Calculating a sum of total cases and total deaths (case fatality ratio)

[Julia Silver]

Hi Stata Community,

I am getting stuck on something I feel has a more succinct way of coding than what I am currently doing. I would like to calculate case fatality ratios. My time variables are discharge year and discharge quarter. So far, I've calculated the total # cases and total #deaths by repeating the code below. However, the egen total function scores the result in all of the observations meeting the condition--that's not quite what I want. Is there a way I can store the total number of deaths in just one cell? I would like to generate a case fatality ratio for each unique year/quarter by dividing the cases variable/death variable. Thank you in advance!


sort discharge_year discharge_qtr
egen deaths_2016_q1=total(expired) if discharge_year==2016 &discharge_qtr==1
gen cases_2016_q1=_n if discharge_year==2016 &discharge_qtr==1

* Example generated by -dataex-. To install: ssc install dataex
clear
input int encounter_id long(discharge_year discharge_qtr) byte expired
 187 2016 1 0
 998 2016 1 0
 128 2016 1 1
  19 2016 1 0
  83 2016 1 0
 586 2016 1 1
 127 2016 1 1
1170 2016 1 0
 141 2016 1 0
1236 2016 1 0
 359 2016 1 0
1537 2016 1 0
1203 2016 1 0
1045 2016 1 0
 314 2016 1 0
 866 2016 1 0
 582 2016 1 0
 433 2016 1 0
 981 2016 1 0
1008 2016 1 1
 730 2016 1 0
 158 2016 1 0
 853 2016 1 0
1137 2016 1 0
1069 2016 1 0
1197 2016 1 0
1230 2016 1 0
1173 2016 1 0
1144 2016 1 1
 400 2016 1 0
1051 2016 1 1
1174 2016 1 0
 784 2016 1 0
 825 2016 1 0
 915 2016 1 1
1181 2016 1 0
 392 2016 1 0
1585 2016 1 0
1289 2016 1 0
1207 2016 1 0
 919 2016 1 0
 113 2016 1 0
 758 2016 1 0
 883 2016 1 1
 928 2016 1 0
 278 2016 1 0
 788 2016 1 1
1642 2016 1 0
1504 2016 1 0
1446 2016 1 0
1531 2016 1 0
1740 2016 1 0
 123 2016 1 0
 994 2016 1 0
 467 2016 1 0
  40 2016 1 0
1350 2016 1 0
  71 2016 1 0
  10 2016 1 0
1560 2016 1 0
1141 2016 1 0
 877 2016 1 1
1416 2016 1 0
 703 2016 1 0
1284 2016 1 0
1739 2016 1 0
1227 2016 1 0
  13 2016 1 0
 639 2016 1 0
 805 2016 1 0
1211 2016 1 0
 234 2016 1 0
1534 2016 1 0
1118 2016 1 0
1399 2016 1 0
 414 2016 1 0
 504 2016 1 0
 197 2016 1 0
1021 2016 1 0
 213 2016 1 0
 472 2016 1 0
1044 2016 1 0
 742 2016 1 0
1728 2016 1 0
 257 2016 1 0
 117 2016 1 1
1722 2016 1 0
1025 2016 1 0
 252 2016 1 0
1557 2016 1 0
 383 2016 1 0
 707 2016 1 0
1123 2016 1 0
 203 2016 1 0
 355 2016 2 0
 403 2016 2 0
  55 2016 2 0
1039 2016 2 0
1704 2016 2 0
1409 2016 2 1
end

[Ken Chui]

How about just aggregating over them?

collapse (count) case = expired (sum) expired = expired, by(discharge_year discharge_qtr)
gen cf_ratio = expired / case * 100

Also, regarding

I would like to generate a case fatality ratio for each unique year/quarter by dividing the cases variable/death variable

I think it's the other way, death / cases.

[Julia Silver]

Ken Chui oh whoops! Yes, definitely meant death/cases.

[Julia Silver]

Thank you so much Ken Chui ! That worked beautifully

[Ken Chui]

Thank you so much Ken Chui ! That worked beautifully

You are welcome. Glad that it worked out fine.