Keeping multiple dates for different events per individual ID.

[Matthew Milano]

Hi my fellow stata users. I have multiple diagnosis dates for different conditions per individual patient. I am trying to keep the earliest diagnosis date per condition per patient. I have tried the following command -bysort ID ( CCI_date) : keep if _n == 1-
but I incorrectly keep the earliest diagnosis date overall and not by each condition. I want to end up with several diagnosis dates per patient, but each diagnosis date will correspond to a different condition. I hope that's not too confusing!

Thanks you in advance. Below is sample data:

* Example generated by -dataex-. For more info, type help dataex
clear
input str1 ID byte charlson_category float CCI_date
"1" 1 14976
"1" 3 15721
"1" 3 15734
"1" 6 21375
"2" 6 21375
"2" 4 11688
"2" 6 20977
"2" 6 20576
"2" 6 20233
"2" 6 20233
"2" 4 20030
"3" 4 19710
"3" 6 20627
"3" 6 20258
"4" 6 20258
"4" 4 21339
"4" 6 21025
"4" 4 11688
"5" 4 19682
"5" 1 19780
end
format %td CCI_date
label values charlson_category charlson_category
label def charlson_category 1 "Any malignancy", modify
label def charlson_category 3 "Chronic pulmonary disease", modify
label def charlson_category 4 "Congestive heart failure", modify
label def charlson_category 6 "Diabetes", modify

[Øyvind Snilsberg]

to keep the first observation by ID and charlson_category,

by ID charlson_category (CCI_date), sort: keep if _n==1

[Matthew Milano]

Thank you, that has worked perfectly.

[Matthew Milano]

Hi I have a similar issue. In the sample below. For duplicate ID's where they have both Dementia_Type of both 'Alzheimer's' and 'unspecified'. I want to prioritise and keep the Alzheimer's disease regardless of diagnosis date. I.e. if for the same ID a diagnosis of Alzheimer's was 5 years after a diagnosis of Unspecified dementia I want to keep the Alzheimer's defined diagnosis date only for that ID.

* Example generated by -dataex-. For more info, type help dataex
clear
input float ID int Diagnosis_Date long Dementia_Type
 1 13808 6
 2 15299 1
 3 20792 6
 3 20773 1
 4 15594 6
 4 18176 1
 5 20076 6
 6 14748 1
 7 13082 1
 7 15461 6
 8 18120 1
 8 14243 6
 9 20881 1
 9  9313 6
10 18652 6
10 20527 1
11 15932 6
12 12799 1
13 15257 6
14 17224 6
15 11299 6
end
format %td Diagnosis_Date
label values Dementia_Type subtype_n
label def subtype_n 1 "Alzheimers Disease", modify
label def subtype_n 6 "Unspecified Dementia", modify

[Øyvind Snilsberg]

by ID (Dementia_Type), sort: drop if Dementia_Type==6 & Dementia_Type[1]!=Dementia_Type[_N]

[Matthew Milano]

Thank you again!