Missing evaluations often occur in studies on terminal patients for various reasons but mostly because of patients' poor health conditions. The main potential consequences of missing data are the loss of power of the study to demonstrate what it was designed for, and the bias of the estimates performed on complete cases. Some tips regarding standards of data collection aimed at improving compliance will be given. Terminology definition will cover the meaning of "attrition", "non compliance", "missing", "dropout" and "censoring" so as the distinction between "missing items" and "missing forms". Understanding the missing data mechanism is important for the analysis and interpretation of the study results. For this reason missing data will be classified as Missing Completely At Random (MCAR), Missing At Random (MAR) and Non Missing At Random (NMAR) depending upon the reasons which lead to missing assessment and the potential relationship between this reason and the missing value.