Computing a new variable by selecting on 100's of different diagnosis codes
Locked
Computing a new variable by selecting on 100's of different diagnosis codes
 
|
Hi,
I am working with a large medical claims file and need to compute a new variable that represents patients with one of over 300 different diagnosis codes. I would normally use the wizard to compute a new variable and enter the individual diagnosis codes that meet the criteria for inclusion. However, with hundreds of codes, it will take a very long time to enter them and there will likely be a lot of typos. Is there an efficient and accurate way to tell SPSS to select cases that have one of over 300 codes without having to type them in in the syntax editor or the wizard? Thank you! Mary -- Sent from: http://spssx-discussion.1045642.n5.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 |
Re: Computing a new variable by selecting on 100's of different diagnosis codes
Administrator
|
Where are these target diagnosis codes stored?
Are they in an electronic format or in a hard copy? [If not electronic you will have to type them]. Is the criteria variable a single field or a set of fields? How many cases in the file? THESE ARE THE TYPE OF THINGS YOU SHOULD POST IN INITIAL QUERIES. General solution: Codes are in a sorted SPSS dataset DSCritCode with a single variable CritCode. Raw data are a dataset Raw sorted on a unique ID variable [IDvar]. Criteria variable a set of contiguous fields CV1 TO CV10. Length of strings? CritCode and CV1 TO CV10 need to be identical. UNTESTED off the top of my head. DATASET ACTIVATE Raw . DATASET COPY CopyRaw . DATASET ACTIVATE CopyRaw . VARSTOCASES /ID=IDvar/MAKE CritCode FROM CV1 TO CV10. SORT CASES BY CritCode . MATCH FILES /FILE * /TABLE=DSCritCode /IN=@FLAG@/BY CritCode . AGGREGATE OUTFILE */BREAK IDvar / @FLAG@=MAX(@FLAG@). MATCH FILES /FILE Raw /FILE * / BY IDvar . Active file will have a new variable @FLAG@ with a 1 for cases with one or more of the values of interest. mcgannmary wrote > Hi, > > I am working with a large medical claims file and need to compute a new > variable that represents patients with one of over 300 different diagnosis > codes. I would normally use the wizard to compute a new variable and > enter > the individual diagnosis codes that meet the criteria for inclusion. > However, with hundreds of codes, it will take a very long time to enter > them > and there will likely be a lot of typos. > > Is there an efficient and accurate way to tell SPSS to select cases that > have one of over 300 codes without having to type them in in the syntax > editor or the wizard? > > Thank you! > Mary > > > > -- > Sent from: http://spssx-discussion.1045642.n5.nabble.com/ > > ===================== > To manage your subscription to SPSSX-L, send a message to > LISTSERV@.UGA > (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 ----- Please reply to the list and not to my personal email. Those desiring my consulting or training services please feel free to email me. --- "Nolite dare sanctum canibus neque mittatis margaritas vestras ante porcos ne forte conculcent eas pedibus suis." Cum es damnatorum possederunt porcos iens ut salire off sanguinum cliff in abyssum?" -- Sent from: http://spssx-discussion.1045642.n5.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
Please reply to the list and not to my personal email.
Those desiring my consulting or training services please feel free to email me. --- "Nolite dare sanctum canibus neque mittatis margaritas vestras ante porcos ne forte conculcent eas pedibus suis." Cum es damnatorum possederunt porcos iens ut salire off sanguinum cliff in abyssum?" |
|
|
In reply to this post by mcgannmary
Here is an example that matches the codes in one command. Any keys with no match in the lookup table will be sysmis. No sorting or restructuring of the dataset required. * The lookup table. data list free/ value(F8.0) akey(A1). begin data 10 'a' 20 'b' 100 'z' end data. dataset name lookup. * The main dataset. data list free/x(f8.0) y(A2). begin data 1 'a' 2 'b' 5 'a ' 10 '' 1 'b' end data. dataset name main. dataset activate main. * This command creates the result code by looking up the key in the lookup table. spssinc trans result = resultcodealpha /initial "extendedTransforms.vlookup('akey', 'value', 'lookup')" /formula func(y). list. x y resultcodealpha 1 a 10.00 2 b 20.00 5 a 10.00 10 . 1 b 20.00 The spssinc trans extension command can be installed via the Extensions > Extension Hub menu if you don't already have it. On Tue, May 7, 2019 at 1:07 PM mcgannmary <[hidden email]> wrote: Hi, |
Re: Computing a new variable by selecting on 100's of different diagnosis codes
|
|
Thank you, Jon. This is perfect! Mary From: SPSSX(r) Discussion <[hidden email]>
On Behalf Of Jon Peck Here is an example that matches the codes in one command. Any keys with no match in the lookup table will be sysmis. No sorting or restructuring of the dataset required. * The lookup table. data list free/ value(F8.0) akey(A1). begin data 10 'a' 20 'b' 100 'z' end data. dataset name lookup. * The main dataset. data list free/x(f8.0) y(A2). begin data 1 'a' 2 'b' 5 'a ' 10 '' 1 'b' end data. dataset name main. dataset activate main. * This command creates the result code by looking up the key in the lookup table. spssinc trans result = resultcodealpha /initial "extendedTransforms.vlookup('akey', 'value', 'lookup')" /formula func(y). list. x y resultcodealpha 1 a 10.00 2 b 20.00 5 a 10.00 10 . 1 b 20.00 The spssinc trans extension command can be installed via the Extensions > Extension Hub menu if you don't already have it. On Tue, May 7, 2019 at 1:07 PM mcgannmary <[hidden email]> wrote:
-- Jon K Peck ===================== 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
|
Re: Computing a new variable by selecting on 100's of different diagnosis codes
|
In reply to this post by Jon Peck
Jon
perhaps before the SPSSINC line set a default something like this *initialize to a user missing code. numeric ResultCodeAlpha (f2). compute ResultCodeAlpha = -1. missing values ResultCodeAlpha (low thru -1). value labels ResultCodeAlpha -1 'has codes but no code on list' . . . If there are input cases with all blank codes should that be treated as a different kind of user missing? ----- Art Kendall Social Research Consultants -- Sent from: http://spssx-discussion.1045642.n5.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
Art Kendall
Social Research Consultants |
|
|
That would have no effect on the result as it would be overwritten. All the code knows is that the lookup did not find a value. A user could prepare the lookup table with a specific user missing value if they know what values would be looked up, but it would be easier to post process the lookup results. On Thu, May 9, 2019 at 12:44 PM Art Kendall <[hidden email]> wrote: Jon |
Re: Computing a new variable by selecting on 100's of different diagnosis codes
|
|
I should have thought through the process a little more. post makes more
sense My main idea was the user knows why the result was missing so symis should be corrected either there were valid codes but none matched the list or there were no valid codes. In many instances the reasoning about missingness would need to take that into account. ----- Art Kendall Social Research Consultants -- Sent from: http://spssx-discussion.1045642.n5.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
Art Kendall
Social Research Consultants |
«
Return to SPSSX Discussion
|
1 view|%1 views
| Free forum by Nabble | Edit this page |