I want to know if it's possible to calculate the difference in two incidence rates from a negative binomial regression? We are wanting a measure of absolute (not relative) difference between two time points (26 and 60 - identified as the sequence variable here). This would be the difference (with 95% C.I.'s) between the two values generated by the margins command below (i.e. 7.23 - 1.37 = 5.86). I have figured out how to calculate the relative change with the lincom command but can't see how to do this for the difference. Any help would be appreciated.
Thank you.
Some sample date is provided.
Code:
nbreg n seq, exposure(popn) dispersion(mean) irr margins, at(seq=(26 60)) exp(predict(ir)*100000) lincom _cons + 26*seq, irr // IR at seq = 26 lincom _cons + 60*seq, irr // IR at seq = 60 lincom _cons + 60*seq - _cons - 26*seq, irr // IRR for seq 26 -> seq 60
Code:
* Example generated by -dataex-. To install: ssc install dataex clear input byte month int year float(seq n) double popn 1 2011 1 11 5193457 2 2011 2 29 5193457 3 2011 3 83 5193457 4 2011 4 22 5217264 5 2011 5 20 5217264 6 2011 6 46 5217264 7 2011 7 19 5245972 8 2011 8 27 5245972 9 2011 9 21 5245972 10 2011 10 18 5269429 11 2011 11 22 5269429 12 2011 12 44 5269429 1 2012 13 14 5302688 2 2012 14 37 5302688 3 2012 15 42 5302688 4 2012 16 55 5329763 5 2012 17 66 5329763 6 2012 18 59 5329763 7 2012 19 23 5363139 8 2012 20 32 5363139 9 2012 21 46 5363139 10 2012 22 55 5397799 11 2012 23 41 5397799 12 2012 24 41 5397799 1 2013 25 41 5447630 2 2013 26 30 5447630 3 2013 27 29 5447630 4 2013 28 23 5481356 5 2013 29 46 5481356 6 2013 30 44 5481356 7 2013 31 108 5519750 8 2013 32 117 5519750 9 2013 33 182 5519750 10 2013 34 97 5686981 11 2013 35 142 5686981 12 2013 36 142 5686981 1 2014 37 217 5741247 2 2014 38 162 5741247 3 2014 39 262 5741247 4 2014 40 225.75 5779641 5 2014 41 292.75 5779641 6 2014 42 207 5779641 7 2014 43 288.25 5820836 8 2014 44 215.75 5820836 9 2014 45 200.75 5820836 10 2014 46 239.5 5852229 11 2014 47 193.25 5852229 12 2014 48 238.25 5852229 1 2015 49 209.5 5903810 2 2015 50 303.25 5903810 3 2015 51 338.25 5903810 4 2015 52 548.25 5938587 5 2015 53 238.25 5938587 6 2015 54 223.25 5938587 7 2015 55 230.75 5982816 8 2015 56 462 5982816 9 2015 57 448.25 5982816 10 2015 58 229.5 6016309 11 2015 59 187 6016309 12 2015 60 188.25 6016309 end
0 Response to NBReg - How to calculate difference between two incidence rates
Post a Comment