On the Accelerated Failure Time Model for Current Status and Interval Censored Data

This paper introduces a novel approach to making inference about the regression parameters in the accelerated failure time (AFT) model for current status and interval censored data. The estimator is constructed by inverting a Wald type test for testing a null proportional hazards model. A numerically efficient Markov chain Monte Carlo (MCMC) based resampling method is proposed to simultaneously obtain the point estimator and a consistent estimator of its variance-covariance matrix. We illustrate our approach with interval censored data sets from two clinical studies. Extensive numerical studies are conducted to evaluate the finite sample performance of the new estimators.

Analyzing Panel Count Data with Informative Observation Times

In this paper, we study panel count data with informative observation times. We assume nonparametric and semiparametric proportional rate models for the underlying recurrent event process, where the form of the baseline rate function is left unspecified and a subject-specific frailty variable inflates or deflates the rate function multiplicatively. The proposed models allow the recurrent event processes and observation times to be correlated through their connections with the unobserved frailty; moreover, the distributions of both the frailty variable and observation times are considered as nuisance parameters. The baseline rate function and the regression parameters are estimated by maximizing a conditional likelihood function of observed event counts and solving estimation equations. Large sample properties of the proposed estimators are studied. Numerical studies demonstrate that the proposed estimation procedures perform well for moderate sample sizes. An application to a bladder tumor study is presented to illustrate the use of the proposed methods.

Modeling multilevel sleep transitional data via Poisson log-linear multilevel models

This paper proposes Poisson log-linear multilevel models to investigate population variability in sleep state transition rates. We specifically propose a Bayesian Poisson regression model that is more flexible, scalable to larger studies, and easily fit than other attempts in the literature. We further use hierarchical random effects to account for pairings of individuals and repeated measures within those individuals, as comparing diseased to non-diseased subjects while minimizing bias is of epidemiologic importance. We estimate essentially non-parametric piecewise constant hazards and smooth them, and allow for time varying covariates and segment of the night comparisons. The Bayesian Poisson regression is justified through a re-derivation of a classical algebraic likelihood equivalence of Poisson regression with a log(time) offset and survival regression assuming piecewise constant hazards. This relationship allows us to synthesize two methods currently used to analyze sleep transition phenomena: stratified multi-state proportional hazards models and log-linear models with GEE for transition counts. An example data set from the Sleep Heart Health Study is analyzed.

Nonparametric Estimation of the Bivariate Recurrence Time Distribution

This paper considers statistical models in which two different types of events, such as the diagnosis of a disease and the remission of the disease, occur alternately over time and are observed subject to right censoring. We propose nonparametric estimators for the joint distribution of bivariate recurrence times and the marginal distribution of the first recurrence time. In general, the marginal distribution of the second recurrence time cannot be estimated due to an identifiability problem, but a conditional distribution of the second recurrence time can be estimated non-parametrically. In literature, statistical methods have been developed to estimate the joint distribution of bivariate recurrence times based on data of the first pair of censored bivariate recurrence times. These methods are efficient in the current model because recurrence times of higher orders are not used. Asymptotic properties of the estimators are established. Numerical studies demonstrate the estimator performs well with practical sample sizes. We apply the proposed method to a Denmark psychiatric case register data set for illustration of the methods and theory.

Kernel Estimation of Rate Function for Recurrent Event Data

Recurrent event data are largely characterized by the rate function but smoothing techniques for estimating the rate function have never been rigorously developed or studied in statistical literature. This paper considers the moment and least squares methods for estimating the rate function from recurrent event data. With an independent censoring assumption on the recurrent event process, we study statistical properties of the proposed estimators and propose bootstrap procedures for the bandwidth selection and for the approximation of confidence intervals in the estimation of the occurrence rate function. It is identified that the moment method without resmoothing via a smaller bandwidth will produce curve with nicks occurring at the censoring times, whereas there is no such problem with the least squares method. Furthermore, the asymptotic variance of the least squares estimator is shown to be smaller under regularity conditions. However, in the implementation of the bootstrap procedures, the moment method is computationally more efficient than the least squares method because the former approach uses condensed bootstrap data. The performance of the proposed procedures is studied through Monte Carlo simulations and an epidemiological example on intravenous drug users.

Chlorhexidine and preservation of sound tooth structure in older adults. A placebo-controlled trial

The Trial to Enhance Elderly Teeth Health (TEETH) was designed to test the impact of regular rinsing with a 0.12% chlorhexidine (CHX) solution on tooth loss, and the causes of tooth loss (caries, periodontal disease and trauma) were also investigated. This paper reports on the effectiveness of a 0.12% CHX solution for controlling caries using a tooth surface (coronal and root) survival analysis. A total of 1,101 low income elders in Seattle (United States) and Vancouver (Canada), aged 60-75 years, were recruited for a double-blind clinical trial and assigned to either a CHX (n = 550) or a placebo (n = 551) mouth rinse. Subjects alternated between daily rinsing for 1 month, followed by weekly rinsing for 5 months. All sound coronal and root surfaces at baseline were followed annually for up to 5 years. At each follow-up examination, those tooth surfaces with caries, restored, or extracted were scored as 'carious'. The hazard ratio associated with CHX for a sound surface to become filled, decayed, or extracted was 0.87 for coronal surfaces (95% confidence interval: 0.71-1.14, p = 0.20) and 0.91 for root surfaces (95% confidence interval: 0.73-1.14, p = 0.41). These findings suggest that regular rinsing with CHX does not have a substantial effect on the preservation of sound tooth structure in older adults.

