Hi all, I'm trying to graph mean utility scores collected over time in a maximum of four surveys, where the dataset has been xtset with the panel variable of patient ID, and time variable of survey response. I am currently using
to get the mean utility score for both groups by long covid (0=no, 1=yes) by survey response. This is a snapshot of what the (dummy) data looks like, with patient ID, the survey response, exposure dummy and finally utility score.
The issue I have is that some participants (denoted by patient_id) have completed 2 surveys, some 3 and finally the maximum of 4, and combining them using
produces a line that appears to show a downward gradient, but this is more likely loss to follow-up. The question I therefore have is how would you change the graph so as to have one line that shows the mean scores if participants have only completed 2 surveys, then further lines for 3 and 4 responses respectfully. I'm hoping this will show the current gradient just to be from the most affected filling all surveys, whereas least affected may only do 2. I am using Stata MP v. 16.1 and within a secure research environment so user created commands are not allowed unfortunately. Many thanks for any help! Ollie
Code:
egen mean = mean(utility), by(survey_response long_covid)
Code:
input double(patient_id survey_response) float(long_covid utility) 15 2 0 .9865437 21 1 0 .975916 38 1 0 .7132339 41 1 0 .5275346 42 1 0 .9831794 49 1 0 .6651402 51 1 0 .52302635 52 1 0 .6507832 52 2 0 .8427859 57 1 0 -.07101776 57 2 0 .5238919 57 3 0 .57239664 62 3 0 .9750341 62 2 0 .9750341 62 1 0 .9750341 62 4 1 .9750341 63 2 0 .4941897 63 1 0 .6041356 65 2 0 .987404 68 1 0 -.16829173 70 1 0 .353614 72 1 1 -.06054749 75 1 0 .9796785 76 2 0 .10147964 77 1 0 .9751386 77 2 0 .9751386 82 1 0 .648912 84 1 0 .7297567 94 1 0 .9783098 94 2 0 .9783098 94 3 0 .9783098 94 4 0 .9783098 97 3 1 .9755546 100 2 1 .9879804 100 3 1 .9879804 100 1 0 .9879804 102 3 0 -.3323491 102 1 0 .2568085 104 2 0 -.04498811 104 1 0 .3425053 108 1 1 .27844697 110 2 0 .9834457 116 4 0 .987404 116 1 0 .987404 118 1 0 .5605557 119 1 1 .08590872
Code:
twoway (tsline mean if long_covid==0) (tsline mean if long_covid==1)
