Miracle Tracker

Re: How to output p value for covariance parameter estimates when using PROC GLIMMIX?

Miracle — Mon, 13 Dec 2021 02:32:55 GMT

Hi @di_niu0 .
This output section can be obtained by using the ods output statement where it is stored in CovParms in SAS.
HTH.

ods trace on;

ods output CovParms=outds;

proc glimmix data=schools method=quad;

class school_id class_id;

model score = class_id / solution dist=binary link=logit;

random intercept / subject=school_id solution;

run;

ods output close;

ods trace off;

Re: Exposure-adjusted event rate

Miracle — Wed, 25 Aug 2021 16:44:06 GMT

Thanks @MelissaM.

Re: Assign id

Miracle — Wed, 25 Aug 2021 16:42:36 GMT

Thanks everyone for the advice.

Re: Assign id

Miracle — Fri, 20 Aug 2021 08:49:55 GMT

Thanks @Kurt_Bremser for your solution using datastep.
But I need to perform this manipulation using proc sql.
BTW, this is not homework. I try to use proc sql more for my work now.
Thanks a lot.

Assign id

Miracle — Wed, 18 Aug 2021 09:15:11 GMT

Dear All,

May I know how do go from 'have' to 'want' using proc sql?

Thanks a lot.

have
studyid test
abc110  1
abc110  2
abc110  3
abc201  1
abc201  2
abc405  1
abc405  2

want
studyid test   n
abc110  1           1
abc110  2          1
abc110  3          1
abc201  1          2
abc201  2         2
abc405  1         3
abc405  2         3

Re: proc sql: combine data with string dates

Miracle — Wed, 11 Aug 2021 14:36:08 GMT

Dear @Patrick and All,
I have another question please.
How do I combine the demographic dataset that contains more patients to this sql please? So every patient from the demographic dataset will show in the final dataset regardless whether they are present in x or y dataset.

Thanks a lot. Again!

Re: Replicate every id by all possible unique group present in data

Miracle — Thu, 05 Aug 2021 15:00:57 GMT

Thanks @maguiremq for your quick reply.
I was thinking something like datastep by reading a macro variable containing all the unique grp then using do loop to repeat for every row but I dunno how exactly.
This is much easier and should do for now.

Re: proc sql: combine data with string dates

Miracle — Thu, 05 Aug 2021 14:27:24 GMT

Thanks a lot @Patrick.

It's a very detailed explanation for someone who isn't a sql person.
I really appreciate it.

Replicate every id by all possible unique group present in data

Miracle — Thu, 05 Aug 2021 14:25:04 GMT

Dear All,
Can I please seek your help on how to achieve this?
For example, I would like turn data X into data Y.

Thanks a lot.

Data X	Data Y
id grp	id grp
1 A	1 A
1 B	1 B
2 A	1 C
2 C	2 A
3 A	2 B
3 B	2 C
3 C	3 A
4 C	3 B
	3 C
	4 A
	4 B
	4 C

proc sql: combine data with string dates

Miracle — Wed, 04 Aug 2021 11:00:06 GMT

Dear All,
Can I please seek your help on how to combine this 2 proc sql together?
The first proc sql is to obtaine the date of first and last treatment.

The second proc sql is to identify treatment emergent AE.
Instead of two separate proc sql, how can I achieve this by using one proc sql?
Thanks a lot.

proc sql;
	create table txdat as 
	select id, 
	       min(input(ECSTD,yymmdd10.)) format=is8601da. length=8 as trtstd,
				 max(input(ECEND,yymmdd10.)) format=is8601da. length=8 as trtend
	from ec
	group by id;
quit;
proc sql;
	create table chk as 
	select coalesce (x.id,y.id) as id, 
				 x.trtstd, 
				 y.AEYN,
				 y.aeterm, 
				 case when AESTDT <> "" then input(y.AESTDT,is8601da.) else . end as aestd format=is8601da. length=8,
				 case when AESTDT <> "" and trtstd <> . and calculated aestd >= trtstd then 1 else 0 end as teae,
				 y.tmp
	from txdat as x
	full join ae as y
	on x.id=y.id;
quit;

Incorrect year from character date to SAS date

Miracle — Wed, 14 Jul 2021 11:40:44 GMT

Dear All,

Could you please point me out the what's wrong with my code?
Why it is giving me the year of 2020 instead of 2021?
Thanks a lot.

data _null_;
strdate="13 APR 2021";
sasdate=input(strdate,date9.);
format sasdate yymmdd10.;
put strdate=; 
put sasdate=;
run;

strdate=13 APR 2021
sasdate=2020-04-13

Re: Exposure-adjusted event rate

Miracle — Wed, 14 Apr 2021 12:59:08 GMT

Thanks @ChrisNZ for your reply.
I can see many on exposure-adjusted incidence rate only.
Would you mind to share me the link specifically for exposure-adjusted event rate please?

