Hua L.,Center for Biostatistics in Research |
Zhang Y.,University of Iowa
Biostatistics | Year: 2012
We propose to analyze panel count data using a spline-based semiparametric projected generalized estimating equation (GEE) method with the proportional mean model E((t)Z) = Λ0(t) eβ0TZ. The natural logarithm of the baseline mean function, logΛ0(t), is approximated by a monotone cubic B-spline function. The estimates of regression parameters and spline coefficients are obtained by projecting the GEE estimates into the feasible domain using a weighted isotonic regression (IR). The proposed method avoids assuming any parametric structure of the baseline mean function or any stochastic model for the underlying counting process. Selection of the working covariance matrix that accounts for overdispersion improves the estimation efficiency and leads to less biased variance estimations. Simulation studies are conducted using different working covariance matrices in the GEE to investigate finite sample performance of the proposed method, to compare the estimation efficiency, and to explore the performance of different variance estimates in presence of overdispersion. Finally, the proposed method is applied to a real data set from a bladder tumor clinical trial. The Author 2011. Published by Oxford University Press. All rights reserved.
Li J.Z.,Harvard University |
Gallien S.,Harvard University |
Ribaudo H.,Center for Biostatistics in Research |
Heisey A.,Harvard University |
And 2 more authors.
AIDS | Year: 2014
Objectives:To evaluate the relationship between incomplete antiretroviral therapy (ART) adherence and levels of residual HIV-1 viremia. Design: Medication adherence and residual viremia less than 50 copies/ml were quantified in participants of a cohort of homeless and marginally housed individuals with HIV/AIDS. Methods: Participants had at least 6 months of virologic suppression of less than 50 copies/ml and were in the adherence monitoring cohort with monthly unannounced pill counts. Residual viremia was measured by the single-copy assay. Results: The median average ART adherence over the prior 1 and 2 months were 94% [interquartile range (IQR) 79-100%] and 93% (IQR 82-98%), respectively. Average ART adherence over the past 2 months was significantly associated with levels of residual HIV viremia (Spearman r=-0.25, P = 0.04). One-third of participants with 100% ART adherence over the past 2 months had detectable residual viremia. On multivariate regression analysis, ART adherence over the past 2 months, but not duration of virologic suppression, CD4 + T-cell count or or ART regimen, was significantly associated with levels of residual HIV viremia. Detectable residual viremia was associated with virologic failure (50 copies/ml) on univariate Cox proportional hazard analysis (hazard ratio 2.08, P = 0.02). However, on multivariate analysis, only ART adherence was associated with risk of virologic failure. Conclusion: Incomplete ART adherence is associated with higher levels of residual HIV-1 viremia, but detectable residual viremia can be present despite 100% measured ART adherence. © 2014 Wolters Kluwer Health.
Siberry G.K.,National Health Research Institute |
Li H.,Rush University |
Jacobson D.,Center for Biostatistics in Research
AIDS Research and Human Retroviruses | Year: 2012
The objective of this study was to examine the incidence of fractures in HIV-infected children and comparable HIV-exposed, uninfected (HEU) children in a multicenter, prospective cohort study (PACTG 219/219C) in the United States. The main outcome was first fracture during the risk period. Nine fractures occurred in 7 of 1326 HIV-infected and 2 of 649 HEU children, corresponding to incidence rates of 1.2 per 1000 person-years and 1.1 per 1000 person-years, respectively. The incidence rate ratio was 1.1 (95% CI 0.2, 5.5). There was no evidence of a substantially increased risk of fracture in HIV-infected compared to HEU children. © 2012, Mary Ann Liebert, Inc.
Noubary F.,Center for Biostatistics in Research
Journal of acquired immune deficiency syndromes (1999) | Year: 2012
To evaluate factors affecting antiretroviral therapy (ART) start time when triggered by a CD4 count <350 cells/μL while monitoring counts over time. Measurement frequency, requirement for confirmatory counts, and precision and accuracy of CD4 enumeration technology were considered. Using a model of CD4 count trajectories among seroconverters in the Multicenter AIDS Cohort Study, sequences of counts were simulated for a large hypothetical population monitored for 5 years from seroconversion. Time of first count <350 cells/μL was defined as ART start time. The simulation was adapted to evaluate the effect of the above factors on these times. ART initiation was considered "very late" among patients whose underlying trajectory declined less than 200 cells/μL during the period simulated if no previous observed count was <350 cells/μL. For 12-, 6-, 4-, and 3-monthly measurements, median start time was 48, 36, 32, and 30 months after seroconversion and proportion of patients starting ART very late was 11.5%, 1.6%, 0.2%, and 0.1%. For 6-monthly measurements, requiring confirmation increased the median to 49 months and proportion to 8.9%. Changes in standard deviation of short-term variability in counts of 25% and measurement bias for a novel technology of ±10% changed median time by ±6 months with modest change in the proportion very late (range, 0.5%-3.2%). : 6-monthly measurements appear adequate in achieving low rates of very late ART whereas confirmation affects rates adversely. Studies comparing new versus standard measurement technologies should focus on ruling out modest bias, particularly proximal to important thresholds for treatment management.
Krishnan S.,Center for Biostatistics in Research
Journal of acquired immune deficiency syndromes (1999) | Year: 2012
Metabolic syndrome (MetS) is a cluster of risk factors for cardiovascular disease and diabetes, many of which are associated with HIV and antiretroviral therapy (ART). We examined prevalence and incidence of MetS and risk factors for MetS in ART-naive HIV-infected individuals starting ART. MetS, defined by the Adult Treatment Panel III criteria, was assessed at and after ART initiation in HIV-infected individuals who enrolled in selected AIDS Clinical Trials Group trials and were followed long-term after these trials as part of the AIDS Clinical Trials Group Longitudinal Linked Randomized Trials cohort. Cox proportional hazards models were used to examine risk factors of incident MetS. Adjusted hazard ratios (aHRs) and 95% confidence intervals (CI) are reported. At ART initiation, the prevalence of MetS was 20%. After ART initiation, the incidence of MetS was 8.5 per 100 person-years. After adjusting for demographics and body mass index, the risk of MetS was decreased for CD4+ T-cell counts >50 cells per cubic millimeter (aHR = 0.62, 95% CI = 0.43 to 0.90 for CD4 >500), and the risk was increased for HIV-1 RNA >400 copies per milliliter (aHR = 1.55 (95% CI = 1.25 to 1.92) and use of a protease-inhibitor (PI)-based regimen [relative to no PI use, aHR = 1.25 (95% CI = 1.04 to 1.51) for any PI use]. In HIV-infected individuals on ART, virologic suppression and maintenance of high CD4+ T-cell counts may be potentially modifiable factors that can reduce the risk of MetS. The effect of MetS on the risk of cardiovascular disease and diabetes needs to be evaluated.