Abstract:ObjectiveTo quantify the impact of tuberculosis (TB) co-infection on death and loss to follow-up (LTFU) 12 months after entry into an ART program. DesignProspective intervention study. MethodsFrom May 2007 to 2008, patients undergoing pre-ART training in Durban, South Africa, were screened for pulmonary TB using mycobacterial culture. Subjects missing appointments for >3 months were phoned. Patients who could not be reached were considered LTFU. Deaths were ascertained by report from family members. We used the Kaplan–Meier method to estimate time to LTFU or death for 3 groups at enrollment as follows: (1) newly diagnosed with TB by sputum culture; (2) on TB treatment (ie, previously diagnosed); and (3) TB free. We evaluated the role of TB on mortality and LTFU using Cox proportional hazards models. ResultsNine hundred fifty-one HIV-infected subjects were enrolled; 59% were female, and median baseline CD4 count was 90 cells per microliter (IQR: 41–148 cells/&mgr;L). One hundred forty-four (15%) were newly diagnosed with TB by sputum culture; an additional 199 (21%) were already on TB treatment. By 12 months, 26% newly diagnosed with TB at enrollment died or were LTFU, compared with 19% already on TB treatment, and 14% who were TB free (P = 0.001). Controlling for age, sex, smoking, CD4, and opportunistic infection history, subjects newly diagnosed with pulmonary TB were 76% more likely to die or be LTFU (hazard ratio: 1.76, 95% confidence interval: 1.20 to 2.60) than those without TB. ConclusionsHIV/TB co-infected individuals are more likely to die or be LTFU within 12 months of ART clinic entry in South Africa. These patients require intensive follow-up during ART initiation.

Maximum likelihood estimation of the Weibull distribution with reduced bias

The Maxwell Length-Biased Distribution: Properties and Estimation

A new approach to estimate parameters of the two-parameter Weibull distribution

Penalized bias reduction in extreme value estimation for censored Pareto-type data, and long-tailed insurance applications

Maximum Likelihood Estimations and EM Algorithms with Length-biased Data

A NEW PERSPECTIVE ON ROBUST M-ESTIMATION: FINITE SAMPLE THEORY AND APPLICATIONS TO DEPENDENCE-ADJUSTED MULTIPLE TESTING

Analysis of Weibull progressively first-failure censored data with beta-binomial removals

On Birnbaum-Saunders Inference

Empirical Likelihood-Based Inference with Missing and Censored Data

On the Maximum Penalized Full Likelihood Approach for Cox Model with Extreme Value for Heavily Censored Survival Data

Firth adjustment for Weibull current-status survival analysis

Loss to Follow-Up and Mortality Among HIV-Infected People Co-Infected With TB at ART Initiation in Durban, South Africa

Empirical Likelihood-base Inference in Linear Models with Censored Data

Analysis of the new complementary unit Weibull model from adaptive progressively type-II hybrid

Robust estimation and bias-corrected empirical likelihood in generalized linear models with right censored data

Shortest Confidence Intervals for Weibull Modulus for Small Samples in Materials Reliability Analysis

A maximum-mean-discrepancy goodness-of-fit test for censored data

A method of maximum likelihood fit to data with non-uniform efficiencies

Mean and median bias reduction: A concise review and application to adjacent-categories logit models

Empirical bias-reducing adjustments to estimating functions

Estimating the failure rate of the log-logistic distribution by smooth adaptive and bias-correction methods