survan.doc

完整的生存分析Fortran语言源代码。有5个大的功能模块。

                       ===== End Commentary =====

Please specify the desired method of data entry:
    1) Survival times entered by groups from the terminal
    2) Free field -- from terminal or ascii files
? 2
                      ===== Begin Commentary =====

      The user can input the data in two ways. The first option allows
the user to  be  prompted for the  data by groups.   The second option
allows  the  data to be  entered  free  field, i.e.,  spaces  separate
variables.  If the data is in a file (as is usual) choose option two.

                       ===== End Commentary =====

Please select one of the following options
     1  (*) Variable names are V1 ... Vn
     2  ( ) Variable names typed in from terminal
     3  ( ) Variable names read from a file
? 2





                      ===== Begin Commentary =====

    The  suggested  default   (option 1)   is  that the  names of  the
variables read will be called  v1, v2,  ..., vn.  This  saves the user
the bother  of  entering his names.   However,  she  may want  to give
mnemonic names to the variables in which case options 2 or 3 should be
chosen.  Option  2 allows names  to be entered interactively; option 3
allows names to be written to a file which is read  to find the names.
Names  consist  of  1-8   alphanumeric  characters  and  are not  case
sensitive (i.e., NaMe1 is the same as name1).
    Note that  if the  user  specifies a set  of  names, the number of
names determines the number of variables to be read.  If the user does
not specify the names, then  he has the choice  of entering the number
of variables  to be  read or  of having it  determined  from the first
record of the data  file which  must contain  all observations  from a
complete case.

                       ===== End Commentary =====

Enter the names of the variables in the observations.
You must enter the names in the order in which the
values appear in the observations file.
Enter a blank line after all names have been entered.
? surv status group
?
Please select one of the following options.
     1 ( ) Abort if number of values on a record doesn't match
             the number of variables
     2 (*) Do not abort
? 1
                      ===== Begin Commentary =====

     Values of  variables   are  separated by  blanks   in  the SURVAN
program.   The number of variables in  a record  is determined by  the
number of names  entered, or  by the number  of  values  on the  first
record, or by the number explicitly specified (see above).

     This option  is useful in  the case  in which all values for  one
record fit on one line.  If option 1 is chosen, SURVAN will abort with
an error message if  there  are  too few  or  too many variables on  a
particular line.  Too few or too many  is usually an  input error.  If
option 1 is  not chosen,  the   specified number of variables  is read
sequentially for each record ignoring input line boundaries.

                      ===== End Commentary =====

Enter the name of the file with the data
Specify file name or * for terminal input:
? sexample.dat
Opened file sexample.dat --
   -- P(roceed) or R(etry file specification)?
? p
Do you want to specify a value to signify missing values? (y/n)
(This value will be considered missing for each variable.)
? n
*** DATA ENTERED ***





                     ===== Begin Commentary =====

     SURVAN  requests the  name  of  the   file containing the   data,
confirms that it was found and gives the user the option of proceeding
or entering a different file name.

    A value may be entered to indicate a missing value.  The data must
have this value as missing  for  all variables.  That is,  SURVAN does
not allow SEX to have a missing value of -9  but AGE to have a missing
value of -99.  All missing values must be the same.

                      ===== End Commentary =====

Press Return / Enter to continue:
?
Enter option number:
    (1) Print data
    (2) Define new survival variable
    (3) Define event for logistic regression model
    (4) Save current status to a file
    (5) Set/review/change group, strata, and compulsory variables
    (6) Define new print level for regressions
    (7) Compute frequencies
    (8) Compute condescriptive statistics
    (9) Run Kaplan-Meier estimation
   (10) Run log rank test
   (11) Run Gehan Breslow test
   (12) Run proportional hazards model estimation
   (13) Run logistic regression model estimation
   (14) End program

                     ===== Begin Commentary =====

     This is the main menu of SURVAN.  We  will visit and describe the
options under each entry.

