About Banke

Banke

Both codes work, thank you both so much! Is it possible to accept two responses as the solution?

Banke

Hello all, I found a solved question relating to identifying gaps in therapy (https://communities.sas.com/t5/SAS-Programming/Data-cleaning-for-Cox-regression-analysis/td-p/730972). I have a similar problem but I don't know how to: 1. create the days and days_int variables. 2. censor patients whose dose changed. I appreciate your assistance. Thank you ID Dose Med_start_date Days Days_int 1 2 01JAN2015 0 . 1 4 16JAN2015 15 15 2 2 01JAN2010 0 . 2 4 11APR2010 100 100 3 8 01JAN2012 0 . 3 16 31JAN2012 30 30 3 24 10MAY2012 130 100 4 8 01MAR2017 0 . 4 8 30APR2017 60 60 4 8 29JUN2017 120 60 4 16 28AUG2017 180 60 4 16 27OCT2017 240 60 5 8 01FEB2018 0 . 5 8 02APR2018 60 60 5 16 01JUN2018 120 60 5 16 21JUL2018 170 50 5 32 20AUG2018 200 30 5 32 18NOV2018 300 90 6 1 01JAN2011 0 . 6 2 20FEB2011 50 50 6 2 21APR2011 110 60 6 4 20JUN2011 170 60 6 6 18SEP2011 260 90 7 8 01JAN2019 0 . 7 16 28OCT2019 300 300

Banke

Yes, you are correct. It worked perfectly. Thank you so much!

Banke

Hello everyone, I need your help please: I want to calculate the average daily dose (ADD) for each patient (= quantity_supplied/days_supplied) for the first 6 months of prescription fills (fill_date), such that for each patient, I have a single value of ADD that corresponds to the average number of tablets they took per day from the first time they received the medication (index date) till up to 6 months. Your help will be greatly appreciated. Thank you patient_id fill_date index_date quantity_supplied days_supplied 1 01/11/2018 01/11/2018 140 30 1 01/12/2018 01/11/2018 140 30 1 01/22/2018 01/11/2018 160 30 1 02/16/2018 01/11/2018 90 90 1 03/19/2018 01/11/2018 90 90 1 06/18/2018 01/11/2018 90 90 2 02/12/2017 02/12/2017 120 30 2 03/23/2017 02/12/2017 120 30 2 06/14/2017 02/12/2017 90 90 2 07/30/2017 02/12/2017 90 90 2 08/31/2017 02/12/2017 60 60

Banke · ‎10-28-2022

Thank you so much!

Banke · ‎10-28-2022

Hello everyone, I have two variables: start_year and end_year. start_year is in recorded in years and in format numeric 8. while end_year is mmddyy10. i have tried to subtract and get the age but i keep getting funny values even with some of the codes i found. Attached is the simplest code i tried, your help is highly appreciated. sample data. start_date (8.) end_date(mmddyy10.) 2015 01/02/2017 2013 05/07/2014 2018 06/06/2019 data want; set have; format end_date year4.; end_date2 = end_date; format end_date2 8.; age = end_date2 - start_date; run;

Banke · ‎10-28-2022

I understand it better now, thank you

Banke · ‎10-27-2022

Sorry for the confusion. The baseline characteristics i want to identify are age, sex, region, diseases they had in that period and medications they took. I am not really worried about the first 3 because they usually stay the same but they are likely to have fewer diseases and take fewer medications before the index date. The disease variables are like 10 e.g. icd1 icd2 icd3 icd4 etc. each icd variable have list of conditions the patients have. I dont know how to link this because the index, start, and end dates are fixed for each patient. But checking now, there is a medical treatment date variable. Is that what i will use to identify if the condition was diagnosed before the index date? There is also fill date for prescription. Thank you

Banke · ‎10-27-2022

Hello everyone, Please, I am monitoring patietns for 1 year: 6months before they start taking medications (start_date) and 6 months after (end_date). The first day of medication intake is the index_date. Each patient has unique start, index, and end dates. I am asked toidentify baseline characteristics of my patients 6 months before they start taking medications i.e before the index date.My data consists of multiple observations for each patient. I am trying to reduce these to observations occuring 6 months before the index date.I am confused at what to do. I will really appreciate your help. Thank you My data looks like this patient start_date index_date end_date 1 04/2015 10/24/2015 04/2016 1 04/2015 10/24/2015 04/2016 1 04/2015 10/24/2015 04/2016 2 01/2015 07/17/2015 01/2016 2 01/2015 07/17/2015 01/2016

