Page 384 - HIVMED_v21_i1.indb

Page 9 of 9 Original Research

Appendix 1

The Cox proportional hazard regression model
The fitted Cox proportional hazard regression model was:
()
 ht 
log   =−0.499 X + 0.266 X − 0.467 X − 2.031 X −1.099 X − 0.012 X
 ht  1 2 3 4 5 6
() 

0
Or equivalently,
ht) (
ht) ( = exp(− 0.499 X + 0.266 X − 0.467 X − 2.031 X − 1.099 X − 0.012 X )
4
3
2
5
6
1
0
=
0.607 X + 1.305 X + 0.627 X + 0.131 X + 0.333 X + 0.988 X
1 2 3 4 5 6
where
h(t) = expected hazard of retention at time t
h (t) = baseline hazard of retention at time t
0
X = model of care
1
X = age group
2
X = sex
3
X = disclosure
4
X = regimen
5
X = duration on ART.
6

http://www.sajhivmed.org.za 376 Open Access