                      ===== End Commentary =====

? 1
***** PRINT DATA *****
Enter the variable name(s) to be printed.
(A '?' causes the list of names to be displayed.)
?
     SURV       STATUS     GROUP
Enter the variable name(s) to be printed.
(A '?' causes the list of names to be displayed.)
all
Do you wish to print in an 80-column format? (y/n)
(If not, printing will be in a 132-column format.)
? y






    ***** Listing of Unsorted Data *****


  (All data listed as real numbers)



    SURV          STATUS        GROUP

  43.0000       1.00000       0.00000
  64.0000       1.00000       0.00000
  88.0000       1.00000       0.00000
  88.0000       1.00000       0.00000
  90.0000       1.00000       0.00000
  92.0000       1.00000       0.00000
  106.000       1.00000       0.00000
  109.000       1.00000       0.00000
  113.000       1.00000       0.00000
  116.000       1.00000       0.00000
  116.000       0.00000       0.00000
  120.000       1.00000       0.00000
  127.000       1.00000       0.00000
  130.000       1.00000       0.00000
  134.000       1.00000       0.00000
  144.000       0.00000       0.00000
  146.000       1.00000       0.00000
  165.000       1.00000       0.00000
  204.000       1.00000       0.00000
  42.0000       1.00000       1.00000
  56.0000       1.00000       1.00000
  63.0000       1.00000       1.00000
  98.0000       1.00000       1.00000
  104.000       0.00000       1.00000
  105.000       1.00000       1.00000
  132.000       1.00000       1.00000
  132.000       1.00000       1.00000
  133.000       1.00000       1.00000
  133.000       1.00000       1.00000
  133.000       1.00000       1.00000
  133.000       1.00000       1.00000
  139.000       1.00000       1.00000
  140.000       1.00000       1.00000
  161.000       1.00000       1.00000
  180.000       1.00000       1.00000
  180.000       1.00000       1.00000
  196.000       1.00000       1.00000
  196.000       1.00000       1.00000
  223.000       1.00000       1.00000
  244.000       0.00000       1.00000



*** Data printed ***





                     ===== Begin Commentary =====

     Menu selection 1 allows the user to print the data  in one of two
formats.  80-column  format is  suitable  for printing to  a terminal.
132-column format is suitable for printing on a line printer.

      Since   the user   may have   forgotten  the names  assigned  to
variables, entry of a '?',  when asked  for variable  names causes the
list of names to be printed. This  ability is  available in all places
where a variable  name is to  input.  The specification of 'all', when
asked for variable names  cause  the   operation to  be done  for  all
variables.  This ability is available where it makes  sense  and where
more than one variable name may be entered.

                      ===== End Commentary =====

Press Return / Enter to continue:
?
Enter option number:
    (1) Print data
    (2) Define new survival variable
    (3) Define event for logistic regression model
    (4) Save current status to a file
    (5) Set/review/change group, strata, and compulsory variables
    (6) Define new print level for regressions
    (7) Compute frequencies
    (8) Compute condescriptive statistics
    (9) Run Kaplan-Meier estimation
   (10) Run log rank test
   (11) Run Gehan Breslow test
   (12) Run proportional hazards model estimation
   (13) Run logistic regression model estimation
   (14) End program
? 2
***** NEW SURVIVAL DEFINITION *****
Enter the survival variable name.
(A '?' causes the list of names to be displayed.)
surv
Enter the event variable name.
(A '?' causes the list of names to be displayed.)
status
Enter the event variable values indicating an uncensored
(dead) survival time. (Terminate list with 'END')
1
end
*** Survival definition entered ***
                     ===== Begin Commentary =====

      Menu option 2 allows the user to specify which variable contains
the survival time,  and which variable  contains  the status for  each
observation.  This option must  be invoked before a Kaplan-Meier,  log
rank,  Gehan  Breslow,   or  proportional  hazards analysis     can be
performed.

