AbstractRecurrent event data are frequently encountered in longitudinal studies. In many applications, the times between successive recurrent events (gap times) are often of interest and lead to problems that have received much attention recently. In this article, using the approach of inverse probability-of-censoring weights (IPCW), we propose nonparametric estimators for the estimation of the bivariate distribution and survival functions for gap times of recurrent event data. We also consider the estimation of Kendall’s tau for two gap times by expressing it as an integral functional of the bivariate survival function. The asymptotic properties of the proposed estimators are established. Simulation studies are conducted to investigate their finite sample performance. 相似文献
If the years of late adolescence and early adulthood are depicted as rocky, disorderly, and experimental, then how should the life course stage that immediately follows be characterized? Using data from the National Longitudinal Study of the High School Class of 1972, we focus on work and school activities in the latter half of the young adult years (ages roughly 25–31) to empirically address whether the diversity of work and schooling patterns diminishes. Finding a lack of orderliness in work and schooling trajectories, and that holding two or more activities simultaneously is quite common, we question how this stage of the life course should be conceptualized. Moving away from the traditional sequential roles approach, we use multinomial logistic regression to differentiate among activity sequences commonly held by young men and women at these ages, and address how family of origin background factors and prior life experiences influence life course experiences in the later 20s and early 30s.
Joint modeling of recurrent and terminal events has attracted considerable interest and extensive investigations by many authors. The assumption of low-dimensional covariates has been usually applied in the existing studies, which is however inapplicable in many practical situations. In this paper, we consider a partial sufficient dimension reduction approach for a joint model with high-dimensional covariates. Some simulations as well as three real data applications are presented to confirm and assess the performance of the proposed model and approach. 相似文献
For analyzing recurrent event data, either total time scale or gap time scale is adopted according to research interest. In particular, gap time scale is known to be more appropriate for modeling a renewal process. In this paper, we adopt gap time scale to analyze recurrent event data with repeated observation gaps which cannot be observed completely because of unknown termination times of observation gaps. In order to estimate termination times, interval-censored mechanism is applied. Simulation studies are done to compare the suggested methods with the unadjusted method ignoring incomplete observation gaps. As a real example, conviction data set with suspensions is analyzed with suggested methods. 相似文献
Despite findings that depression is a risk factor for heart disease and for death following cardiac events and that depressed
cardiac patients experience significantly reduced quality of life and are less likely to follow treatment regimens, depression
is neither adequately identified nor treated in cardiac patients. Recent calls in the literature for the use of standardized
screening measures and sensitivity/specificity studies to identify useful measures compelled us to examine the sensitivity
and specificity of the Beck Depression Inventory-II (BDI-II) and Geriatric Depression Scale (GDS) and recommend appropriate
cut-scores for identifying depression in post-myocardial infarction or unstable angina patients. A total of 119 patients who
met criteria for either acute myocardial infarction or unstable angina pectoris were recruited from coronary care units at
three hospitals and interviewed in their homes approximately 2 weeks post-admission. The criterion used in the study was the
Structured Clinical Interview for DSM-IV-TR Axis I Disorders (SCID-I/NP). Administration of the SCID-I/NP, BDI-II, and GDS was counterbalanced using a digram-balanced
approach and blinded comparison was used. Alphas were .89 for BDI-II and .88 for GDS. For major depression, a BDI-II cut-score
of 10 produced a sensitivity = 100%, specificity = 75%, and PPV = 18% whereas a GDS cut-score of 14 produced sensitivity = 100%,
specificity = 94%, and PPV = 50%. For major/double depression, a BDI-II cut-score of 10 produced sensitivity = 100%, specificity = 75%,
and PPV = 21% whereas a GDS cut-score of 13 produced sensitivity = 100%, specificity = 91%, and PPV = 41%. Although both measures
demonstrated excellent reliability and sensitivity, the GDS showed better specificity and PPV and is recommended as the better
screen for major depression or double depression with cardiac patients. 相似文献
Longitudinal network data recording the moment at which ties appear, change, or disappear are increasingly available. Event history models can be used to analyze the dynamics of time-stamped network data. This paper adapts the discrete-time event history model to social network data. A discrete-time event history model can easily incorporate a multilevel design and time-varying covariates. A multilevel design is needed to account for dependencies among ties and vertices, which should not be ignored in a small longitudinal network. Time-varying covariates are required to analyze network effects, that is, the impact of previous ties. In addition, a discrete-time event history model handles constraints on who can act or who can be acted upon in a straightforward way. The model can be estimated with multilevel logistic regression analysis, which is illustrated by an application to book reviews, so network evolution can be analyzed with a fairly standard statistical tool. 相似文献