Intervalo de confiança para avaliação de bioequivalencia individual utilizando a medida de Kullback-Leibler

AUTOR(ES)
DATA DE PUBLICAÇÃO

2004

RESUMO

Bioequivalence trials are conducted to show that two formlllations, a reference dmg R and a test T, produces similar bioavailabitity. In Brazil, the introduction of the generic drug started in 1999 and the current analysis of the pharmacokinetics measures for rate and extent of the absorption is the average bioequivalence (ABE) which is not sufficient to guarantee that an individual present similar responses when sllbmitted to the two formulations. Alternatives methods as individual bioequivalence (IBE) and populational bioequivalence (PBE) were proposed in order to a measure the interchangeability, cla..c;sified as prescribability and switchability, between the reference and test formulations. Several studies for the assessment of individual bioeqllivalence between two formulations of a drug have been proposed recently. The study of Dragalin V. et alo (2003) proposed the Kullback-Leibler divergence as the mea..c;ure of the discrepancy between the distriblltions of the two formulations for assessing the individual bioequivalence. This study seems to have advantages sllch as the hierarchical property of BEI --. BEP --. BEM, satisfies the property of a tme metric distance, is invariant to monotonic transformations and can be generalized to multivariate setting and finally, can be applied to a wide range of distributions of the response. The assessment of individual bioequivalenee is made using a simulation study. Hyslop T. et alo ( 2000) proposed an alternative eonfidenee interval proeedure to assess IBE by the FDA eriteria, which uses the bootstrap-pereentil method. To ealculate the eonfidence interval, used the Howes approximation I method (1974) based on Cornish - Fisher expansion. Sinee the limits of the eonfidenee interval are based on the exact distributions of its eomponents, it ean be ea..<;ily ealeulated by available software. The advantages presented in Kullback-Leibler divergenee and the closed form of the upper limits of the eonfidence interval for the FDA eriteria motivated the present study. The present study foeus on the Cornish- Fisher expansion technique, as used by Hys lop(2000), for development of the procedure for the a..<;sessment of individual bioequivalenee using Kullba.ck-Leibler divergence as the measure of the diserepaney between the two formulations. Numerieal studies with three data sets ITom 2 x 4 crossover design and simulation studies for the power of the proposed alternative method for Kullbaek-Leibler divergenee are presented. Results of the Kullback-Leibler-confidence interval proposed indicate sensitive to ehanges in mean response between formulations and to the deteetion of subject-by-formulation interaction. The advantages of the proposed method are easy programming and no requirement of computational intensive method as require the bootstrap percentile method

ASSUNTO(S)

medicamentos planejamento de experiencias (estatistica)

ACESSO AO ARTIGO

Documentos Relacionados