When you have missing data, you can use the MISSING= option in the PROC BGLIMM statement as follows to specify how you want to handle the missing response values:[30]
If you specify MISSING=CC (CC stands for complete cases), PROC BGLIMM discards all observations that have missing or partial missing values (for example, in a repeated measures model) before carrying out the simulation. This is equivalent to assuming that the missing values are missing completely at random (MCAR).
If you specify MISSING=CCMODELY, PROC BGLIMM treats missing response values as parameters and includes the sampling of the missing data as part of the simulation. The procedure discards all observations that have missing covariates. This is equivalent to assuming that the missing values are missing at random (MAR).
Different types of missing data were first defined by Rubin (1976). For a comprehensive treatment of missing data analysis, see Little and Rubin (2002). PROC BGLIMM does not model the missing not at random (MNAR) type of missing data.
[30] A missing value is usually, although not necessarily, represented by a single period (.) in the input data set.