Missing covariate values are a common problem in survival studies, and the method of choice when handling such incomplete data is often multiple imputation. However, it is not obvious how this can be used most effectively when an incomplete covariate is a function of other covariates. For example, body mass index (BMI) is the ratio of weight and height-squared. In this situation, the following question arises: Should a composite covariate such as BMI be imputed directly, or is it advantageous to impute its constituents, weight and height, first and to construct BMI afterwards? We address this question through a carefully designed simulation study that compares various approaches to multiple imputation of composite covariates in a survival context. We discuss advantages and limitations of these approaches for various types of missingness and imputation models. Our results are a first step towards providing much needed guidance to practitioners for analysing their incomplete survival data effectively.
13 September 2012 in Journal of Applied Statistics
We decompose the score statistic for testing for shared finite variance frailty in multivariate lifetime data into marginal and covariance-based terms. The null properties of the covariance-based statistic are derived in the context of parametric lifetime models. Its non-null properties are estimated using simulation and compared with those of the score test and two likelihood ratio tests when the underlying lifetime distribution is Weibull. Some examples are used to illustrate the covariance-based test. A case is made for using the covariance-based statistic as a simple diagnostic procedure for shared frailty in a parametric exploratory analysis of multivariate lifetime data and a link to the bivariate Clayton–Oakes copula model is shown.