Lilford RJ, Edwards SJ, Braunholtz DA; International Society of Technology Assessment in Health Care. Meeting.
Annu Meet Int Soc Technol Assess Health Care Int Soc Technol Assess Health Care Meet. 1999; 15: 42.
Department of Public Health and Epidemiology, The University of Birmingham, England.
The received wisdom concerning the analysis of interim data coming from clinical trials is that its should be done in secret, primarily so that any premature change in practice is prevented. A sufficiently precise measurement of therapeutic effectiveness can thus be achieved. However, the empirical evidence of whether or not unfettered dissemination of interim results would, in fact, have this effect is neither convincing nor consistent. We will present arguments to show that making interim results publicly accessible need not slow recruitment to the point that sufficient statistical precision is unlikely to be achieved. In particular, because patients assign different values to clinical outcomes and because their doctors have different estimates of what treatments are likely to achieve, some patients will want to participate in a trial on the basis of interim data while some patients will want to drop out. Following the 'expected utility' rule, patients should want to participate in a trial when they are in a state of 'equipoise,' i.e., when the expected utility of the trial aims are equal to, or better than, all available treatment alternatives. So, as some patients will be moving out of equipoise so some will be moving into equipoise. We will illustrate this particular argument with a real example, approaching the analysis from a Bayesian way of thinking. We will argue, moreover, that Data Monitoring Committees, which currently keep patients and their doctors in the dark by withholding interim data, are behaving unethically whatever the effects on recruitment rates to trials. That said, the duties, bestowed on members of Data Monitoring Committees, often involve pitting the needs of trial patients and those who might want to participate on the basis of interim data against the needs of future patients and society at large. These duties are not easily discharged. Given that knowledge is increasingly seen as a continuous variable rather than a dichotomous outcome, the current activities of Data Monitoring Committees are unethical and politically unsustainable.
Publication Types:
Keywords:
- Clinical Trials Data Monitoring Committees
- Clinical Trials as Topic
- Humans
- Research Design
- statistics & numerical data
- hsrmtgs
Other ID:
UI: 102193908
From Meeting Abstracts