We found a match
Your institution may have access to this item. Find your institution then sign in to continue.
- Title
USO DE MODELOS LINEALES MIXTOS PARA LA ESTIMACIÓN DEL TIEMPO DE VIDA ÚTIL DE UN MEDICAMENTO.
- Authors
PAZMIÑO, Jhonathan; GONZALEZ, Nelfi
- Abstract
Background: Statistical reliability has made significant advances in the modeling of data from stability studies for different equipment and products, including home appliances, car components and electrical circuits. However, even considering the complexity of the pharmaceutical stability tests and the socioeco-nomic impact of the definition of the drug shelf-life, its permeation into the pharmaceutical area has not been promoted. Additionally, it has been found that due to the manufacturing processes of batch drugs, variation effect associated to them should be considered when performing the statistical analysis of stability data, something that the Colombian guide does not take into account. Objectives: This study is aimed to show the convenience of using an alternative analysis method to analyze the information derived from an accelerated stability testing of pharmaceutical products. Methods: Statistical simulation was used to compare two proposed shelf-life estimators based on linear mixed models versus another existing estimator based on simple linear regression models called classical estimator, being the latter in accordance with the Colombian methodology. The effect that the inherent variability in the manufacturing process of a drug could have was contemplated, considering the variability between and intra batch. In addition to this, different levels of the factors that the researcher can control in the stability test such as number of lots, number of observation times and number of samples per batch and time were also used. Results: It was determined that the two proposed estimators provide estimates with lesser degrees of bias than those found by the classical estimator; in addition, the behavior of the estimates based on the intrinsic variance components of the study and the effect of experimental factors were presented. Conclusions: The estimates made using linear mixed models have greater accuracy in comparison with those made in the traditional manner; this decrease in bias could improve the productivity of pharmaceutical companies without compromising public health.
- Publication
Vitae (01214004), 2017, Vol 24, p80
- ISSN
0121-4004
- Publication type
Article
- DOI
10.17533/udea.vitae.v24n2(2)a09