*------------------------------------------------------------------*
| Purpose: To identify outliers based on n standard deviations     |
|          from the mean.                                          |
| Arguments: DSN    - Data set name                                |
|            VAR    - Numeric variable to be checked               |
|            IDVAR  - ID variable name                             |
|            N_SD   - The number of standard deviation units for   |
|                     declaring an outlier                         |
|                                                                  |
| Example: %SD_ALL(CLEAN.PATIENTS,HR,PATNO,2)                      |
*------------------------------------------------------------------*;
%MACRO SD_ALL(DSN,VAR,IDVAR,N_SD);

   TITLE1 "Outliers for Variable &VAR Data Set &DSN";
   TITLE2 "Based on &N_SD Standard Deviations";

   PROC MEANS DATA=&DSN NOPRINT;
   VAR &VAR ;
   OUTPUT OUT=MEANS(DROP=_TYPE_ _FREQ_)
          MEAN=M
          STD=S;
   RUN;

   DATA _NULL_;
      FILE PRINT;
      SET &DSN;
      IF _N_ = 1 THEN SET MEANS;
         IF &VAR LT M - &N_SD*S AND &VAR NE .
         OR &VAR GT M + &N_SD*S THEN PUT &IDVAR= &VAR=;
   RUN;

   PROC DATASETS LIBRARY=WORK NOLIST;
      DELETE MEANS;
   RUN;
   QUIT;

%MEND SD_ALL;