Use of the results of PROC PHREG in PROC MIANALYZE

[paola1]

Hello everybody, I'm using SAS Studio and I'm a beginner, but I have the following problem.

The option CLASS in PROC MIANALYZE after the PROC PHREG doesn't work: where am I doing wrong?

If I run the following code:

PROC MI DATA=data NIMPUTE=10 out=out_data seed=54321;
mcmc plots=trace plots=acf ;
VAR hpcr cystatin_c;
RUN;

PROC PHREG DATA=out_data;
CLASS sex (REF=FIRST) ipertensione (REF='No') diabete (REF='No')
ipercolesterolemia (REF='No') cvd (REF='No') q_fumo (REF='No') ALBUMINURIA (REF='< 3.4') / PARAM=REF;
MODEL GIORNI*morto_vivo2(0) = ALBUMINURIA sex age ipertensione diabete ipercolesterolemia q_fumo cvd gfr p_ac_urico colesterolo c_ldl hpcr cystatin_c/ RISKLIMIT COVB;
BY _imputation_;
OUTPUT OUT=mi ParameterEstimates=a_mvn CovB=mxcovb;
RUN;
QUIT;

PROC MIANALYZE PARMS(CLASSVAR=full)=a_mvn COVB(effectvar=rowcol)=mxcovb;
CLASS ALBUMINURIA;
MODELEFFECTS ALBUMINURIA sex age ipertensione diabete ipercolesterolemia q_fumo cvd gfr p_ac_urico colesterolo c_ldl hpcr cystatin_c;
RUN;

The LOG follows:

1 OPTIONS NONOTES NOSTIMER NOSOURCE NOSYNTAXCHECK;

55

56 PROC MI DATA=INCIPE_SURV NIMPUTE=10 out=OUT_INCIPE seed=54321;

57 mcmc plots=trace plots=acf ;

58 var hpcr cystatin_c;

59 run;

NOTE: The EM algorithm (MLE) converges in 4 iterations.

NOTE: The EM algorithm (posterior mode) converges in 2 iterations.

NOTE: The data set WORK.OUT_INCIPE has 37550 observations and 76 variables.

NOTE: PROCEDURE MI used (Total process time):

real time 1.38 seconds

cpu time 0.55 seconds

60

61 /* 2. Analysis Phase:*/

62 PROC PHREG DATA=OUT_INCIPE;

63 FORMAT ALBUMINURIA albFmt. sex sexFmt. ipertensione ipertensioneFmt. diabete diabeteFmt.

64 q_fumo q_fumoFmt. ipercolesterolemia ipercolesterolemiaFmt. cvd cvdFmt.;

65 CLASS sex (REF=FIRST) ipertensione (REF='No') diabete (REF='No')

66 ipercolesterolemia (REF='No') cvd (REF='No') q_fumo (REF='No') ALBUMINURIA (REF='< 3.4') / PARAM=REF;

67 MODEL GIORNI*morto_vivo2(0) = ALBUMINURIA sex age ipertensione diabete ipercolesterolemia q_fumo cvd gfr p_ac_urico

67 ! colesterolo c_ldl hpcr cystatin_c/ RISKLIMIT COVB;

68 BY _imputation_;

69 OUTPUT OUT=mi ParameterEstimates=a_mvn CovB=mxcovb;

__________________

22

76

ERROR 22-322: Errore di sintassi, previsto uno dei seguenti: ;, (, /, ATRISK, CIF, DFBETA, LD, LMAX, LOGLOGS, LOGSURV, NUM_LEFT,

OUT, RESDEV, RESMART, RESSCH, RESSCO, STDXBETA, SURVIVAL, WTRESSCH, XBETA.

ERROR 76-322: Errore di sintassi, l'istruzione verrà ignorata.

70 RUN;

NOTE: The SAS System stopped processing this step because of errors.

NOTE: PROCEDURE PHREG used (Total process time):

real time 0.01 seconds

cpu time 0.00 seconds

71 QUIT;

72

73 /* 3. Pooling Phase: */

74 TITLE 'Multivariate Cox Proportional Hazard Models - Full Model 2';

75 PROC MIANALYZE PARMS(CLASSVAR=full)=a_mvn COVB(effectvar=rowcol)=mxcovb;

ERROR: File WORK.A_MVN.DATA does not exist.

ERROR: File WORK.MXCOVB.DATA does not exist.

76 CLASS ALBUMINURIA;

77 FORMAT ALBUMINURIA albFmt. sex sexFmt. ipertensione ipertensioneFmt. diabete diabeteFmt.

78 q_fumo q_fumoFmt. ipercolesterolemia ipercolesterolemiaFmt. cvd cvdFmt.;

79 MODELEFFECTS ALBUMINURIA sex age ipertensione diabete ipercolesterolemia q_fumo cvd gfr p_ac_urico colesterolo c_ldl hpcr

79 ! cystatin_c;

80 RUN;

NOTE: The SAS System stopped processing this step because of errors.

NOTE: PROCEDURE MIANALYZE used (Total process time):

real time 0.00 seconds

cpu time 0.01 seconds

81

82 OPTIONS NONOTES NOSTIMER NOSOURCE NOSYNTAXCHECK;

94

So, if I change my code in this way:

PROC PHREG DATA=OUT_INCIPE NOPRINT;
CLASS sex (REF=FIRST) ipertensione (REF='No') diabete (REF='No')
ipercolesterolemia (REF='No') cvd (REF='No') q_fumo (REF='No') ALBUMINURIA (REF='< 3.4') / PARAM=REF;
MODEL GIORNI*morto_vivo2(0) = ALBUMINURIA sex age ipertensione diabete ipercolesterolemia q_fumo cvd gfr p_ac_urico colesterolo c_ldl hpcr cystatin_c/ RISKLIMIT COVB;
BY _imputation_;
OUTPUT OUT=a_mvn;
RUN;
QUIT;

PROC MIANALYZE DATA=a_mvn;
MODELEFFECTS ALBUMINURIA sex age ipertensione diabete ipercolesterolemia q_fumo cvd gfr p_ac_urico colesterolo c_ldl hpcr cystatin_c;
RUN;

I receive this message in the LOG file right after the PROC MIANALYZE:

ERROR: The input TYPE= data set is not a valid data set without specifying variables for standard errors in the STDERR statement.

NOTE: The SAS System stopped processing this step because of errors.

NOTE: PROCEDURE MIANALYZE used (Total process time):

real time 0.00 seconds

cpu time 0.01 seconds

[rayIII]

Hi, Paola.

You're trying to pass a table of parameter estimates to proc mianalyze. 

You can get the parameter estimates from PHREG as ODS output. Try changing your OUTPUT statement as follows: 

  ods output parameterEstimates  = a_mvn; 

Ray