A Semiparametric Mixture Model for Analyzing Clustered Competing Risks Data

A dependent Dirichlet process model for survival data with competing risks

In this paper, we first propose a dependent Dirichlet process (DDP) model using a mixture of Weibull models with each mixture component resembling a Cox model for survival data. We then build a Dirichlet process mixture model for competing risks data without regression covariates. Next we extend this model to a DDP model for competing risks regression data by using a multiplicative covariate effect on subdistribution hazards in the mixture components. Though built on proportional hazards (or subdistribution hazards) models, the proposed nonparametric Bayesian regression models do not require the assumption of constant hazard (or subdistribution hazard) ratio. An external time-dependent covariate is also considered in the survival model. After describing the model, we discuss how both cause-specific and subdistribution hazard ratios can be estimated from the same nonparametric Bayesian model for competing risks regression. For use with the regression models proposed, we introduce an omnibus prior that is suitable when little external information is available about covariate effects. Finally we compare the models' performance with existing methods through simulations. We also illustrate the proposed competing risks regression model with data from a breast cancer study. An R package "DPWeibull" implementing all of the proposed methods is available at CRAN.

Inequalities in HAART uptake and differential survival according to exposure category in Rio de Janeiro, Brazil

Despite substantial improvement in prognosis and quality of life among people living with HIV/AIDS (PLWHA) in Brazil, inequalities in access to treatment remain. We assessed the impact of these inequalities on survival in Rio de Janeiro over a 12-year period (2000/11). Data were merged from four databases that comprise the national AIDS monitoring system: SINAN-AIDS (Brazilian Information System for Notificable Diseases; AIDS cases), SISCEL (laboratory tests), SICLOM (electronic dispensing system), and SIM (Brazilian Mortality Information System), using probabilistic linkage. Cox regressions were fitted to assess the impact of HAART (highly active antiretroviral therapy) on AIDS-related mortality among men who have sex with men (MSM), people who inject drugs (PWID), and heterosexuals diagnosed with AIDS, between 2000 and 2011, in the city of Rio de Janeiro, RJ, Brazil. Among 15,420 cases, 60.7% were heterosexuals, 36.1% MSM and 3.2% PWID. There were 2,807 (18.2%) deaths and the median survival time was 6.29. HAART and CD4+ > 200 at baseline were associated with important protective effects. Non-whites had a 33% higher risk of dying in consequence of AIDS than whites. PWID had a 56% higher risk and MSM a 11% lower risk of dying of AIDS than heterosexuals. Non-white individuals, those with less than eight years of formal education, and PWID, were more likely to die of AIDS and less likely to receive HAART. Important inequalities persist in access to treatment, resulting in disparate impacts on mortality among exposure categories. Despite these persistent disparities, mortality decreased significantly during the period for all categories under analysis, and the overall positive impact of HAART on survival has been dramatic.

Differential survival benefit of universal HAART access in Brazil: a nation-wide comparison of injecting drug users versus men who have sex with men

OBJECTIVE

Brazil accounts for approximately 70% of injection drug users (IDUs) receiving highly active antiretroviral therapy (HAART) in low-income/middle-income countries. We evaluated the impact of HAART availability/access on AIDS-related mortality among IDUs versus men who have sex with men (MSM).

DESIGN

Nation-wide analysis on Brazilian IDU and MSM diagnosed with AIDS in 2000-2006.

METHODS

Four national information systems were linked, and Cox regression was used to assess impact of HAART availability/access on differential AIDS-related mortality.

RESULTS

Among 28,426 patients, 6777 died during 87,792 person-years of follow-up. Compared with MSM, IDU were significantly less likely to be receiving HAART, to have ever had determinations for CD4 or viral load. After controlling for confounders, IDU had a significantly higher risk of death (adjusted hazard ratio: 1.94; 95% confidence interval: 1.84 to 2.05). Among the subset that had at least 1 CD4 and viral load determination, higher risk of death among IDU persisted (hazard ratio: 1.82; 95% confidence interval: 1.58 to 2.11). Nonwhite ethnicity significantly increased this risk, whereas prompt HAART uptake after AIDS diagnosis reduced the risk of death. After controlling for spatially correlated survival data, AIDS-related mortality remained higher in IDU than in MSM.

CONCLUSIONS

Despite free/universal HAART access, differential AIDS-related mortality exists in Brazil. Efforts are needed to identify and eliminate these health disparities.

Statistical methods for analysing longitudinal data in delirium studies

Delirium is an often acute and highly fluctuating syndrome that can be transient or in some cases associated with prolonged disturbances. The best way to capture its natural course is to conduct studies with longitudinal design, but data analysis in longitudinal studies is difficult, as often the measured variables of each subject are correlated over the course of time. As such, there has been limited application of such methods for analysing longitudinal data in the study of delirium. This overview considers simple traditional approaches along with more complex methods that involve modelling of data. The relative merits of survival analysis, structural equation modelling, and path analysis are reviewed. Furthermore, two flexible modelling techniques are considered; the mixed effects model and generalized estimating equations with emphasis on their use with binary outcomes, as often the outcome in delirium studies is delirium/no delirium. Their contrasting approach to parameter interpretation and methods for accounting for correlation and dealing with missing data are detailed. Information on available software is provided. Delirium research will be substantially enhanced by incorporating such methods.

Effectiveness of tibolone on the reduction of menopausal problems—a Bayesian semi-parametric interpretation

Two major physiological problems women experience at the time of menopause are hot flush and vaginal dryness. Exploratory investigations reveal that these two binary outcomes are very much dependent as both of them have predominant oestrogenic effects. A primary interest is to investigate how the bivariate association and the marginal univariate risks are affected by repeated measurements on each woman over several months. To achieve this we propose a very general class of bivariate binary models. Parametric inference is drawn on the basis of full non-parametric Bayesian approach under Dirichlet process mixture. Study addresses some more interesting phenomena on the effectiveness of tibolone treatment in reducing menopausal problems. A simulation study further strengthens the proposed methodology.