Matching by age, sex and service date between two cohorts?

[Bane Ling]

Hi, I am trying to do the followings.

I have a cohort who all have enrolled into a service centre (which could influence subsequent services received) with each of these exposed cases receiving at different enrolment date. The people not receiving the clinical service is a much larger cohort, the overall goal of the study is to access the difference in time between a diagnostic test and the final diagnosis between the exposed (with service centre) and unexposed (without service centre). Say I want to do it in 2 scenarios:

1) Which is a simplified version, in which I will ignore the matching of time, but just matching by age and sex. I have age and sex information about the exposed and unexposed cases, the matching should be 1:3 (one exposed: 3 unexposed).

2) The other scenario will be a bit more complicated, the difficulty is to match the date of diagnostic test in unexposed cases VS. the first diagnostic test date after enrolment of the service centre in exposed cases. For example, if the case is 19 year old woman received the initial diagnostic test on 01Mar2012 after the enrolment into the service centre in 15Feb2012, I would like to find 3 unexposed cases of 19 year women receiving a diagnostic test between 01Feb2012 and 01Apr2012.

Any tips on how I can tackle these or referral to source I can learn will be greatly appreciated!

Thanks.

[David Marso]

Have you carefully studied the MATCH FILES command?
If you have not then you really need to do so.
If you have and can not apply it then post the syntax you have attempted.
The second is more complicated so I will let you solve the first piece first and then return for that.
--

Bane Ling wrote

Hi, I am trying to do the followings.

I have a cohort who all have enrolled into a service centre (which could influence subsequent services received) with each of these exposed cases receiving at different enrolment date. The people not receiving the clinical service is a much larger cohort, the overall goal of the study is to access the difference in time between a diagnostic test and the final diagnosis between the exposed (with service centre) and unexposed (without service centre). Say I want to do it in 2 scenarios:

1) Which is a simplified version, in which I will ignore the matching of time, but just matching by age and sex. I have age and sex information about the exposed and unexposed cases, the matching should be 1:3 (one exposed: 3 unexposed).

2) The other scenario will be a bit more complicated, the difficulty is to match the date of diagnostic test in unexposed cases VS. the first diagnostic test date after enrolment of the service centre in exposed cases. For example, if the case is 19 year old woman received the initial diagnostic test on 01Mar2012 after the enrolment into the service centre in 15Feb2012, I would like to find 3 unexposed cases of 19 year women receiving a diagnostic test between 01Feb2012 and 01Apr2012.

Any tips on how I can tackle these or referral to source I can learn will be greatly appreciated!

Thanks.

[Jon K Peck]

Jon Peck (no "h") aka Kim
Senior Software Engineer, IBM
[hidden email]
new phone: [phone]




From:        Bombartang <[hidden email]>
To:        [hidden email],
Date:        12/11/2012 08:58 AM
Subject:        [SPSSX-L] Matching by age,              sex and service date between two cohorts?
Sent by:        "SPSSX(r) Discussion" <[hidden email]>

---

Hi, I am trying to do the followings.

I have a cohort who all have enrolled into a service centre (which could
influence subsequent services received) with each of these exposed cases
receiving at different enrolment date. The people not receiving the clinical
service is a much larger cohort, the overall goal of the study is to access
the difference in time between a diagnostic test and the final diagnosis
between the exposed (with service centre) and unexposed (without service
centre). Say I want to do it in 2 scenarios:

1) Which is a simplified version, in which I will ignore the matching of
time, but just matching by age and sex. I have age and sex information about
the exposed and unexposed cases, the matching should be 1:3 (one exposed: 3
unexposed).
>>>This can be done with the FUZZY extension command.  The input would be two datasets, and there are various options for the output.  You can specify any number of controls for a case.  Matching can be exact or within a "fuzz" tolerance.

2) The other scenario will be a bit more complicated, the difficulty is to
match the date of diagnostic test in unexposed cases VS. the first
diagnostic test date after enrolment of the service centre in exposed cases.
For example, if the case is 19 year old woman received the initial
diagnostic test on 01Mar2012 after the enrolment into the service centre in
15Feb2012, I would like to find 3 unexposed cases of 19 year women receiving
a diagnostic test between 01Feb2012 and 01Apr2012.
>>>If you have a variable in the exposed set for the first test date and variables in the unexposed set for test dates all in the same record, you can define a custom matching function to use with FUZZY.  Data manipulation might be required to set up the records depending on how your data is currently structured.

FUZZY is part of the Python Essentials, which can be downloaded from the SPSS Community website at www.ibm.com/developerworks/spssdevcentral.  Depending on your Statistics version, you might have to get a newer version of FUZZY from the Community site than then one installed with the Essentials.

HTH,
Jon Peck

Any tips on how I can tackle these or referral to source I can learn will be
greatly appreciated!

Thanks.



--
View this message in context: http://spssx-discussion.1045642.n5.nabble.com/Matching-by-age-sex-and-service-date-between-two-cohorts-tp5716852.html
Sent from the SPSSX Discussion mailing list archive at Nabble.com.

=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD

[Bane Ling]

Thanks Jon,

So the only way to accommodate the inexact date matching is via fuzzy commands. Is there a site I can learn all about the command structure and examples of fuzzy commands?

[Bane Ling]

Thanks David,

I have used Match Files before to merge different files, didn't know they can do matching as well. Just to make sure, can the Match Files command do both 3:1 matching and the "Fuzzy" (or inexact) matching by service dates?