Thanks in advance.

Conditionally run proc mixed within macro

Miracle — Mon, 12 Apr 2021 09:53:19 GMT

Dear All,
Could you please to point out what have I done wrong given the following errors?

Thanking you in advance.

ERROR: Required operator not found in expression: &sub eq
ERROR: The macro REPEAT will stop executing.

ods output LSMeans=LSM_all_&param
SliceDiffs=SDiff_all_&param;
proc mixed data=inds;
%if &sub ne %then by &byvar;
where PARNAM1C="&param" and 1<=vis<=13;
class id resp tx sex vis / ref=first;
model chg=resp tx sex resp*vis / s ddfm=kenwardroger;
lsmeans resp*vis / cl;
repeated vis / subject=id type=&cov;
slice resp*VIS / sliceby=vis pdiff cl;
run;
ods output close;

Re: Exposure-adjusted event rate

Miracle — Mon, 12 Apr 2021 03:34:39 GMT

Hi @ChrisNZ .

From clinicaltrials.gov.

Exposure-adjusted event rate

Miracle — Fri, 02 Apr 2021 15:29:36 GMT

Dear All,
My apologies that this is not a SAS question.
But I'm wondering if anyone is able to explain to me on how exposure-adjusted event rate is calculated? Especially the denominator. Is the denominator the sum of differences between treatment start date and discontinuation/completion date from all patients?
Thanking you in advance.

Re: Infinite Likelihood error Proc Mixed

Miracle — Wed, 17 Mar 2021 04:50:59 GMT

Thanks @QiQi_22 for your reply.
I have already tried that but still getting the warning.

Re: Infinite Likelihood error Proc Mixed

Miracle — Mon, 15 Mar 2021 13:13:42 GMT

Hi @SteveDenham.
I am also getting this warning - WARNING: Stopped because of infinite likelihood.
I have tried different covariance structure as advised here.
And there is only one observation row for every VIS from ID.
Any advice from anybody will be very much appreciated.

proc mixed data=final;
by labparam;
class ID resp sex tx VIS;
model chg=base resp age sex tx VIS tx*VIS / s ddfm=kenwardroger;
lsmeans tx*VIS / cl;
repeated VIS / subject=ID type=ar(1);
/* repeated VIS / subject=ID type=vc;*/
/* repeated VIS / subject=ID type=cs;*/
/* repeated VIS / subject=ID type=un(1);*/
/* repeated VIS / subject=ID type=toep;*/
slice tx*VIS / sliceby=VIS pdiff cl;
run;.

Re: Conditional flagging id by few variables

Miracle — Fri, 05 Mar 2021 07:41:19 GMT

Yes, you are correct. The data is stored in a database.
Thanks @andreas_lds for pointing out.

Re: Conditional flagging id by few variables

Miracle — Fri, 05 Mar 2021 06:23:03 GMT

Thanks @andreas_lds for your response.

It is not homework.

It is a claim data with >1billion data rows.

I was advised to use sql because it will run faster but I am not good at sql.
Thanks.

Conditional flagging id by few variables

Miracle — Fri, 05 Mar 2021 03:20:00 GMT

Dear All,

Can I please seek your help on this by using sql?

I have a dataset containing id, date_1, date_2 and indication.

I would like to derive get, keep, date_3 and group.

The rationale is to flag keep=Y for the all the data rows with the earliest dates of both date_1 and date_2 for all the ids
If the data rows of keep=Y having both the same month and year i.e. id=1, then date_3=date_2 and group=Gen
If the data rows of keep=Y having different month and year i.e. id=2, then group=Indication.

Thanking you in advance.

id	date_1	        date_2	       indication	keep	date_3	       group
1	01-02-2020	20-02-2020	Y			
1	01-02-2018	03-03-2018	Y			
1	01-02-1990	20-02-1990	N	        Y	       20-02-1990	Gen
1	01-02-1990	20-02-1990	Y	        Y	       20-02-1990	Gen
1	01-02-1990	20-02-1990	Y	        Y	       20-02-1990	Gen
1	01-02-1990	20-02-1990	N	        Y	       20-02-1990	Gen
2	01-06-2009	25-08-2010	Y	        Y		                        Y
2	01-06-2009	25-08-2010	N	        Y		                        N
2	01-06-2009	25-08-2010	N	        Y		                        N
2	01-06-2009	25-08-2010	Y	        Y		                        Y
2	01-05-2018	08-06-2019	Y			
2	01-05-2018	04-12-2019	Y