About sjarvis847

sjarvis847 · ‎05-20-2022

Hello! I am attempting to search through 44 procedure codes (proc1-proc44) using an array to find specific codes related to a hip fracture (hipproc). I am using the following code and getting and error that the numeric variables are being converted to characters. I see that character variables can be used if you type in a $ after the subscript, but I still get the same error with that. Any suggestions: data gfp.main; set gfp.main; array procedure {43} $ proc1 - proc43; do i = 1 to 43; if procedure[i] = ('0QB4' or '0QB5' or '0QB6' or '0QB7' or '0QB8' or '0QB9' or '0QBB' or '0QBC' or '0SB9' or '0SBB') then hipproc= 1; if hipproc NE 1 then hipproc= 0; END; RUN; thanks so much for your help!

sjarvis847 · ‎05-20-2022

It would be great if you could post the code that worked! I am running into the same issue and trying to start with your not working code then revise per the other peoples instructions has been a puzzle in itself. Any chance you'd be willing to share what worked for you?

sjarvis847 · ‎01-08-2021

Hello SAS Community! I am hoping someone might have a trick to shorten the following if then statement. I am working in a file that contains 44 diagnosis codes, I need to go through those and make some of the diagnosis codes a new variable with the specific diagnosis (for example femoral neck fracture = yes1/no0), but am not interested in all of the diagnosis codes (there are over 5k) so I don't want to transpose the variable. What I have been doing is as follows: data.x; set.x; if diagcode1 = 'S72.0' then femoralneckfx = 1; if diagcode2 = 'S72.0' then femoralneckfx = 1; ....(and so on for all 44 diagnosis codes).... if femoralneckfx NE 1 then femoralneckfx = 0; run; then repeating for a similar process for the 9 diagnoses of interest. I found I could slightly shorten it a little by typing "if diagcode1 = 'S72.0' or diagcode2 = 'S72.0' of diagcode3 = 'S72.0' .......then femoralneckfx = 1;" But I was wondering if there is a way to write it even shorter, can you look for the text across all 44 variables at the same time so you don't have to repeat the ='S72.0'? I tried " if 'S72.0' in (diagcode1, diagcode2, diagcode3...) then femoralneckfx = 1; " but that didn't work either. Thanks for any help you can provide! Steph

sjarvis847 · ‎08-06-2020

Yes, we are comparing during COVID this year, to same period in 2019 and same period in 2018. There exists no rows for dates without a patient enrolled which is why those dates don't have a value....

sjarvis847 · ‎08-06-2020

Hi Paige, The code creates the number of patients enrolled per day. Then I take the average and compare if the average number of patients per day changed during the enrollment periods. Currently there is a difference observed in the average number of patients enrolled per day (in 2019 = 4 patients came in per day, during COVID = 2 patients came in per day). However because the days with 0 patients per day don't exist, the difference could be even greater that what I previously wrote. As there are not statistical tools to compare raw counts (I have the total volume [#] of patients during each enrollment period), I am comparing the number of patients who came in by day. I am sorry to hear this doesn't make sense to you. Thanks anyway, Steph Steph

sjarvis847 · ‎08-06-2020

Hi Reeza, Thanks for your reply! Yes it is operating at the daily level and is currently providing the accurate number of patients enrolled per day, except for days when no patients are enrolled. The values are not the same for each day. For example, 4/1/2018 4 patients were enrolled, 4/2/2018 1 patient was enrolled, 4/8/2018 no patients were enrolled but currently there exists no value for 4/8/2018 in the dataset for 4/82018 as it's only count days with patients enrolled. The goal is get a dataset with number of patients per day. Then assign the groups using enrollment date. Check normality of data, determine median or mean patients per day for each group. Use proper statistics to compare the mean or median patients per day in each group. Without the 0 values entered I won't have the true average number of patients per day, I will have the average number of patients enrolled per day for days which enrollment occured if that makes sense (which is not what I want). We are comparing admission of patients before and after COVID, assumption that COVID resulted in fewer admissions/ Thanks again for your help!!! Stephanie

sjarvis847 · ‎08-06-2020

Hi Paige, Thanks for your reply. I am hoping to look at patients/ day not per month & eventually want to use statistical methods to compare the patients per day enrolled per group. I know how to calculate an average by hand. I need it in sas so I can determine the p-value when I compare. Thanks for the suggestion though! Steph

sjarvis847 · ‎08-06-2020

Thanks for your reply! I just tried this and it gave me an error that there is an invalid DO loop, either the initial or TO expression is missing or the by expression is missing, zero or invalid.

