Add line to sts graph (kaplan meier)

[Lytfi Krasniqi]

Im having trouble with overlaying two graphs.


I have 3208 observations for my survival graph (curve1).
I have 16 coordinates for curve2

---

 *bagX
set obs 3224
gen bagX = .
replace bagX = 0 if _n == 3209
replace bagX = 1 if _n == 3210
replace bagX = 2 if _n == 3211
replace bagX = 3 if _n == 3212
replace bagX = 4 if _n == 3213
replace bagX = 5 if _n == 3214
replace bagX = 6 if _n == 3215
replace bagX = 7 if _n == 3216
replace bagX = 8 if _n == 3217
replace bagX = 9 if _n == 3218
replace bagX = 10 if _n == 3219
replace bagX = 11 if _n == 3220
replace bagX = 12 if _n == 3221
replace bagX = 13 if _n == 3222
replace bagX = 14 if _n == 3223
replace bagX = 15 if _n == 3224

*bagY
gen bagY = .
replace bagY = 100 if bagX == 0
replace bagY = 96.5 if bagX == 1
replace bagY = 92.9 if bagX == 2
replace bagY = 89 if bagX == 3
replace bagY = 85 if bagX == 4
replace bagY = 80.8 if bagX == 5
replace bagY = 76.3 if bagX == 6
replace bagY = 71.7 if bagX == 7
replace bagY = 66.9 if bagX == 8
replace bagY = 62 if bagX == 9
replace bagY = 57 if bagX == 10
replace bagY = 52 if bagX == 11
replace bagY = 46.9 if bagX == 12
replace bagY = 41.9 if bagX == 13
replace bagY = 37 if bagX == 14
replace bagY = 32.3 if bagX == 15


sts graph, by(ES2)ylabel(1 "100" .8 "80" .6 "60" .4 "40" .2 "20" .0 "0") ytitle("Mortality") xlabel(0(2.5)15) tmax(15) addplot(line bagY bagX)

---

The result is a abnorm graph all over the place.
What am I missing?

[Andrew Musau]

The -addplot- option amounts to

tw line bagY bagX

which produces

So with your data example, there is no way to determine what

The result is a abnorm graph all over the place.

means. I think the sts graph is therefore important in explaining your problem.

[Lytfi Krasniqi]

The sts graph works fine on its own. When I add addplot(line bagY bagY), the graph from sts graph still work the way it should, but the line is "all over the place" meaning that it swings out of the white area, and swings twice.

twoway doesn't work when using sts graph, so im looking for another way.. and I can't get addplot to work.
Here is the code for sts graph:

use "F:\Projekter\FSEID00003877\STATA\MACCE1604_KMK_Efter_de4.dta", clear

gen OPDATO2 = date(OPDATO, "YMD"), after(OPDATO)
format %td OPDATO2

gen DEATHDATE2 = date(DEATHDATE, "YMD"), after(DEATHDATE)
format %td DEATHDATE2

gen date_difference = (DEATHDATE2 - OPDATO2)/365.25

replace date_difference = 0.0000001 if date_difference == 0

stset date_difference, failure(ALIVE1_DEAD2==2)


***ES
gen ES2=.
replace ES2=1 if EUROSCORE2 < 2
replace ES2=2 if EUROSCORE2 >= 2

*****

expand 2, gen(dub)
replace ES2 = 9 if dub == 1     //  EUROSCORE2 in three groups (9 = total)

sts graph, by(ES2)ylabel(1 "100" .8 "80" .6 "60" .4 "40" .2 "20" .0 "0") ytitle("Mortality") xlabel(0(2.5)15) tmax(15)  addplot(line bagY bagX)
stdescribe

[Andrew Musau]

I will need to see your data to be able to find out what is going on. If you are not able to create a reproducible example using one of the Stata datasets (or fake data that you have created), post your real data using dataex or upload it.

[Lytfi Krasniqi]

I'm afraid that it's not possible to upload my data and I'm not able to reproduce an example..
Is there a way to add the coordinates as a function with a range of 0 to 15 to the sts graph?

Because the coordinates can be describes as: y=(0.0046x^3)-(0.1602x^2)-(3.1362x)+99.862)

[Andrew Musau]

For the added plot to fit exactly into the overlaid graph, the ranges of the y axis and x axis should match. It may be just a simple scaling problem, but I cannot help much without a reproducible example.

[Lytfi Krasniqi]

I understand Thank you for your response.