Academic Journals Database
Disseminating quality controlled scientific knowledge

Binomial Distribution Sample Confidence Intervals Estimation 10. Relative Risk Reduction and RRR-like Expressions

Author(s): Sorana BOLBOACĂ

Journal: Leonardo Electronic Journal of Practices and Technologies
ISSN 1583-1078

Volume: 4;
Issue: 6;
Start page: 60;
Date: 2005;
Original page

Keywords: Confidence Intervals Estimation | Relative Risk Reduction | Relative Risk Increase | Relative Benefit Increase

In trial, when the results are reported as dichotomies variables, the most important measure of effect are represented by the relative risk reduction, absolute risk reduction and number needed to treat, providing the basis for clinicians to balance the benefits and harms of therapy for their patients. The relative risk reduction is a very useful parameter in assessment of a treatment effect if it is accompanied by confidence intervals. The only method used in medical article for computing the confidence intervals for relative risk reduction is the asymptotic method. The aim the research was to propose some new methods of computing confidence intervals for relative risk reduction and relative risk reduction like parameters and to compare these methods with the asymptotic one in order to assess their performance.In order to estimate the confidence intervals for relative risk reduction we proposed based on the literature definitions and on our experiences in confidence intervals five methods called here ARPWald, ARPAC, ARPWaldC1, ARPWaldC2, and ARPWaldC3. The criterions of assessment were represented by the upper and lower boundaries, the average of experimental errors and standard deviations, and the deviation relative to imposed significance level α equal with 5%. All methods were assessed on random variables (X, Y) and random samples sizes (n, m).Chousing a method of computing confidence intervals for relative risk reduction (RRR) and RRR-like functions depend on objective of research. If we desire a method which to obtain performance in estimating the average of the experimental errors we can chouse the ARPWald method, while if we need a method with the smallest deviation we will chouse the ARPAC method.
     40% off discount for PC Health Advisor