Time-dependent Variables (time-dependent + variable)

Distribution by Scientific Domains

Selected Abstracts

Predicting short-term disease progression among HIV-infected patients in Asia and the Pacific region: preliminary results from the TREAT Asia HIV Observational Database (TAHOD)

HIV MEDICINE, Issue 3 2005
J Zhou
Objectives HIV disease progression has been well documented in Western populations. This study aimed to estimate the short-term risk of AIDS and death from the TREAT Asia HIV Observational Database (TAHOD), a prospective, multicentre cohort study in Asia and the Pacific region. Methods Prospective data were analysed to estimate short-term disease progression. Endpoints were defined as the time from study entry to diagnosis with AIDS or death. Antiretroviral treatment was fitted as a time-dependent variable. Predictors of disease progression were assessed using Cox proportional hazards models, and prognostic models were developed using Weibull models. Results A total of 1260 patients with prospective follow-up data contributed 477 person-years of follow-up, during which 18 patients died and 34 were diagnosed with AIDS, a combined rate of 10.1 per 100 person-years. Compared with patients receiving antiretroviral treatment, patients not on treatment had a higher rate of disease progression (17.6 vs. 8.1 per 100 person-years, respectively). Baseline CD4 count was the strongest predictor of disease progression. Prognostic models, using either a baseline CD4 count as the sole marker or markers including baseline haemoglobin, AIDS-related symptoms and previous or current antiretroviral treatment, were successful at identifying patients at high risk of short-term disease progression. Conclusions Similar to the situation in Western countries, baseline CD4 count was the strongest predictor of short-term disease progression. Prognostic models based on readily available clinical data and haemoglobin level should be useful in estimating short-term clinical risk in HIV-infected patients in Asia and the Pacific region. [source]

Time-Dependent ROC Curves for Censored Survival Data and a Diagnostic Marker

BIOMETRICS, Issue 2 2000
Patrick J. Heagerty
Summary. ROC curves are a popular method for displaying sensitivity and specificity of a continuous marker, X, for a binary disease variable, D. However, many disease outcomes are time dependent, D(t, and ROC curves that vary as a function of time may be mire appropriate. A common examples of a time-dependent variable is vital status, where D(t) = 1 if a patient has died prior to time t and zero otherwise. We propose summarizing the discrimination potential of a marker X, measured at baseline (t= 0), by calculating ROC Curves for cumulative disease or death incidence by time t, which we denote as ROC(t). A typical complexity with survival data is that observations may be censored. Two ROC curve estimators are proposed that can accommodate censored data. A simple estimator is based on using the Kaplan-Meier estimated for each possible subset X > c. However, this estimator does not guarantee the necessary condition that sensitivity and specificity are monotone in X. An alternative estimator that does guarantee monotonicity is based on a nearest neighbor estimator for the bivariate distribution function of (X, T), where T represents survival time (Akritas, M. J., 1994, Annals of Statistics22, 1299,1327). We present an example where ROC(t) is used to compare a standard and a modified flow cytometry measurement for predicting survival after detection of breast cancer and an example where the ROC(t) curve displays the impact of modifying eligibility criteria for sample size and power in HIV prevention trials. [source]

A Semivirtual Watershed Model by Neural Networks

James C. Y. Guo
A semivirtual watershed model is presented in this study. This model places the design rainfall distribution on the input layer and the predicted runoff hydrograph on the output layer. The optimization scheme developed in this study can train the model to establish a set of weights under the guidance of the kinematic wave theory. The weights are time-dependent variables by which rainfall signals can be converted to runoff distributions by weighting procedures only. With the consideration of time dependence, the computational efficiency of virtual watershed models is greatly enhanced by eliminating unnecessary visitations between layers. The weighting procedure used in the semivirtual watershed model expands the rational method from peak runoff predictions to complete hydrograph predictions under continuous and nonuniform rainfall events. [source]

