A New Extended Intestinal Release Formulation of Rifaximin, 400 mg Tablets, for the Treatment of Moderately Active Crohn's Disease
| Autor: | Cosimo Prantera, Maria Grimaldi, Silvio Danese, Paolo Gionchetti, Maria Vittoria Fogli, Herbert Lochs |
|---|---|
| Rok vydání: | 2011 |
| Předmět: | |
| Zdroj: | Gastroenterology. 140:S-572 |
| ISSN: | 0016-5085 1728-1733 |
| DOI: | 10.1016/s0016-5085(11)62370-6 |
| Popis: | Introduction. Results from clinical trials often employ the odds ratio or relative risk to compare the proportion of subjects responding to experimental treatment to the proportion responding to control treatment. These measures have the advantage of using a single index to summarize the study results, but neither measure provides information that is intuitively meaningful to clinicians, and both can vary in different ways depending on the response to control treatment. Methods. We calculated the odds ratio and relative risk for responses to control treatment varying from 5% to 75% with a fixed difference of 20% between experimental and control responses. We then compared these values to those for two measures of effect size: number-needed-to-treat (NNT; N Engl J Med 1988; 318:1728-1733) and Probabilistic Index (Statist Med 2006; 25:591-602). Results. The Figure illustrates that with a fixed difference between responses to control and experimental treatment, as the response to control treatment increases, the odds ratio decreases to a plateau and then increases, whereas the relative risk decreases progressively. Since NNT and Probabilistic Index depend only on the difference between response to control and experimental treatment, values for these measures do not change with changing response to control treatment. That is, for a 20% difference between control and experimental responses, the NNT of 5 gives the number of experimental subjects needed to obtain one more success than treating the same number of control subjects. The Probabilistic Index for these same data is 0.6, and gives the probability that a randomly selected experimental subject has an outcome preferable to that from a randomly selected control subject. Conclusion. The NNT and Probabilistic Index depend only on the difference between responses to control and experimental treatment and can help clinicians assess the potential clinical significance of different values of the odds ratio and relative risk. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|