Throughout the last couple of years, multiple imputation has become a popular and widely accepted technique to handle missing data. Although various multiple imputation procedures have been implemented in all major statistical packages, currently available software is still highly limited regarding the imputation of incomplete count data. As count data analysis typically makes it necessary to fit statistical models that are suited for count data like Poisson or negative binomial models, also imputation procedures should be specially tailored to the statistical specialities of count data. We present a flexible and easy-to-use solution to create multiple imputations of incomplete multilevel count data, based on a generalized linear mixed-effects Poisson model with multivariate normal random effects, using penalized quasi-likelihood. Our procedure works as an add-on for the popular and powerful multiple imputation package MICE (van Buuren & Groothuis-Oudshoorn, 2011) for the R language and environment for statistical computing

Missing data pose a threat to the validity of statistical inferences when they are numerous and not missing completely at random (Little & Rubin, 1987; Schafer, 1997a). In this case, multiple imputation (MI; e.g. Schafer, 1997a; van Buuren & Groothuis-Oudshoorn, 2011) is a state-of-the-art procedure to address the missing data problem, as it allows the use of all available information in the dataset to predict missing information. In comparison to simple solutions like case deletion, unconditional mean imputation, or last value carried forward, MI typically yields far better and mostly unbiased parameter estimates and standard errors (Schafer & Graham, 2002; Graham, 2009; Kleinke, Stemmler, Reinecke, & Lösel, 2011).