                      ===== End Commentary =====






Press Return / Enter to continue:
?
Enter option number:
    (1) Print data
    (2) Define new survival variable
    (3) Define event for logistic regression model
    (4) Save current status to a file
    (5) Set/review/change group, strata, and compulsory variables
    (6) Define new print level for regressions
    (7) Compute frequencies
    (8) Compute condescriptive statistics
    (9) Run Kaplan-Meier estimation
   (10) Run log rank test
   (11) Run Gehan Breslow test
   (12) Run proportional hazards model estimation
   (13) Run logistic regression model estimation
   (14) End program
? 3
***** EVENT FOR LOGISTIC REGRESSION MODEL *****
An event must be defined whose probability is a
function of the independent variables.
This event can consist of:
    (1) Values of a variable above (or below) some criterion value
    (2) A specified set of values of a variable
Enter the desired type of event:
? 2
Enter the name of the variable whose values will
determine the occurrence of the event.
  (A '?' causes the list of names to be displayed.)
status
Enter a list of values indicating the occurrence of the event:
(Terminate list with "END")
1

*** Logistic regression event entered ***
                     ===== Begin Commentary =====

      Option 3 allows the user to specify which variable is  to define
the  logistic event.  This option  must be invoked before  a  logistic
regression is performed.

                      ===== End Commentary =====






Press Return / Enter to continue:
?
Enter option number:
    (1) Print data
    (2) Define new survival variable
    (3) Define event for logistic regression model
    (4) Save current status to a file
    (5) Set/review/change group, strata, and compulsory variables
    (6) Define new print level for regressions
    (7) Compute frequencies
    (8) Compute condescriptive statistics
    (9) Run Kaplan-Meier estimation
   (10) Run log rank test
   (11) Run Gehan Breslow test
   (12) Run proportional hazards model estimation
   (13) Run logistic regression model estimation
   (14) End program
? 4
***** SAVE CURRENT STATUS TO A FILE *****
Enter the name of the state file
where the state of this run will be saved
Specify file name or * for terminal output:
? sexample.sve
Created file sexample.sve --
   -- P(roceed) or R(etry file specification)?
? p
*** saved current status to a file ***
                     ===== Begin Commentary =====

      Option 4 allows the user to save the current program state  to a
file which can be read by SURVAN during another run. After reading the
state  file  SURVAN will   analyize the same data set   with the  same
survival and event variables as before. Any other variable definitions
or program options, such as the regression  print  level, will also be
set. The binary data file, the machine readable data file  produced by
SURVAN, must be in the same place when the option to restore is chosen
as when this option was invoked.

                      ===== End Commentary =====






Press Return / Enter to continue:
?
Enter option number:
    (1) Print data
    (2) Define new survival variable
    (3) Define event for logistic regression model
    (4) Save current status to a file
    (5) Set/review/change group, strata, and compulsory variables
    (6) Define new print level for regressions
    (7) Compute frequencies
    (8) Compute condescriptive statistics
    (9) Run Kaplan-Meier estimation
   (10) Run log rank test
   (11) Run Gehan Breslow test
   (12) Run proportional hazards model estimation
   (13) Run logistic regression model estimation
   (14) End program
? 5
***** SET/REVIEW/CHANGE GROUP,STRATA,COMPULSORY VARIABLES *****
Want to:
    (1) enter/change group variable
    (2) enter/change strata variable
    (3) enter/change compulsory variables
    (4) view group, strata, compulsory variables
    (5) return to option list
? 4
                     ===== Begin Commentary =====

      Menu  selection 2 allows the  user  to  specify  or view various
variable definitions.

      Choice 1 allows the user to  specify which variable will specify
the groups.    The  values  of  this variable   must be   integral and
consecutive.   SURVAN has    the capability  of  running  Kaplan-Meier
estimation separately    on groups and   groups  are  the   objects of
comparison in the Gehan Breslow and log rank test.

      Choice 2 allows the user  to specify which variable will specify