an intervention entails changing the organisation of health care in a unit all patients attending that unit will be affected. Similarly, when interventions are targeted at changing the behaviour of health professionals — for example, a training package — the care given to all subsequent patients may be modified once staff have been exposed to the intervention. In both situations it would be difficult or impossible for staff to revert …