Predicting the evolution of Kaposi sarcoma, in the highly active antiretroviral therapy era

BACKGROUND: The outcome of Kaposi sarcoma varies. While many patients do well on highly active antiretroviral therapy, others have progressive disease and need chemotherapy. In order to predict which patients are at risk of unfavorable evolution, we established a prognostic score. METHOD: The survival analysis (Kaplan-Meier method; Cox proportional hazards models) of 144 patients with Kaposi sarcoma prospectively included in the Swiss HIV Cohort Study, from January 1996 to December 2004, was conducted. OUTCOME ANALYZED: use of chemotherapy or death. VARIABLES ANALYZED: demographics, tumor staging [T0 or T1 (16)], CD4 cell counts and HIV-1 RNA concentration, human herpesvirus 8 (HHV8) DNA in plasma and serological titers to latent and lytic antigens. RESULTS: Of 144 patients, 54 needed chemotherapy or died. In the univariate analysis, tumor stage T1, CD4 cell count below 200 cells/microl, positive HHV8 DNA and absence of antibodies against the HHV8 lytic antigen at the time of diagnosis were significantly associated with a bad outcome.Using multivariate analysis, the following variables were associated with an increased risk of unfavorable outcome: T1 [hazard ratio (HR) 5.22; 95% confidence interval (CI) 2.97-9.18], CD4 cell count below 200 cells/microl (HR 2.33; 95% CI 1.22-4.45) and positive HHV8 DNA (HR 2.14; 95% CI 1.79-2.85).We created a score with these variables ranging from 0 to 4: T1 stage counted for two points, CD4 cell count below 200 cells/microl for one point, and positive HHV8 viral load for one point. Each point increase was associated with a HR of 2.26 (95% CI 1.79-2.85). CONCLUSION: In the multivariate analysis, staging (T1), CD4 cell count (<200 cells/microl), positive HHV8 DNA in plasma, at the time of diagnosis, predict evolution towards death or the need of chemotherapy.

Incidence and outcome of progressive multifocal leukoencephalopathy over 20 years of the Swiss HIV Cohort Study

BACKGROUND: We investigated the incidence and outcome of progressive multifocal leukoencephalopathy (PML) in human immunodeficiency virus (HIV)-infected individuals before and after the introduction of combination antiretroviral therapy (cART) in 1996. METHODS: From 1988 through 2007, 226 cases of PML were reported to the Swiss HIV Cohort Study. By chart review, we confirmed 186 cases and recorded all-cause and PML-attributable mortality. For the survival analysis, 25 patients with postmortem diagnosis and 2 without CD4+ T cell counts were excluded, leaving a total of 159 patients (89 before 1996 and 70 during 1996-2007). RESULTS: The incidence rate of PML decreased from 0.24 cases per 100 patient-years (PY; 95% confidence interval [CI], 0.20-0.29 cases per 100 PY) before 1996 to 0.06 cases per 100 PY (95% CI, 0.04-0.10 cases per 100 PY) from 1996 onward. Patients who received a diagnosis before 1996 had a higher frequency of prior acquired immunodeficiency syndrome-defining conditions (P = .007) but similar CD4+ T cell counts (60 vs. 71 cells/microL; P = .25), compared with patients who received a diagnosis during 1996 or thereafter. The median time to PML-attributable death was 71 days (interquartile range, 44-140 days), compared with 90 days (interquartile range, 54-313 days) for all-cause mortality. The PML-attributable 1-year mortality rate decreased from 82.3 cases per 100 PY (95% CI, 58.8-115.1 cases per 100 PY) during the pre-cART era to 37.6 cases per 100 PY (95% CI, 23.4.-60.5 cases per 100 PY) during the cART era. In multivariate models, cART was the only factor associated with lower PML-attributable mortality (hazard ratio, 0.18; 95% CI, 0.07-0.50; P < .001), whereas all-cause mortality was associated with baseline CD4+ T cell count (hazard ratio per increase of 100 cells/microL, 0.52; 95% CI, 0.32-0.85; P = .010) and cART use (hazard ratio, 0.37; 95% CI, 0.19-0.75; P = .006). CONCLUSIONS: cART reduced the incidence and PML-attributable 1-year mortality, regardless of baseline CD4+ T cell count, whereas overall mortality was dependent on cART use and baseline CD4+ T cell count.