The use of interval estimators as a basis for decision-making in medicine

✍ Scribed by Reidar K. Lie

• Publisher: Springer Netherlands
• Year: 1984
• Tongue: English
• Weight: 506 KB
• Volume: 5
• Category: Article
• ISSN: 1573-1200
• DOI: 10.1007/bf00489472

✦ Synopsis

Decision analysts sometimes use the results of clinical trials in order to evaluate treatment alternatives. I discuss some problems associated with this, and in particular I point out that it is not valid to use the estimates from clinical trials as the probabilities of events which are needed for decision analysis. I also attempt to show that an approach based on objective statistical theory may have advantages over commonly used methods based on decision theory. These advantages include the recognition of uncertain data, the introduction of a third alternative, namely suspension of judgement, and the possibility of modifying the choice of probabilities based on a clinical trial with reference to other available knowledge. I have not, however, shown in detail how this modification is done, but I think the concept is sufficiently promising to be applied to an actual clinical decision problem.