Skip to main content
. Author manuscript; available in PMC: 2016 Oct 1.
Published in final edited form as: J R Stat Soc Ser A Stat Soc. 2015 Mar 17;178(4):903–923. doi: 10.1111/rssa.12101

Table 1.

Estimates (standard errors) from the KPNC data. Covariate coefficients are in AFT parameterization (i.e. positive coefficients imply decreased risk).

submodel effect linear HPV duration model log(HPV duration) model
HPV
duration
submodel		 Intercept (β0) 6.63 (0.033) 6.58 (0.036)
baseline age (β1) −0.224 (0.046) −0.24 (0.048)
baseline age2 (β2) 0.064 (0.013) 0.070 (0.013)
scale (τ) 0.79 (0.023) 0.68 (0.033)
frailty SD (σ) 1.065(0.029) 1.21 (0.035)
precancer/
cancer
submodel		 Intercept (α0) 9.10 (0.13) 21.28 (0.81)
baseline age (α1) −0.0042 (0.11) 0.015 (0.12)
baseline age2 (α2) 0.054 (0.032) 0.063 (0.033)
scale (κ) 0.19 (0.017) 0.52 (0.040)
duration (δ) −0.25 (0.021) −1.7 (0.10)
HR: HPV lasts 2 vs. 1 years 3.9 (0.51) 9.9 (1.8)
HR: HPV lasts 3 vs. 2 years 3.9 (0.51) 3.8 (0.41)
HR: HPV lasts 4 vs. 3 years 3.9 (0.51) 2.6 (0.20)
Deviance 24606 24485
AIC 24626 24505