Banke · ‎10-24-2022

PROC TRANSPOSE DATA = ENROLLMENT1 OUT = ENROLLMENT2; BY PATIENT_ID; VAR ENROLLED_01_2014--ENROLLED_12_2016; RUN; DATA ENROLLMENT2(rename=(col1=type)); SET ENROLLMENT2; MONTH = INPUT(SUBSTR(_NAME_, 10,2), 8.); YEAR = INPUT(SUBSTR(_NAME_, 13,4), 8.); DATE = MDY( MONTH,1,YEAR); FORMAT DATE YYMMS7.; DROP _NAME_ MONTH YEAR; RUN; Thank you, I agree with your thought process. I was able to convert the variable names to date format in the long format using the attached code.

Banke · ‎09-30-2022

Oh sorry, thats not my data, i just wanted to show what it looks like.

Banke · ‎09-30-2022

TYou are absolutely right about the arrangement in my array not making sense. I just observed that. Thank you for the code you shared, i am not familiar with the functions and dont really understand it. Will i use it in a data step?

Banke · ‎09-30-2022

Thank you so much, i dont know how i missed your response. Been waiting..lol. So i used your code and there was no error. I think what SAS did was not look at patients with value "B" in my array. What i really want is to identify patients that had value "B" from their start to end dates i.e. are continouslly enrolled in that period. Do you have an idea of how i can go about it please?

Banke · ‎09-30-2022

My data is like this: data have; infile datalines ; informat patient_id $. startdate indexdate enddate mmddyy10.;; format patient_id $. startdate indexdate enddate mmddyy10.; input patient_id $. startdate indexdate enddate mon1 mon2 mon3 mon4 mon5 mon6 mon7 mon8 mon9 mon10 mon11 mon12; datalines; AA 12/30/2016 07/25/2014 01/21/2015 A B C B C C C C ...... BB 11/29/2016 01/07/2014 07/06/2014 A B C B C B B C ...... CC 08/08/2015 05/10/2015 11/06/2015 A B C B C C C C ...... DD 12/06/2016 10/23/2015 04/20/2016 A C C C C C C C ...... EE 11/27/2016 05/07/2014 11/03/2014 A B C B C C D ......

Banke · ‎09-30-2022

START and END_1 are date formats MMDDYY10. The array elements are character variables with values of either A,B,C and D. I am interested in "C" because it is the value that shows that the patient have twe two types of insurance i want. I want to select patients that are continously enrolled with that kind of insurance from the START to END_1 date. I had previously calculated STAT_1 and END_1 by subtracting and adding 160 days repsectively from the variable INDEX_date which is the first day of using the insurance. Thank you

Online Status	Offline
Date Last Visited	10 hours ago

Re: creating a cummulative variable based on dates

creating a cummulative variable based on dates

Re: calculating mean dose taken for the first 6 months of therapy

calculating mean dose taken for the first 6 months of therapy

Re: computing age in years from SAS dates

computing age in years from SAS dates

Re: identifying diseases before a particular date

Re: identifying diseases before a particular date

identifying diseases before a particular date

Re: array to identify patients with insurance

Re: how to pad character variable with leading zeroes?

Re: creating a cummulative variable based on dates

Re: creating a cummulative variable based on dates

Re: Extract year from a date

Re: Rename all variables starting with the same preffix to a different...

Re: Using macros to import ssp file

Re: creating a cummulative variable based on dates

creating a cummulative variable based on dates

Re: calculating mean dose taken for the first 6 months of therapy

calculating mean dose taken for the first 6 months of therapy

Re: computing age in years from SAS dates

computing age in years from SAS dates

Re: identifying diseases before a particular date

Re: identifying diseases before a particular date

identifying diseases before a particular date

Re: array to identify patients with insurance

Re: array to identify patients with insurance

Re: array to identify patients with insurance

Re: array to identify patients with insurance

Re: array to identify patients with insurance

Re: array to identify patients with insurance