The Probability of success (PoS) of a trial is conventionally defined as the expected value of the power function of a test with respect to a design prior assigned to the parameter under scrutiny. Even though this quantity is widely used in clinical trials for experimental design, the definition of probability of success is not univocal. In this presentation we review and compare the main types of PoS; specifically, we focus on a unifying, decision-theoretic approach that yields a new type of PoS as the expected utility of the trial (u-PoS), that is the expected probability of making the correct choice between two hypotheses. We base our comparisons on properties of the probability distributions of the power-related random variables associated to these definitions. Our conclusion is that u-PoS shows a conceptual advantage over previous versions of PoS; moreover, when suitable design priors are used, it may produce smaller optimal sample sizes with respect to its competitors.
7 Novembre 2025, ore 12
Fulvio De Santis
Sapienza University of Rome
In person: Room 34 (4th floor) building CU002 Scienze Statistiche
Webinar: https://uniroma1.zoom.us/j/83625004899?pwd=bXCtz0mp759PUh2lkqT0BUoVa0Uegg.1
ID riunione: 836 2500 4899
Passcode: 123456