Adding extra years to an existing date variable

[Sigrid Vikjord]

Dear Statalist,

I have a date variable created the following way:

gen PartDat_gli = .
replace PartDat_gli = PartDat_NT2BLQ1 if copd_HUNT==1
replace PartDat_gli = PartDat_NT3BLQ1 if copd_HUNT==2

* Example generated by -dataex-. To install: ssc install dataex
clear
input long(PartDat_NT2BLQ1 PartDat_NT3BLQ1) float PartDat_gli
13444 17079     .
13114 17293     .
13619 17675     .
13040     .     .
13537     .     .
13065     .     .
13300 17289     .
13661 17640     .
13522 17538     .
13116 17342     .
13468     .     .
13394     .     .
13212 17448     .
13117 17385     .
13425     .     .
13670 17674     .
13377     .     .
13403     .     .
13654 17689     .
13459 17274     .
13535 17580 17580
13647     .     .
13207 17140 13207
    . 17651     .
    . 17644     .
13236 17498     .
13124 17282     .
13082 17475     .
    . 17324 17324
13388 17237     .
    . 17336     .
    . 17553     .
13201     . 13201
13216 17489     .
13675 17644     .
    . 17189     .
13480 17672     .
13677 17660     .
13472 17667     .
13614     .     .
13065 17343     .
    . 17632     .
    . 17623     .
13403     .     .
13675 17688     .
13543 17230     .
13569     .     .
13683     .     .
13457     .     .
13440     . 13440
13293     .     .
13678 17660     .
13055 17321     .
13430 17092     .
13165 17395     .
13115     .     .
13135     .     .
13293     .     .
    . 17391     .
13079 17447     .
    . 17707     .
13193     .     .
13461 17317     .
13317 17238     .
13123     .     .
13206     .     .
13591 17630 17630
13475 17566     .
    . 17491     .
    . 17328     .
13032     .     .
13152     .     .
13289 17589     .
13499 17505     .
    . 17406     .
13212     .     .
13130 17205     .
13191     . 13191
13310 17283     .
13221 17447     .
13234 17492     .
13038 17407     .
13565     .     .
    . 17660     .
13255     .     .
13416     .     .
    . 17315     .
    . 17632     .
13212     .     .
13534     .     .
13375     .     .
    . 17290     .
13235 17174     .
    . 17475     .
13424 17120     .
13052     .     .
    . 17269     .
13628     . 13628
13212 17496     .
13289     .     .
end
format %tdD_m_Y PartDat_NT2BLQ1
format %tdD_m_Y PartDat_NT3BLQ1

It is the participation date for individuals with COPD who participated in either of/both the first (NT2) or the second (NT3) cycle of a population study (their final PartDat_gli depends on when they were diagnosed with COPD, that's why some have the date from NT3 despite participating in both cycles). I use the date variable to -stset- before -stcox-.
I now want to run a sensitivity analysis excluding the first 3 years of the follow-up time. How could I add 3 years for each participant to this date variable, taking into account the different participation dates? I tried to have a look at earlier threads in the forum, but I couldn't find a solution to my exact problem.

Best regards,
Sigrid Vikjord

[Andrew Musau]

How could I add 3 years for each participant to this date variable

There is no participant id in your data example. How do you identify a participant in your data?

[William Lisowski]

For the purposes of your question, after reading in your example data

. format %tdD_m_Y PartDat_gli

. list if PartDat_gli !=. , clean abbreviate(16)

       PartDat_NT2BLQ1   PartDat_NT3BLQ1   PartDat_gli  
 21.         21 Jan 97         18 Feb 08     18 Feb 08  
 23.         28 Feb 96         05 Dec 06     28 Feb 96  
 29.                 .         07 Jun 07     07 Jun 07  
 33.         22 Feb 96                 .     22 Feb 96  
 50.         18 Oct 96                 .     18 Oct 96  
 67.         18 Mar 97         08 Apr 08     08 Apr 08  
 78.         12 Feb 96                 .     12 Feb 96  
 98.         24 Apr 97                 .     24 Apr 97

So is your objective to increase the values of PartDat_gli by 3 years? If so, this is going to pose a problem for any February 29, because three years later there is no February 29. Perhaps you would be content to increase the values of PartDat_gli by 3*365 + 1 days, because the odds are 3:1 that a 3-year span will include a leap year day. If so, this will do.

. generate PartDat_gli3 = PartDat_gli + 365*3 + 1
(92 missing values generated)

. format %tdD_m_Y PartDat_gli3

. list if PartDat_gli !=. , clean abbreviate(16)

       PartDat_NT2BLQ1   PartDat_NT3BLQ1   PartDat_gli   PartDat_gli3  
 21.         21 Jan 97         18 Feb 08     18 Feb 08      18 Feb 11  
 23.         28 Feb 96         05 Dec 06     28 Feb 96      28 Feb 99  
 29.                 .         07 Jun 07     07 Jun 07      07 Jun 10  
 33.         22 Feb 96                 .     22 Feb 96      22 Feb 99  
 50.         18 Oct 96                 .     18 Oct 96      19 Oct 99  
 67.         18 Mar 97         08 Apr 08     08 Apr 08      09 Apr 11  
 78.         12 Feb 96                 .     12 Feb 96      12 Feb 99  
 98.         24 Apr 97                 .     24 Apr 97      24 Apr 00

Or we could get more accurate by adding another line of code. I think (but have not confirmed) that February 29 2000 would become February 28 2003.

. replace PartDat_gli3 = PartDat_gli3-1 if day(PartDat_gli3)!=day(PartDat_gli)
(2 real changes made)

. list if PartDat_gli !=. , clean abbreviate(16)

       PartDat_NT2BLQ1   PartDat_NT3BLQ1   PartDat_gli   PartDat_gli3  
 21.         21 Jan 97         18 Feb 08     18 Feb 08      18 Feb 11  
 23.         28 Feb 96         05 Dec 06     28 Feb 96      28 Feb 99  
 29.                 .         07 Jun 07     07 Jun 07      07 Jun 10  
 33.         22 Feb 96                 .     22 Feb 96      22 Feb 99  
 50.         18 Oct 96                 .     18 Oct 96      18 Oct 99  
 67.         18 Mar 97         08 Apr 08     08 Apr 08      08 Apr 11  
 78.         12 Feb 96                 .     12 Feb 96      12 Feb 99  
 98.         24 Apr 97                 .     24 Apr 97      24 Apr 00