ASSIGNING ADVERSE EVENTS TO VISITS (ADDING 2 IMPUTED DAYS)

[michtka]

Hi guys, i've got the next problem, I need to assigned the AE of subjects 1 and 2 given by the dataset AE,

with the VISIT records given by the dataset ASSIGNED.

I am doing PROC SQL to assigned  most of the VISITS, but there are two specific VISITS that I want to include

in the AE that I cant get and I need too.

fist one (01DEC2008): I explain why...

The last VISIT in my final dataset SORTWANT (see final result obtained using PROC PRINT) is 20OCT2008,

but because the AE for this subject stop the 26NOV2008, I would like extend this AE until the next record 01DEC2008.

second one (20JAN2009):...the same history...

The last VISIT in my final dataset is 22DEC2008,

but because the AE for this subject 2 stop the 01JAN2009, I would like extend this AE until the next record 20JAN2009.

Then, the question is:

Anyway to get this two extra imputed VISITS: (01DEC2008) and (20JAN2009) in my final dataset SORTWANT?

Thnaks,

V.

data AE;

informat START date9. STOP date9.;

input SUBJID  START STOP;

format START STOP date9.;

cards;

1 07APR2008 26NOV2008

2 06DEC2008 01JAN2009

run;

data ASSIGNED;

informat VISIT date9.;

input SUBJID  VISIT;

format VISIT date9.;

cards;

1 02APR2008

1 07APR2008

1 12MAY2008

1 19MAY2008

1 30JUN2008

1 01JUL2008

1 07JUL2008

1 07AUG2008

1 29SEP2008

1 13OCT2008

1 20OCT2008

1 01DEC2008  *IMPUTED VISIT THAT I AM INTERESTED TO INCLUDED IN MY FINAL DATASET SORTWANT;

1 22FEB2009

2 27NOV2008

2 08DEC2008

2 22DEC2008

2 20JAN2009   *IMPUTED VISIT THAT I AM INTERESTED TO INCLUDED IN MY FINAL DATASET SORTWANT;

run;

PROC SQL;

CREATE table want AS

SELECT ae.SUBJID, ae.START, ae.STOP, ASSIGNED.visit

FROM ae

LEFT join ASSIGNED a

ON ae.SUBJID=a.SUBJID

AND a.VISIT BETWEEN ae.START AND ae.STOP;

QUIT;

PROC SORT data=want out=sortwant;

  by SUBJID VISIT;

RUN;

PROC PRINT data=sortwant NOOBS;

RUN;

The Result obtained is:

    SUBJID        START         STOP        VISIT

                              1      07APR2008    26NOV2008    07APR2008

                              1      07APR2008    26NOV2008    12MAY2008

                              1      07APR2008    26NOV2008    19MAY2008

                              1      07APR2008    26NOV2008    30JUN2008

                              1      07APR2008    26NOV2008    01JUL2008

                              1      07APR2008    26NOV2008    07JUL2008

                              1      07APR2008    26NOV2008    07AUG2008

                              1      07APR2008    26NOV2008    29SEP2008

                              1      07APR2008    26NOV2008    13OCT2008

                              1      07APR2008    26NOV2008    20OCT2008  *last visit subject 1;

                              2      06DEC2008    01JAN2009    08DEC2008

                              2      06DEC2008    01JAN2009    22DEC2008  *last visit subject 2;

[michtka]

adding a LAG period of 30 to the ae.STOP  can help, but is not like i really want, I only want to get the next visit.

Any ideas?

Thanks.

PROC SQL;

CREATE table want AS

SELECT ae.SUBJID, ae.START, ae.STOP, ASSIGNED.visit

FROM ae

LEFT join ASSIGNED a

ON ae.SUBJID=a.SUBJID

AND a.VISIT BETWEEN ae.START AND ae.STOP + 30;

QUIT;

[Jagadishkatam]

Hi,

i think the below code will help you to get those missing dates in the final dataset.

proc sql;

    create table ae_ as select a.*,b.visit from ae as a left join assigned as b on

    a.subjid=b.subjid where ((start<=visit<=stop) or (visit>=stop));

quit;

Thanks,

Jag