sjarvis847 · ‎08-06-2020

Hello all! I used the following code to create a database which had the number of patients enrolled per day, that way I could compare the average patients per day during each enrollment period. However it's not accounting for the days with 0 patients enrolled. Anyone have an idea on how to add rows for days with 0 patients? proc sql; create table strcovid.new as select arrdate, count (mrn) as patients from strcovid.cov group by arrdate; create table average as select month(arrdate) as month, day(arrdate) as day, avg (patients) from strcovid.new group by month, day; quit; Thank you! Stephanie

sjarvis847 · ‎10-08-2019

Hello, I am using fishers or chi-squared to compare proportions across three groups, when appropriate. I was wondering if anyone knows how to get indivdiual p=values for each comparison? I know there is the option for cellchi2, but I need the actual p-values per journal request. Here is an example of one comparison made, proc freq data = reversal.all; tables hemiarthroplasy * studyarm/ exact; run; Thanks! Stephanie

sjarvis847 · ‎01-30-2019

Hello everyone, I recently received data from the CDC's national death index. They provide death data for subjects and if duplicate records are found they provide back all of the data found and let you feed through it. There is one variable they return called exact match; however if there is an exact match and there are multiple records pulled up, they still send all the matching files. I would like to keep the exact match if found and delete any subsequent records for the same patient, without deleting duplicates for remaining records without exact matches. Below is an example of some data (not the real data of course). See patient ID #1 has an exact match, so I would want to keep that match and delete the following two. Patient ID #2 does not have an exact match, so I have to look at more information to see what row I want to keep, therefore I do not want to delete any records for patient ID #2. ID DOB State DeathDate ExactMatch 1 9/15/2010 CO 10/26/2016 * 1 9/15/2010 IL 10/28/2016 1 9/15/2010 MO 1/6/2016 2 10/2/2013 KY 3/16/2016 2 10/2/2013 NY 2/26/2016 2 10/2/2013 CO 11/46/2016 Thanks for any help you can provide! Stephanie

sjarvis847 · ‎01-15-2019

This worked as well. I can't make all the replies a solution. Thanks for all the help, very much appreciated!!

sjarvis847 · ‎01-15-2019

This worked with retaining the row with the most responses, but did not always keep the first entry when multiple rows contained the same amount of missing variables. Thank you for your reply and help!

sjarvis847 · ‎01-15-2019

Hello! I am working on a study that we collected 6 different admission vitals for: systolic blood pressure, diastolic blood pressure, heart rate, respiratory rate, temp, and oximetry. The data came with up to three entries per patient and I would like to keep only one row that is the most complete with all the data from that timepoint. If there are multiple rows with the same amount of values collected the first recorded is the row we would like to keep. Any ideas on how to complete only the row with that is the most complete? Example of data Patient ID SBP DBP HR RR TEMP OXIMETRY 1 120 19 36.6 94 1 139 103 75 20 1 122 101 98 18 36.6 94 <------ This row is the desired row to keep 2 110 19 36.6 94 <------ This row is the desired row to keep 2 109 76 20 3 102 81 18 <------ This row is the desired row to keep 3 90 82 17 Thanks for any help you can provide! Stephanie

Online Status	Offline
Date Last Visited	‎05-24-2022 03:56 PM

Re: Array to search multiple character variables for variable differen...

Re: Searching multiple variables for multiple character strings

Shortened if then statement

Re: Average Patients Per Day

Re: Average Patients Per Day

Re: Average Patients Per Day

Re: Average Patients Per Day

Re: Average Patients Per Day

Average Patients Per Day

Proc Freq Multiple Comparisons p-value

Re: Select & keep the most complete row

Re: Select & keep the most complete row

Re: Array to search multiple character variables for variable differen...

Re: Searching multiple variables for multiple character strings

Shortened if then statement

Re: Average Patients Per Day

Re: Average Patients Per Day

Re: Average Patients Per Day

Re: Average Patients Per Day

Re: Average Patients Per Day

Average Patients Per Day

Proc Freq Multiple Comparisons p-value

deleting duplicate subject records by a variable

Re: Select & keep the most complete row

Re: Select & keep the most complete row

Select & keep the most complete row