Patients must be told that treatment will be randomisedBMJ 1995; 310 doi: https://doi.org/10.1136/bmj.310.6972.125c (Published 14 January 1995) Cite this as: BMJ 1995;310:125
- James Gilbert
- Macmillan consultant in palliative medicine University of Exeter, Exeter and District Hospice, Exeter EX2 5JJ
EDITOR,—Henry McQuay and Andrew Moore refer to equipoise being the requirement for an ethically sound randomised controlled trial.1 Although equipoise is a necessary condition, it is not sufficient. In addition, those whose participation is sought in a randomised controlled trial should be informed about all relevant features of that trial, an important one of which is the randomisation process. For an ethically sound trial of the sort reported by Ian R McWhinney and colleagues2 it would be necessary to recruit only those patients who were genuinely ambivalent (or were prepared, knowingly, to forgo their choice) about whether to receive palliative care support immediately1 or after a delay of one month. Although the figure is not explicitly stated in the report, it seems unlikely that fewer than 20 of the 166 eligible patients refused to participate in the trial on this basis.
The moral and practical difficulties of informed consent in medical research frequently hinge on disclosure of the fact of randomisation. When the interventions being compared are identical in form (such as active and placebo drugs presented as indistinguishable tablets) genuine ambivalence is much more likely.
When dissimilar treatments—for instance, surgery and radiotherapy—are being compared many people will have an innate preference based on personal considerations rather than scientific evidence. In such circumstances there may be genuine clinical equipoise but the right ofautonomous choice should not be removed by failure to disclose the fact of randomisation.
While many individual interventions in palliative care may feasibly be subject to randomised controlled trials, the assessment of a whole service by such means seems a tall order indeed.