delete records if none of the date observations occurred after a specific date

[tbe7]

I have a dataset that looks like this:

Patient_ID	Age	Obs_date	Obs_value	treatment
1	19	03/04/2020		A
1	.	12/09/2018		A
1	.	01/05/2019		.
2	.	12/22/2021		B
2	35	06/12/2018		A
2	.	04/12/2014		.
2	.	09/17/2020		.
2	.	08/28/2015		.
2	.	09/09/2021		C
2	.	03/31/2000		.
3	16	02/04/2017		A
3	.	09/12/2018		A
3	.	10/10/2020		A
4	44	05/22/2016		B
5	28	05/08/2016		B
5	.	09/05/2016		.
5	.	01/25/2017		.
5	.	01/09/2014		.
5	.	12/12/2014		A
5	.	07/18/2016		A
5	.	09/12/2012		.
5	.	06/07/2013		A
5	.	01/30/2017		C

I am trying to create a dataset with patients who:

• Aged≥18 years old
• Have Obs_date after from 03/01/2017 onwards

The new dataset should exclude patient_ID 3 (below 18 years old) and patient_IDs 4&5 (all Obs_dates are prior to 03/01/2017)

I tried the following:

data want;

set have;

if age =>18;

if '1mar2017'd <= obs_date;

run;

The output successfully deleted patient_IDs 3, 4, and 5; however, for patient_IDs 1&2, only one observation is kept for each. I’d like to keep all observations for patient_IDs 1&2 (screenshot below)

I'm looking for an output like the following:

Patient_ID	Age	Obs_date	Obs_value	treatment
1	19	03/04/2020		A
1	.	12/09/2018		A
1	.	01/05/2019		.
2	.	12/22/2021		B
2	35	06/12/2018		A
2	.	04/12/2014		.
2	.	09/17/2020		.
2	.	08/28/2015		.
2	.	09/09/2021		C
2	.	03/31/2000		.

[ballardw]

---

@tbe7 wrote:

I have a dataset that looks like this:

Patient_ID	Age	Obs_date	Obs_value	treatment
1	19	03/04/2020		A
1	.	12/09/2018		A
1	.	01/05/2019		.
2	.	12/22/2021		B
2	35	06/12/2018		A
2	.	04/12/2014		.
2	.	09/17/2020		.
2	.	08/28/2015		.
2	.	09/09/2021		C
2	.	03/31/2000		.
3	16	02/04/2017		A
3	.	09/12/2018		A
3	.	10/10/2020		A
4	44	05/22/2016		B
5	28	05/08/2016		B
5	.	09/05/2016		.
5	.	01/25/2017		.
5	.	01/09/2014		.
5	.	12/12/2014		A
5	.	07/18/2016		A
5	.	09/12/2012		.
5	.	06/07/2013		A
5	.	01/30/2017		C

I am trying to create a dataset with patients who:

• Aged≥18 years old
• Have Obs_date after from 03/01/2017 onwards

The new dataset should exclude patient_ID 3 (below 18 years old) and patient_IDs 4&5 (all Obs_dates are prior to 03/01/2017)

I tried the following:

data want;

set have;

if age =>18;

if '1mar2017'd <= obs_date;

run;

The output successfully deleted patient_IDs 3, 4, and 5; however, for patient_IDs 1&2, only one observation is kept for each. I’d like to keep all observations for patient_IDs 1&2 (screenshot below)

I'm looking for an output like the following:

Patient_ID	Age	Obs_date	Obs_value	treatment
1	19	03/04/2020		A
1	.	12/09/2018		A
1	.	01/05/2019		.
2	.	12/22/2021		B
2	35	06/12/2018		A
2	.	04/12/2014		.
2	.	09/17/2020		.
2	.	08/28/2015		.
2	.	09/09/2021		C
2	.	03/31/2000		.

---

Maybe if made sure AGE was on all of the records...

Though you have dates far enough apart that if the "age" shown is as of that date the subject would be 19 or older for later observations.

[tbe7]

The "age" was calculated from each patient's date of birth and it represents patient's age as of 2/28/2019 (as a baseline date). I then want to include only patients that have obs_date during the prior 24 months of the baseline date (i.e., only patients with obs_date back through 3/2/2017). Only patients that had ALL obs_date prior to the baseline date should be excluded. Hope that helps

[tbe7]

delete records if none of the date observations occurred after a specific date 

the value 35 for age should go for the 4th observation i.e., when patient_ID first turn to '2'. Sorry for the confusion!