cancel
Turn on suggestions
Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.
Showing results for 
Search instead for 
Did you mean: 
Wolverine
Friday
  • 212 Posts
  • 10 Likes Given
  • 4 Solutions
  • 12 Likes Received

Re: Proc GLM and weights

by Pyrite | Level 9 in Statistical Procedures
‎03-03-2020 12:07 PM
‎03-03-2020 12:07 PM
I went back and forth on whether weighting was necessary, but eventually I concluded that the model should look to see if any year was statistically significantly different from the mean rate.  And the mean rate should be weighted so that years with more cases would have more effect on the mean.   There is a statement in the draft report: "Rates did not change over time during the study period."  So I'm trying to verify that this statement is true.  I think the strictest interpretation is that no year is different than any other year.  And in that case, the rates would be compared to each other, not the mean, and therefore weighting wouldn't be necessary.  Correct? ... View more

Proc GLM and weights

by Pyrite | Level 9 in Statistical Procedures
‎03-03-2020 11:36 AM
‎03-03-2020 11:36 AM
This should be a simple question, but it's been a long time since I did any actual analysis so I want to double-check.  I have a list of rates of people who have a certain medical condition (the numerator) within a certain population (the denominator).  I need to figure out if the changes in rate across years are statistically significant.  My supervisor directed me to use PROC GLM and to weight appropriately.  I want to make sure that I'm calculating and applying the weights correctly.  Below is the data from the file named "MEDpercent_2005_2012" and the code that I used: /*Calculate total number of denominator cases*/ PROC SQL; Create table temp.MEDpercent_2005_2012_sum as Select *, Sum (Denominator_case_count) as total_denom_05_12 From temp.MEDpercent_2005_2012; QUIT; DATA temp.MEDpercent_2005_2012_wt; SET temp.MEDpercent_2005_2012_sum; Weight = (Denominator_case_count / total_denom_11_18); RUN; /*Model rate as a function of year*/ ODS HTML; ODS graphics on; title1 "MED - GLM regression modeling rate as a function of year"; title2 "A non-significant results indicates no significant differences over years"; PROC GLM data = temp.MEDpercent_2005_2012_wt; weight Weight Model MED_rate = year / solution; RUN; QUIT; ODS graphics off; ODS HTML close;   Year Numerator Denominator MED_rate 2005 114 251 45.56% 2006 101 245 41.17% 2007 113 243 46.72% 2008 116 252 45.78% 2009 107 272 39.31% 2010 134 366 36.75% 2011 112 319 35.21% 2012 141 331 42.54% ... View more

Re: Convert character variable to HEX

by Pyrite | Level 9 in SAS Programming
‎10-23-2019 11:41 AM
‎10-23-2019 11:41 AM
Believe me, you're preaching to the choir.  I waste an inordinate amount of time trying to get SAS and Excel to work together.  But I'm basically in the middle of the flowchart -- the data files are produced upstream by someone else (who technically doesn't even work on this project but has access to the data warehouse where the info comes from).  I process the files, pulling out the records we're interested in, and then pass them along to the principal investigator.  To review, she needs to be able to easily display and sort, add comments and additional variables, etc.  Excel does that and it's something that we already have on our computers.    The upstream person also created the encrypted_ID variable.  I don't know why it has a length of 200.  It displays with a length of 64 characters with no extra spaces.  Here is an example: 20850C1CF365857E237F9D83A9F5DCCD7525F99AB62987BB025C2430ED8B1E90   I also discovered that Excel inserts a line break for some reason.  So that same ID in Excel is displayed as 20850C1CF365857E237F9D83A9F5DCCD7525F99 AB62987BB025C2430ED8B1E90 even though the entire number is contained in the same cell.  But I suspect the line break may explain the extra 2 characters that give it a length of 66 instead of 64.   When I figured this out, it gave me an idea.  I simply created truncated versions of the IDs with a length of 40, and then used those to match.  I had to manually compare the full IDs to make sure there weren't 2 separate IDs that truncated to the same ID (there weren't).  It's ugly, but it worked.   ... View more

