Advances and Applications in Statistics
Volume 57, Issue 1, Pages 1 - 20
(July 2019) http://dx.doi.org/10.17654/AS057010001 |
|
SENSITIVITY ANALYSIS INDEX FOR SHARED PARAMETER MODELS IN LONGITUDINAL STUDIES
Heba A. Abd-ElWahab, Rasha B. Elkholy and Ahmed M. Gad
|
Abstract: In longitudinal studies the response variable(s) are measured for the same subject on several different occasions. Missing data arise when the intended measurement for an individual could not be obtained for any reason. The missing data create a major challenge for data analysis. Methods for handling missing data often depend on the pattern of missingness (monotone and non-monotone pattern) and the mechanism that generates the missing values (missing completely at random, missing at random, and missing not at random). Models, such as selection models, pattern mixture models and shared parameter models, are used to model longitudinal data with non-ignorable missingness. These models are based on assumptions about the missing data mechanism that cannot be verified from the observed data. An appropriate approach to the problem of the unverifiable assumptions is to conduct a sensitivity analysis. One of the tools of the local sensitivity is the Index of Sensitivity to Non-Ignorability (ISNI). This index measures the local sensitivity of the parameter estimate to departures from ignorability. Most of the studies in the literature derived the ISNI for selection models. This study proposes an extension to ISNI in the context of joint modeling of the longitudinal and the missing data in shared parameter models framework. The formula of the extended ISNI is derived and a simulation study is conducted to investigate the sensitivity of the model parameter estimates to missing data assumptions using the ISNI. |
Keywords and phrases: non-ignorable missing data, sensitivity analysis, sensitivity index, shared parameter models.
|
|
Number of Downloads: 358 | Number of Views: 1906 |
|