Date:         Sun, 29 Jun 2008 10:00:14 -0400
Reply-To:     Sigurd Hermansen <HERMANS1@WESTAT.COM>
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         Sigurd Hermansen <HERMANS1@WESTAT.COM>
Subject:      Re: Programming question on viral load dataset
Comments: To: sassql@GMAIL.COM
In-Reply-To:  <2472274b-c212-4b9b-a7cb-7a824d2b536e@x41g2000hsb.googlegroups.com>
Content-Type: text/plain; charset="us-ascii"

sassql:
Catchy e-mail address!

As Ian has observed, writing clear and consistent requirements often
takes more time and attention than writing a program. I'd suggest
reviewing the requirements. In particular, the first and fourth
classification rules,

>1 = Confirmed Virologic Failure: Failure to achieve <400 cp/mL on or
before 20 weeks
>4 = Non Virologic Failure: Not confirmed/suspected >=400 cp/mL after
achieving suppression to <400 cp/mL by week 24

    appear to leave a logical gap between week 20 and week 24. Also,
rules 2 and 3 do not appear to classify patients into mutually exclusive
classes.
S

-----Original Message-----
From: SAS(r) Discussion [mailto:SAS-L@LISTSERV.UGA.EDU] On Behalf Of
sassql@GMAIL.COM
Sent: Friday, June 27, 2008 3:23 PM
To: SAS-L@LISTSERV.UGA.EDU
Subject: Programming question on viral load dataset


Dear all,

I have a viral load dataset for around 400 patients. I want to determine
the patients who had viralload failure as per the following criteria.

1  =    Confirmed Virologic Failure: Failure to achieve <400 cp/mL on
or before 20 weeks'
2 =    Confirmed Virologic Failure: >=400 cp/mL confirmed after
initial supression to <400 cp/mL'
3 =    Suspected Virologic Failure: >=400 cp/mL on last visit after
initial supression to <400 cp/mL'
4 =    Non Virologic Failure: Not confirmed/suspected >=400 cp/mL
after achieving suppression to <400 cp/mL by week 24

For each of the patient i have to derive the following variables:

1) Criteriacode: This variable willl have either value 1,2,3 or 4
depending on which patient's viral load satisfies the criteria mentioned
above.

2) ttvf: Time to virologic failure.

2) Failuredate: The date when the patient failed

Here is the example of how the data looks like.

PTNO   WEEK     labdate               Viralload
101      Baseline   01apr2001            >=400
101      Week2     15apr2001            >=400
101      Week4     30apr2001            >=400
101      Week8     28may2001          >=400
101      Week12   25June2001          <400
101      Week16   23July2001           <400
101      Week20   20aug2001            >=400
101      Week24   17sept2001           >=400

As per Criteria 2 this patient has a confirmed rebound of VL >=400 cp/
ml after initial suppression to <400 cp/ml


PTNO   WEEK     labdate               Viralload
102      Baseline   01apr2001            >=400
102      Week2     15apr2001            >=400
102      Week4     30apr2001            >=400
102      Week8     28may2001          >=400
102      Week12   25June2001          <400
102      Week16   23July2001           <400
102      Week20   20aug2001            <400
102      Week24   17sept2001           >=400
THis patient as per criteria 3 will be assigned criteria 3 since it
rebounded only at the visit 24.

PTNO   WEEK     labdate               Viralload
103      Baseline   01apr2001            >=400
103      Week2     15apr2001            >=400
103      Week4     30apr2001            >=400
103      Week8     28may2001          >=400
103      Week12   25June2001          >=400
103      Week16   23July2001           >=400
103      Week20   20aug2001            >=400
103      Week24   17sept2001           >=400

For this patient, it should be assigned criteria 1 since the viral load
never went below 400 cp/mL. And all other patients who didn't meet
criteria 1,2, and 3 will be assigned criteria 4.

Can someone tell me how to programatically determine each of this
criteria using the above mentioned sample data?

Thanks in advance