Data missing, which occurs for different reasons, is an unavoidable problem in epidemiological studies. It is quite widespread and, therefore, it is considered as a challenge in research design and data analysis by many methodologists. Complete case analysis is often used in studies with missing data; however, this approach may result in inaccurate estimates and inferences due to bias associated with missing data. It is possible to obtain more accurate results through data repairing methods. Ordinary, simple strategies, such as single imputation methods, have drawbacks and limitations in practice. It has, however, been demonstrated that using modern imputation techniques can, despite their complexity, reduce bias dramatically in many situations, if used appropriately and properly. In this review, application of multiple imputations, as a novel technique for handling missing data in health and epidemiological research, is briefly discussed