Search

Modified Poisson regression, which estimates the regression parameters in the log-binomial regression model using the Poisson quasi-likelihood estimating equation and robust variance, is a useful tool for estimating the adjusted risk and prevalence ratio in binary outcome analysis. Although several goodness-of-fit tests have been developed for other binary regressions, few goodness-of-fit tests are available for modified Poisson regression. In this study, we proposed several goodness-of-fit tests for modified Poisson regression, including the modified Hosmer-Lemeshow test with empirical variance, Tsiatis test, normalized Pearson chi-square tests with binomial variance and Poisson variance, and normalized residual sum of squares test. The original Hosmer-Lemeshow test and normalized Pearson chi-square test with binomial variance are inappropriate for the modified Poisson regression, which can produce a fitted value exceeding 1 owing to the unconstrained parameter space. A simulation study revealed that the normalized residual sum of squares test performed well regarding the type I error probability and the power for a wrong link function. We applied the proposed goodness-of-fit tests to the analysis of cross-sectional data of patients with cancer. We recommend the normalized residual sum of squares test as a goodness-of-fit test in the modified Poisson regression. [ABSTRACT FROM AUTHOR]

In clinical and epidemiological studies using survival analysis, some explanatory variables are often missing. When this occurs, multiple imputation (MI) is frequently used in practice. In many cases, simple parametric imputation models are routinely adopted without checking the validity of the model specification. Misspecified imputation models can cause biased parameter estimates. In this study, we describe novel frequentist type MI procedures for survival analysis using proportional and additive hazards models. The procedures are based on non-parametric estimation techniques and do not require the correct specification of parametric imputation models. For continuous missing covariates, we first sample imputation values from a parametric imputation model. Then, we obtain estimates by solving the estimating equation modified by non-parametrically estimated conditional densities. For categorical missing covariates, we directly sample imputation values from a non-parametrically estimated conditional distribution and then obtain estimates by solving the corresponding estimating equation. We evaluate the performance of the proposed procedures using simulation studies: one uses simulated data; another uses data informed by parameters generated from a real-world medical claims database. We also applied the procedures to a pharmacoepidemiological study that examined the effect of antihyperlipidemics on hyperglycemia incidence. [ABSTRACT FROM AUTHOR]