Convert character variable to HEX

by Pyrite | Level 9 in SAS Programming
‎10-23-2019 08:50 AM
‎10-23-2019 08:50 AM
I have a variable called encrypted_id that has the following characteristics:   Type: Char Length: 200 Format: $HEX64.   I had to export a file into Excel for manual review/editing, and then I imported back into SAS.  Now I need to match it with the original files.  But the encrypted_id  variable was changed during the export/import process.  It now has these characteristics and does not match the original variable:   Type: Char Length: 66 Format: $66.   I tried this, but it didn't work and the log says "Invalid argument to function INPUT":   DATA temp.SCA_Atb_NPI_manual_review; SET temp.SCA_Atb_NPI_manual_review; encrypted_id_hex = input(encrypted_id,$HEX64.); RUN; ... View more

Converting number of days since Jan 1 1960 to YYYYMMDD

by Pyrite | Level 9 in SAS Programming
‎04-24-2019 08:54 AM
‎04-24-2019 08:54 AM
Seems like this should be a very simple task, but I've been searching and I can't find a straight answer.   I have a variable named IMGDate that contains the number of days since Jan 1 1960.  For example, 21164 represents Sept 27, 2017.   I want to convert this variable so that 21164 is in 20170927 format. ... View more

Re: Problem converting HEX format to character

by Pyrite | Level 9 in SAS Programming
‎03-11-2019 09:48 AM
‎03-11-2019 09:48 AM
The data is restricted, and beneficiary_id can be linked to patient names.  So beneficiary_id  is converted to encrypted_ID, and then beneficiary_id is dropped from the dataset before I'm allowed to have access to it.  So, going back to the original question, how do I convert encrypted_id so I can I create a test dataset that only includes a few specific encrypted_id's? ... View more

Re: Problem converting HEX format to character

by Pyrite | Level 9 in SAS Programming
‎03-11-2019 08:15 AM
‎03-11-2019 08:15 AM
Unfortunately I didn't create the data sets, and the guy who did no longer works here.  However, he left his code as documentation.  The data were pulled from a restricted-access data warehouse, not from Excel.  Here is the section where the encrypted_ID is created for the PREVIOUS data sets:     /* 1.4 Encrypt beneficiary_id */ proc sql; create table work.scd_data_3 as select sha256(put(input(beneficiary_id, 10.), z10.)) as encrypted_id format $hex64. ,category ,category_description from work.scd_data_2; quit; And this is how he created it for the CURRENT data set: /* 5.2 Create an encrypted ID */ proc sql; create table work.encrypted_ids_2 as select beneficiary_id ,sha256(beneficiary_id) as encrypted_id format $hex64. from work.encrypted_ids_1; quit; ... View more

Re: Problem converting HEX format to character

by Pyrite | Level 9 in SAS Programming
‎03-07-2019 11:26 AM
‎03-07-2019 11:26 AM
You're right -- I got the same warning just above the error: WARNING: Variable encrypted_id has already been defined as numeric.   That means that this particular data file was created differently than the rest of the files in this data collection.  I ran proc contents on this file, and on another file from this collection.    Current file: encrypted_id Num 8 Other file: encrypted_id Char 200 $HEX64.    I don't understand how encrypted_id can be 8-digit numeric when it includes letters and has an actual length of 64.  The long string of numbers and letters in my first bit of syntax was copied directly from the SAS data file and is indeed 64 characters long.  And how do I delete any record that doesn't belong to a specific ID?      ... View more

Problem converting HEX format to character

by Pyrite | Level 9 in SAS Programming
‎03-07-2019 08:42 AM
‎03-07-2019 08:42 AM
I have a collection of data that with IDs that have been encrypted in a hex format.  Sometimes, as part of testing, I find it helpful to follow just a few cases thru the syntax to make the correct cases are being selected at each step.  To do so, I use the following syntax in a data step:   encrypted_id2 = put(encrypted_id, $hex64.); IF encrypted_ID2 ^in ("0720275C8B0B0C376C1B272DA78FD0DF10FB5408EDE76CD21C637EF2F42FC4AB") then delete;   I've used this bit of syntax dozens of times while working with this data.  But on this particular occasion, I got an error that I am at a loss to explain: ERROR 29-185: Width specified for format HEX is invalid.   I've confirmed that the ID is in fact 64 characters.  I tried searching online but couldn't find any useful info.  So what could possibly cause this issue?           ... View more

