|
Abstract:
|
Selecting an appropriate sample size represents a choice that leads to real trade-offs. Smaller sample sizes mean lower costs and quicker results, but they run the risk of not producing a significant result. In contrast, larger sample sizes mean longer studies, higher costs, and potentially costly delays in getting lifesaving interventions into mainstream use. In spite of an illusion of objectivity in most statistical analyses of clinical trials, subjectivity exists in any trial. This talk presents a value-of-information approach to data collection and emphasizes the need for statisticians to recognize the values that exist under seemingly value-free calculations. The talk discusses the challenges to broader implementation of value-of-information approaches and provides insights from prior studies.
|
