, Volume 15, Issue 7, pp 1447-1453
Date: 11 Aug 2011

Medication Adherence: Tailoring the Analysis to the Data

Rent the article at a discount

Rent now

* Final gross prices may vary according to local VAT.

Get Access


The purpose of this paper is to explore more comprehensive methods to analyze antiretroviral non-adherence data. Using illustrative data and simulations, we investigated the value of using binary logistic regression (LR; dichotomized at 0% non-adherence) versus a hurdle model (combination of LR plus generalized linear model for >0% non-adherence) versus a zero-inflated negative binomial (ZINB) model (simultaneously modeling 0% non-adherence and >0% non-adherence). In simulation studies, the hurdle and ZINB models had similar power but both had higher power in comparison to LR alone. The hurdle model had higher power than ZINB in settings where covariate effects were restricted to one or the other part of the model (0% non-adherence or degree of non-adherence). Use of the hurdle and ZINB models are powerful and valuable approaches in analyzing adherence data which yield a more complete picture than LR alone. We recommend adoption of this methodology for future antiretroviral adherence research.


El objetivo de este trabajo es explorar de manera exhaustiva los métodos de análisis de los datos de no adherencia antiretroviral. Utilizando simulaciones y datos representativos, hemos investigado la utilidad de usar la regresión logística binaria (LR; dicotónica con no adherencia del 0%) versus un modelo de hurdle (combinación de LR más el modelo lineal generalizado para no adherencia >0%) versus un modelo de zero-inflated negative binomial (ZINB). En estudios de simulación, los modelos hurdle y ZINB han tenido potencia similar, pero ambos han tenido mayor potencia que el uso exclusivo de LR. El modelo de hurdle ha tenido mayor potencia que ZINB en configuraciones en las que los efectos de las co-variables se limitaban a una u otra parte del modelo (no adherencia del 0% o algún grado de no adherencia). El uso del modelos hurdle y de ZINB constituye un valioso y efectivo enfoque para analizar los datos de no adherencia, y ofrecerá un panorama más completo que el uso exclusivo de LR. Recomendamos que se adopte esta metodología en investigaciones futuras sobre la adherencia antiretroviral.