Re: Eliminate redunant information across 3 variables

by Pyrite | Level 9 in SAS Programming
‎01-24-2019 09:17 AM
‎01-24-2019 09:17 AM
Thank u not only for the code, but also for the explanation of what the code does. This is a great approach that was simple to program and provided exactly what I was looking for! ... View more

Re: Eliminate redunant information across 3 variables

by Pyrite | Level 9 in SAS Programming
‎01-16-2019 10:37 AM
‎01-16-2019 10:37 AM
I see your point.  If I apply my rule to your example, it won't delete any records.  I certainly like the idea of consolidating your example to a single line... now how do I actually program that? ... View more

Re: Eliminate redunant information across 3 variables

by Pyrite | Level 9 in SAS Programming
‎01-16-2019 07:42 AM
‎01-16-2019 07:42 AM
Conceptually, one potential solution would be to have SAS group all records that have the same ID and claim date.  Then, within each of those groups, have SAS look at the DX codes and assign a flag variable.  If a given code appears only once, it would be flagged as 0.  Other codes appearing more than once would be flagged as 1.  Repeat this process for proc and rev codes.   Then an IF statement could use those flags to delete records that don't contain any new information:   IF DX_flag > 0 AND Proc_flag > 0 AND Rev_flag > 0 THEN delete;   So that's the idea... can anyone help me figure out how to actually program that? ... View more

Re: Eliminate redunant information across 3 variables

by Pyrite | Level 9 in SAS Programming
‎01-15-2019 03:02 PM
‎01-15-2019 03:02 PM
I need all of the variables in the example records.   Obs 1 and 3 combined provides the values of Obs 2 in these particular examples.  But the redundant records are not always Obs 2.  Also, there may be any number of records for a particular kid on a particular date. ... View more

Eliminate redunant information across 3 variables

by Pyrite | Level 9 in SAS Programming
‎01-15-2019 02:14 PM
‎01-15-2019 02:14 PM
I'm working with medical claims data, and I'm trying to create a file that will show the medical histories for a select group of kids.  For each claim date for each kid, I need a list of diagnosis, procedure, and revenue codes.  But many of the records do not include any new information.  In each of the 2 examples below, the middle line contains information that is already present in the other lines.  So the middle line is redundant and could be deleted. encrypted_id DATE_OF_BIRTH GENDER CLAIM_SERVICE_FROM_DATE DIAGNOSIS_CODE_t PRCDR_CODE REVENUE_CODE 02D957E3EDF5 23-Aug-05 F 26-Apr-16 H52223 V2715 0611 02D957E3EDF5 23-Aug-05 F 26-Apr-16 H5203 V2715   02D957E3EDF5 23-Aug-05 F 26-Apr-16 H5203 V2020                 encrypted_id DATE_OF_BIRTH GENDER CLAIM_SERVICE_FROM_DATE DIAGNOSIS_CODE_t PRCDR_CODE REVENUE_CODE 14A17629B53944 31-Jul-07 F 20-Apr-17 R480 99174 0610 14A17629B53944 31-Jul-07 F 20-Apr-17 Z6852 99174   14A17629B53944 31-Jul-07 F 20-Apr-17 Z6852 99214     Based on an approach I found in another post, I tried the code below.  But for some reason, the rev_new variable contained diagnosis codes and DX_new contained rev codes.  And there was still redundant information in the output file.   DATA temp.numer_ID_codesort; SET temp.numer_ID_list_dd; LENGTH DX_new proc_new rev_new $10; DX_new = DIAGNOSIS_CODE_t; proc_new = PRCDR_CODE; rev_new = REVENUE_CODE; CALL SORTC(rev_new,proc_new,DX_new); RUN; PROC SORT DATA = temp.numer_ID_codesort OUT = temp.numer_ID_final nodupkey; BY DX_new proc_new rev_new; RUN;   ... View more

Re: PROC PRINT -- observations limited by percentage of total

by Pyrite | Level 9 in SAS Procedures
‎10-02-2018 01:22 PM
‎10-02-2018 01:22 PM
Simple and effective!  Thanks! ... View more
Contact Me
Online Status
Offline
Date Last Visited
Friday
Likes given to
View All
Likes from
Quick links