Additive Hazard Regression Models: An Application to the Natural History of Human Papillomavirus

Table 1

Results from semiparametric additive hazard model fitting for oncogenic HPV and any HPV.

	Oncogenic HPV			Any HPV
	Effect estimate	SE*	P value	Effect estimate	SE	P value
HIV/CD4 Count
HIV- (ref^†)	0			0
CD4 > 500	0.0343	0.0075	<0.0001	0.0878	0.0159	<0.0001
CD4: 200–500	0.0779	0.0086	<0.0001	0.2264	0.0214	<0.0001
CD4 < 200	0.1395	0.0163	<0.0001	0.2954	0.0427	<0.0001
Age
<30 (ref)	0			0
30–34	−0.0213	0.0103	0.04	−0.0472	0.0205	0.02
35–39	−0.0425	0.0097	<0.0001	−0.0547	0.0202	0.01
40–44	−0.0502	0.0106	<0.0001	−0.0282	0.0230	0.22
≥45	−0.0408	0.0137	0.003	−0.0623	0.0274	0.02
Race
White (ref)	0			0
Black	0.0140	0.0096	0.15	0.0517	0.0198	0.01
Hispanic	−0.0030	0.0102	0.77	0.0118	0.0212	0.58
Other	−0.0090	0.0191	0.64	0.0661	0.0485	0.17
Smoking
Never (ref)	0			0
Former	0.0034	0.0098	0.73	0.0060	0.0212	0.78
Current	0.0224	0.0079	0.005	0.0495	0.0166	0.003
Sex^‡
1 (ref)	0			0
0	0.0003	0.0082	0.97	−0.0255	0.0162	0.12
2	0.0140	0.0122	0.25	0.0556	0.0256	0.03
≥3	0.0158	0.0128	0.22	0.1217	0.0330	0.0002

∗Standard error.
^†Reference category.
^‡Number of male sexual partner in past 6 months.