Depression and Retirement in Late Middle-Aged U.S. Workers

Jalpa A. Doshi
Objective. To determine whether late middle-aged U.S. workers with depression are at an increased risk for retirement. Data Source. Six biennial waves (1992,2002) of the Health and Retirement Study, a nationally representative panel survey of noninstitutionalized 51,61-year-olds and their spouses started in 1992. Study Design. Workers aged 53,58 years in 1994 were followed every 2 years thereafter, through 2002. Depression was coded as lagged time-dependent variables measuring active depression and severity of depression. The main outcome variable was a transition to retirement which was measured using two distinct definitions to capture different stages in the retirement process: (1) Retirement was defined as a transition out of the labor force in the sample of all labor force participants (N=2,853); (2) In addition a transition out of full time work was used as the retirement definition in the subset of labor force participants who were full time workers (N=2,288). Principal Findings. In the sample of all labor force participants, the presence of active depression significantly increased the hazard of retirement in both late middle-aged men (adjusted OR: 1.37 [95 percent CI 1.05, 1.80]) and women (adjusted OR: 1.40 [95 percent CI 1.10, 1.78]). For women, subthreshold depression was also a significant predictor of retirement. In the sample of full time workers, the relationship between depression and retirement was considerably weaker for women yet remained strong for men. Conclusions. Depression and depressive symptoms were significantly associated with retirement in late middle-aged U.S. workers. Policymakers must consider the potentially adverse impact of these labor market outcomes when estimating the cost of untreated depression and evaluating the value of interventions to improve the diagnosis and treatment of depression. [source]

Cognitive Status, Muscle Strength, and Subsequent Disability in Older Mexican Americans

Mukaila A. Raji MD
Objectives: To examine the association between Mini-Mental State Examination (MMSE) score and subsequent muscle strength (measured using handgrip strength) and to test the hypothesis that muscle strength will mediate any association between impaired cognition and incident activity of daily living (ADL) disability over a 7-year period in elderly Mexican Americans who were initially not disabled. Design: A 7-year prospective cohort study (1993,2001). Setting: Five southwestern states (Texas, New Mexico, Colorado, Arizona, and California). Participants: Two thousand three hundred eighty-one noninstitutionalized Mexican-American men and women aged 65 and older with no ADL disability at baseline. Measurements: In-home interviews in 1993/1994, 1995/1996, 1998/1999, and 2000/2001 assessed social and demographic factors, medical conditions (diabetes mellitus, stroke, heart attack, and arthritis), body mass index (BMI), depressive symptomatology, handgrip muscle strength, and ADLs. MMSE score was dichotomized as less than 21 for poor cognition and 21 or greater for good cognition. Main outcomes measures were mean and slope of handgrip muscle strength over the 7-year period and incident disability, defined as new onset of any ADL limitation at the 2-, 5-, or 7-year follow-up interview periods. Results: In mixed model analyses, there was a significant cross-sectional association between having poor cognition (MMSE<21) and lower handgrip strength, independent of age, sex, and time of interview (estimate=,1.41, standard error (SE)=0.18; P<.001). With the introduction of a cognition-by-time interaction term into the model, there was also a longitudinal association between poor cognition and change in handgrip strength over time (estimate=,0.25, SE=0.06; P<.001), indicating that subjects with poor cognition had a significantly greater decline in handgrip strength over 7 years than those with good cognition, independent of age, sex, and time. This longitudinal association between poor cognition and greater muscle decline remained significant (P<.001) after controlling for age, sex, education, and time-dependent variables of depression, BMI, and medical conditions. In general estimation equation models, having poor cognition was associated with greater risk of 7-year incident ADL disability (odds ratio=2.01, 95% confidence interval (CI)=1.60,2.52); the magnitude of the association decreased to 1.66 (95% CI=1.31,2.10) when adjustment was made for handgrip strength. Conclusion: Older Mexican Americans with poor cognition had steeper decline in handgrip muscle strength over 7 years than those with good cognition, independent of other demographic and health factors. A possible mediating effect of muscle strength on the association between poor cognition and subsequent ADL disability was also indicated. [source]

Mortality after uveal and conjunctival melanoma: which tumour is more deadly?

Emma Kujala
Abstract. Purpose:, We aimed to model and compare mortality rates for uveal melanoma (UM) and conjunctival melanoma (CM) by adjusting for differences in tumour size and local recurrence. Methods:, Population-based mortality data for 240 and 85 patients with primary UM and CM and 91 and 23 patients with disseminated UM and CM, respectively, were compared with cumulative incidence analysis. Cox proportional hazards multivariate regression with time-dependent variables was used to adjust for differences in tumour diameter, thickness and recurrence rates. Results:, The 10-year cumulative incidences of metastatic death from UM and CM were 39% (95% confidence interval [CI] 33,45) and 32% (95% CI 21,44), respectively. After adjusting for tumour size, risk of death from CM was higher than from UM (hazard ratio [HR] 1.9; p = 0.039). Additional adjustment for more frequent local recurrence of CM diminished the difference (HR 1.5; p = 0.25). Survival periods after systemic metastasis of UM and CM were comparable (median 8 months). Conclusions:, Clinical observations show longer survival after primary CM than after primary UM. This reflects the smaller average size of CM. However, a primary CM of a given size is more deadly than a UM of equivalent size because primary CM tends to recur after treatment and, possibly, because additional lymphatic dissemination occurs with CM. [source]