About buszhangsy

buszhangsy · ‎08-30-2018

Thanks AMSAS! I do have a question. There will be some time periods with overlaps (see below for another example), how should I sort the dataset in order to use the code you provided? As you can see, when SAS processes the last record, it will keep the start_date previously stored but reassign end_date to be 05/20/2018, which is supposed to be 05/31/2018. 00001 01/01/2018 01/31/2018 00001 02/01/2018 02/28/2018 00001 04/01/2018 04/30/2018 00001 05/01/2018 05/31/2018 00001 05/15/2018 05/20/2018

buszhangsy · ‎08-30-2018

Hello, I have monthly enrollment data and would like to roll them up into continuous enrollment segments. Note that some overlap in monthly enrollment data can happen because they are from different sources. I tried to create a date variable, that is, do date=start to end (30 lines if monthly enrollment is Sep01 to Sep30) and check if dif(date)>1 for triggering discontinuation of current enrollment. However, my data set gets extremely large (300B obs) and run time is several hours. I would appreciate if anyone could provide a much efficient approach to tackle this. An example of data input and desired output are shown below. Thanks! Data input PatientID Start End 00001 01/01/2018 01/31/2018 00001 02/01/2018 02/28/2018 00001 04/01/2018 04/30/2018 00001 05/01/2018 05/20/2018 00001 05/15/2018 05/31/2018 Output PatientID Start End 00001 01/01/2018 02/28/2018 00001 04/01/2018 05/31/2018

buszhangsy · ‎08-29-2018

Any help is appreciated!

buszhangsy · ‎08-29-2018

I just put them there for information purpose. It does not matter what's the value of the Coverage variable. It just shows that there could be some records with overlapping time period.

buszhangsy · ‎08-29-2018

Hello, I have monthly enrollment data from commercial and Medicare claims and would like to roll them up into continuous enrollment segments. Tricky part is that an individual can have overlap coverage (commercial+Medicare). I tried to create a date variable (i.e. 30 lines if monthly enrollment is Sep01 to Sep30) and check if dif(date)>1 for triggering discontinuation of current enrollment. However, my data set gets extremely large (300B obs) and run time is several hours. I would appreciate if anyone could provide a much efficient approach to tackle this. An example of data input and desired output are shown below. Data input PatientID Start End Coverage 00001 01/01/2018 01/31/2018 Commercial 00001 02/01/2018 02/28/2018 Commercial 00001 04/01/2018 04/30/2018 Commercial 00001 05/01/2018 05/20/2018 Commercial 00001 05/15/2018 05/31/2018 Medicare Output PatientID Start End 00001 01/01/2018 02/28/2018 00001 04/01/2018 05/31/2018

buszhangsy · ‎06-03-2015

I want to sort prescription records by individual drug and the first service date and assign numbers to each drug based on sorting result but have no idea how to do that. An example is like this: Patient ID Drug name Service date 0001 Drug A 01/01/2007 0001 Drug B 01/10/2007 0001 Drug C 01/20/2007 0001 Drug A 02/01/2007 0001 Drug B 02/10/2007 0001 Drug C 02/20/2007 0001 Drug A 03/01/2007 0001 Drug B 03/10/2007 0001 Drug C 03/20/2007 The desired outcome is like this, Patient ID Drug name Service date Drug number 0001 Drug A 01/01/2007 1 0001 Drug A 02/01/2007 1 0001 Drug A 03/01/2007 1 0001 Drug B 01/10/2007 2 0001 Drug B 02/10/2007 2 0001 Drug B 03/10/2007 2 0001 Drug C 01/20/2007 3 0001 Drug C 02/20/2007 3 0001 Drug C 03/20/2007 3 Thanks!!!

buszhangsy · ‎09-30-2014

Hi PG, I examined the data on 20 patients and the code does its job! Thanks a lot!

buszhangsy · ‎09-29-2014

Thank you PG! I will run it first thing tomorrow morning!! The data can only be accessed through my office desktop computers. I will let you know the result!

buszhangsy · ‎09-29-2014

Omit the patient if we can not find a proper index date. Having an index date is one of the inclusion criteria for the study sample. Thanks!!

buszhangsy · ‎09-29-2014

No, I only need one index date. In most cases, that index date should be the first service date regardless of medications. However, if the corresponding medication does not have a follow-up service within 6 months, its service date can not be used as index date. I need to examine the second service date and see if it meets the criteria. Thanks!

buszhangsy · ‎09-29-2014

For most patients there were 2 index date created....I believe one of them was correct, but not sure how the other one generated.....

buszhangsy · ‎09-29-2014

CODE: proc sql; create table index2 as select baseid, FDB_GNN, min(filldate) as indexdate from index1 t1 where exists(select * from index2 where baseid=t1.baseid and FDB_GNN=t1.FDB_GNN and filldate>t1.filldate and intck('month',t1.filldate,filldate)<=6); quit; LOG: 444 proc sql; 445 create table index2 as 446 select 447 baseid, 448 FDB_GNN, 449 min(filldate) as indexdate 450 from 451 index1 t1 452 where exists(select * from index2 where baseid=t1.baseid and FDB_GNN=t1.FDB_GNN and filldate>t1.filldate 453 and intck('month',t1.filldate,filldate)<=6); ERROR: File WORK.INDEX2.DATA does not exist. 454 quit; NOTE: The SAS System stopped processing this step because of errors. NOTE: PROCEDURE SQL used (Total process time): real time 0.07 seconds cpu time 0.00 seconds

buszhangsy · ‎09-29-2014

Still got an error in the log: File WORK.WANT.DATA does not exist....

buszhangsy · ‎09-29-2014

Thanks! I only learned little SQL language. What does "from have t1" do in the program? I submitted the program and saw error in the log: file work.have.data do not exist, file work.want.data do not exist..

buszhangsy · ‎09-29-2014

I am working on a health-related project and I want to identify a medication use index date for each patient. The medication use index date is defined as the first service date for each patient. In most cases, it's just the first service date regardless of medications. However, the first service date can't be used if there is no additional service for that medication within 6 months. In this case, we have to use the second service date as the medication use index date. Here is an example. Patient ID Medication name Service date 0000120 Qvar 08/07/2007 0000120 SPIRIVA 08/14/2007 0000120 SPIRIVA 11/14/2007 0000120 SPIRIVA 01/16/2008 0000120 QVar 12/05/2008 For this patient, the first service date (08/07/2007) can't be used as the index date because the next service record (12/05/2008) for QVAR is beyond 6 months. I want to use the second service date (08/14/2007) as the index date. How can I write a SAS code to address this? Thanks!!

Online Status	Offline
Date Last Visited	‎09-05-2018 01:46 PM

Re: Roll up monthly enrollment data and create continuous enrollment s...

Roll up monthly enrollment data and create continuous enrollment segme...

Re: Roll up enrollment data

Re: Roll up enrollment data

Roll up enrollment data

Sorting question

Re: SAS Code help!

Re: SAS Code help!

Re: SAS Code help!

Re: SAS Code help!

Re: Roll up monthly enrollment data and create continuous enrollment s...

Roll up monthly enrollment data and create continuous enrollment segme...

Re: Roll up enrollment data

Re: Roll up enrollment data

Roll up enrollment data

Sorting question

Re: SAS Code help!

Re: SAS Code help!

Re: SAS Code help!

Re: SAS Code help!

Re: SAS Code help!

Re: SAS Code help!

Re: SAS Code help!

Re: SAS Code help!

SAS Code help!