Skip to main content
. Author manuscript; available in PMC: 2020 Jul 17.
Published in final edited form as: Stat Methods Med Res. 2018 Jul 31;28(10-11):2897–2911. doi: 10.1177/0962280218790107

Table 3.

Results of fitting frailty models to the OPTIMA Veterans Cohort

Strategy 1 Shared frailty model Joint frailty model Proposed joint model
HR (95% CI) p-value HR (95% CI) p-value HR (95% CI) p-value
Recurrent ADE 1.11 (0.79, 1.56) 0.554 1.08 (0.64, 1.81) 0.771 1.05 (0.64, 1.72) 0.850
Death NA NA 1.04 (0.55, 1.95) 0.903 1.06 (0.62, 1.80) 0.830
Veteran-level Cluster Estimate (SE) p-value Estimate (SE) p-value Estimate (SE) p-value
Heterogeneity 0.46 (0.20) 0.024 1.49 (0.16) <0.001 1.48 (0.16) <0.001
Association with death  NA  NA 1.17 (0.20) <0.001 0.97 (0.15) <0.001
VAMC-level Cluster Estimate (SE) p-value Estimate (SE) p-value Estimate (SE) p-value
Heterogeneity  NA NA NA NA 0.08 (0.23) 0.737
Association with death  NA NA NA NA 0.67 (0.44) 0.878
Strategy 2 Shared frailty model Joint frailty model Proposed joint model
HR (95% CI) p-value HR (95% CI) p-value HR (95% CI) p-value
Recurrent ADE 1.01 (0.72, 1.42) 0.936 0.95 (0.57, 1.59) 0.845 0.92 (0.56, 1.51) 0.748
Death NA NA 1.20 (0.63, 2.27) 0.586 1.15 (0.68, 1.95) 0.610
Veteran-level Cluster Estimate (SE) p-value Estimate (SE) p-value Estimate (SE) p-value
Heterogeneity 0.46 (0.21) 0.026 1.48 (0.16) <0.001 1.48 (0.16) <0.001
Association with death  NA NA 1.19 (0.21) <0.001 0.97 (0.15) <0.001
VAMC-level Cluster Estimate (SE) p-value Estimate (SE) p-value Estimate (SE) p-value
Heterogeneity NA NA NA NA 0.08 (0.23) 0.719
Association with death NA NA NA NA 0.67 (0.43) 0.877

NA: Not Applicable; ART = antiretroviral treatment, HR = hazard ratio, CI = confidence interval, ADE = AIDS defining event; Shared frailty model is analyzed without level-2 clustering; joint frailty model is analyzed without level-2 clustering; Strategy 1 = intensive antiretroviral treatment (ART) vs. standard ART; Strategy 2 = ART treatment interruption vs. no interruption