how to calculate age and sex adjusted incidence rates after poisson

[Mithila shrestja]

Hello Everyone,
I have data sets in the format (fage04:female ages 0-4) up to fage85+ and in similar format for males for years 1970-2013. I want to use stratum specific data not the pooled one.Then enter age and sex as covariate in the poisson. Could you please help me with the command?

Thank you in advance.
Mithila

[Carlo Lazzaro]

Mithila:
you may want to try:

poisson depvar i.sex##i.age_cat, exposure(n) irr///this way you will also obtain the interaction between sex and age_category

[Mithila shrestja]

Hello Carlo,
I do not have either sex as distinct male and female or age catagory .The data looks like this:
f for female(fage04: female age 0-4 and corresponding number 5 is the no of cases)

dg_y	fage04	fage59	fage1014	fage1519	fage2024	fage2529	fage3034
1970	5	5	4	1	5	2	3
1971	2	6	2	2	0	5	5
1972	6	3	1	4	2	4	3
1973	4	1	6	0	4	4	8
1974	4	1	4	3	2	10	8

fpop04(female population 0-4)

fpop04	fpop59	fpop1014	fpop1519	fpop2024	fpop2529
166273	186729	194975	205481	216751	161956
160437	184936	191684	206166	214936	170361
153917	183210	189868	203577	213005	190829
147585	180080	190427	200193	209255	205904

m for male

mage04	mage59	mage1014	mage1519	mage2024
5	6	3	3	4
1	4	4	1	2
3	5	1	5	3
2	4	3	3	5

and population format for male similar as female.
What will be the option for such data type?
Thank you.
Mithila

[Carlo Lazzaro]

Mithila:
your data are very dificult to read. Please post them using the user-written command dataex (install using ssc install dataex) or code delimiters (both recommended by the FAQ). Thanks.
That said, I would advise to -reshape- your dataset from -wide- to -long- format, first.
Then you can deal with -poisson-.

[Carlo Lazzaro]

Mithila:
what follows should merge in the same .dta file glioma cases and population for female patients (please note that data concerning population for -fage3034- and -1974- are faked, because they were missing in your previous post. I would recommend you to cross-check them in your dataset):

drop _all
input dg_y fage04 fage59 fage1014 fage1519 fage2024 fage2529 fage3034
dg_y     fage04     fage59   fage1014   fage1519   fage2024   fage2529   fage3034
  1970 5 5 4 1 5 2 3
  1971 2 6 2 2 0 5 5
  1972 6 3 1 4 2 4 3
  1973 4 1 6 0 4 4 8
  1974 4 1 4 3 2 1 0 8
  end
reshape long fage, i( dg_y ) j(fage_group)
rename fage glioma_cases
replace glioma_cases=fage04 if glioma_cases==.
drop fage04
save "C:\Users\user\Desktop\Glioma_cases_female.dta", replace
drop _all
input dg_y  fpop04 fpop59 fpop1014 fpop1519 fpop2024 fpop2529 fpop3034

          dg_y     fpop04     fpop59   fpop1014   fpop1519   fpop2024   fpop2529   fpop3034
  1970 166273 186729 194975 205481 216751 161956 161956
  1971 160437 184936 191684 206166 214936 170361 170361
  1972 153917 183210 189868 203577 213005 190829 190829
  1973 147585 180080 190427 200193 209255 205904 205904
  1974 147585 180080 190427 200193 209255 205904 205904
 end
reshape long fpop, i( dg_y ) j(fage_group)
replace fpop = fpop04 if fpop ==.
drop fpop04
save "C:\Users\user\Desktop\Glioma_cases_fpopulation.dta", replace
merge 1:1 dg_y fage_group using "C:\Users\user\Desktop\Glioma_cases_female.dta"
label define fage_group 4 "0-4" 59 "5-9" 1014 "10-14" 1519 "15-19" 2024 "20-24" 2529 "25-29" 3034 "30-34"
label val fage_group fage_group
g sex=1
label define sex 1 "Female"
label val sex sex
drop _merge
save "C:\Users\user\Desktop\Glioma_cases_and_fpopulation.dta", replace

You should follow the same approach for male patients. Then you can -append- Glioma_cases_and_fpopulation.dta and Glioma_cases_and_mpopulation.dta files via -sex- variable and save this new file.
The last step will be to apply -poisson- code reported in post #2 to the data populating the last saved .dta file.

[Mithila shrestja]

Hello Carlo,
Now I have done it.I appreciate what you did for me. I could have never done it without your help. Thank you so much.

[Carlo Lazzaro]

Mithila:
thanks for your kind words.
All the best for you and your research.

[Mithila shrestja]

Thank you Carlo,
Now I want to use this command
poisson depvar i.sex##i.age_cat, exposure(n) irr
I have mglioma_cases, fglioma_cases as depvar. For exposure mpop(male population) and fpop(female population).
I am wondering how could I fit all those in poisson?
I have tried with single variable in one time, it does not make any sense.
or do I need to go with dstdize?

Thank you.

[Carlo Lazzaro]

Mithila:
you should -rename- fglioma_cases and fpop in the file concerning female patients:

use "C:\Users\user\Desktop\Glioma_cases_and_fpopulation.dta", replace
rename fglioma_cases glioma_cases
​rename fpop pop
save "C:\Users\user\Desktop\Glioma_cases_and_fpopulation.dta", replace

then do the same for male patients.

Eventually, you should -merge- both files via-sex- variable and save this new file.
The last step will be to apply -poisson- code reported in post #2 to the data populating the last saved .dta file.

[Mithila shrestja]

Why its saying
variable sex does not uniquely identify observations in the master data.

[Carlo Lazzaro]

Mithila:
you're right. My mistake.
You should try:

use "C:\Users\user\Desktop\Glioma_cases_and_fpopulation.dta", clear
append using "C:\Users\user\Desktop\Glioma_cases_and_mpopulation.dta"

:

[Mithila shrestja]

I tried with append also.
I am using stat12, while I use append to combine those two files(male and female), it says variable sex does not uniquely identify observations in the master data. For single male and female files, sex variables are clearly labelled.(I have given 1 for female and 2 for male) But when I append them, if I take male as a master data and append female data with it, in sex variable for female it shows 1 not female and if i take female as master data and append male data, it shows 2 for male.Why its happening like that?

Did you understand my question?

[Carlo Lazzaro]

Mithila:
unfortunately, I cannot replicate your problem, because -append- the two files worked for me and the label for -sex-, was kept as well (this may prove that I understood your question).
Now the issue is: why didn't it work for you?

[Mithila shrestja]

Ok. Thank you Carlo for help.I have options for appending as:
keep all the variables from appending data set/variables to be kept from appending data sets
I have tried both but still I have same problem. What did you mean to append via sex? Is there any difference between different versions of Stata

[Carlo Lazzaro]

Mithila:
I don't think that the problem is that different Stata releases behave differently in this respect.
Probably, you should post both files you tried to -append-, so that we can take a look.