topic Re: Data Quality Check: How to check that age increases or stays the same across visits in SAS Programming

Data Quality Check: How to check that age increases or stays the same across visits

awardell — Mon, 15 Mar 2021 15:19:20 GMT

Hi All,

I am working on a data set that has PatientID, Visit_Date, and Age. I need to check data quality by flagging those who have a decrease in age from an earlier visit_date to a later visit_date, as this is not possible. My data set looks like the following example:

PatientID	Visit_Date	Age
1	1/2/2003	25
1	2/4/2005	23
1	3/7/2006	26
2	3/8/2003	50
2	5/12/2004	51
2	5/13/2005	52
2	5/14/2006	53
3	4/6/2020	72
3	12/7/2020	65

where observations in PatientID 1 and 3 would need to be flagged. How should I go about doing this? Could I use the lag function? I am fairly unfamiliar with the lag function, so any help would be greatly appreciated! Thank you in advance!

Re: Data Quality Check: How to check that age increases or stays the same across visits

ballardw — Mon, 15 Mar 2021 15:34:09 GMT

One way using DIF function:

Data want;
   set have;
   ageflag = (dif(age)<0);
   by patientid;
   if first.patientid then ageflag=.;
run;

DIF returns current value - value on previous record. SAS will return 1 for true and 0 for false, so the comparison will set a flag of 1 when the value on the current record is less than for the previous.

Since you wouldn't want to compare one patient's age with another patient then the first patientid sets the flag to missing (could use 0 but missing is more appropriate in many cases).

The above solution assumes the data is sorted by patientid and date of visit.

If not sorted by patientid but grouped you could use NOTSORTED on the BY patientid statement.

DIF is closely related to LAG but since it is very common to do what you requested, the difference between current and previous, SAS supplies the DIF function as well.

Re: Data Quality Check: How to check that age increases or stays the same across visits

awardell — Tue, 16 Mar 2021 00:23:22 GMT

Thank you so much! This worked perfectly!