Search

Multivariate failure time data arise when the sample consists of clusters and each cluster contains several possibly dependent failure times. The Clayton-Oakes model (Clayton, 1978; Oakes, 1982) for multivariate failure times characterizes the intracluster dependence parametrically but allows arbitrary specification of the marginal distributions. In this paper, we discuss estimation in the Clayton-Oakes model when the marginal distributions are modeled to follow the Cox (1972) proportional hazards regression model. Parameter estimation is based on an approximate generalized maximum likelihood estima- tor. We illustrate the model's application with example datasets.

In comparing the effectiveness of two treatments, suppose that for each patient repeated measurements of an outcome variable are taken at prespecified time points, but some observations may be missing due to the patient's dependent right censoring. In this article, a simple rank estimation procedure, constructed based on an artificial censoring technique, is proposed for the treatment differences over time without imposing a parametric structure on the dependence between the outcome measures and the censoring variable. Our method can be easily implemented and is illustrated by a dataset from an AIDS clinical trial.

To evaluate the reliability of sonographic lung-to-head ratio (LHR) measurement as a predictor of survival in fetuses with congenital diaphragmatic hernia (CDH) and to compare the probability of survival in those with temporary tracheal occlusion (TO) or standard care with respect to the LHR.Fifty-six fetuses with left CDH with liver herniated into the thorax at complete prenatal evaluation were included in logistic regression analyses of antenatal predictors of survival to hospital discharge. Sixteen subjects underwent TO and 40 received standard care.LHR was a significant predictor of survival, with probability of survival increasing with increasing LHR (odds ratio (OR) 8.5, P = 0.04). When subjects with anomalies were excluded, the LHR effect was similar after adjustment for TO (OR 7.1, P = 0.11). Linear spline models suggested a plateau in survival at an LHR of 1.0 and all models suggested increased odds of survival with TO. Minimum LHR measurements had a high degree of inter- and intraobserver agreement (intraclass correlation coefficients of 0.70 and 0.80, respectively).Calculation of the LHR in fetuses with CDH is a reliable and powerful predictor of survival to hospital discharge, although improving odds of survival may plateau at an LHR of 1.0. TO may have an independent benefit on survival to hospital discharge.

The purpose of our study was to determine the effect of tumor-targeted radiation in neuroblastoma by correlating administered (131)I-metaiodobenzylguanidine (MIBG) activity to tumor and whole-body dosimetry, tumor volume change, overall response, and hematologic toxicity.Eligible patients had MIBG-positive lesions and tumor-free, cryopreserved hematopoietic stem cells. Activity was administered according to body weight and protocol as part of a phase I and phase II study. The whole-body radiation dose was derived from daily 1-m exposure measurements, the tumor self-absorbed radiation dose (TSARD) was determined from scintillation-camera conjugate views, and the tumor volume was measured using CT or MRI.Forty-two patients with refractory neuroblastoma (16 with prior hematopoietic stem cell transplant) received a median activity of 555 MBq/kg (15 mCi/kg) (range, 93-770 MBq/kg) and a median total activity of 11,470 MBq (310 mCi) (range, 3,330-30,969 MBq). The median whole-body radiation dose was 228 cGy (range, 57-650 cGy) and the median TSARD was 3,300 cGy (range, 312-30,500 cGy). Responses among evaluable patients included 16 partial response, 3 mixed response, 14 stable disease, and 9 progressive disease. Higher TSARD values predicted better overall disease response (P0.01). The median decrease in tumor volume was 19%; 18 tumors decreased, 4 remained stable, and 5 increased in size. Correlation was seen between administered activity per kilogram and whole-body dose as well as hematologic toxicity (assessed by blood platelet and neutrophil count nadir) (P0.05). The median whole-body dose was higher in the 11 patients who required hematopoietic stem cell infusion for prolonged neutropenia versus the 31 patients who did not (323 vs. 217 cGy; P = 0.03).Despite inaccuracies inherent in dosimetry methods, (131)I-MIBG activity per kilogram correlated with whole-body radiation dose and hematologic toxicity. The TSARD by conjugate planar imaging predicted tumor volume decrease and also correlated with overall tumor response.

This paper describes the properties of a two-stage estimator of the dependence parameter in the Clayton-Oakes multivariate failure time model. The parameter is estimated from a likelihood function in which the marginal hazard functions are replaced by estimates. The method extends the approach of Shih and Louis (1995) and Genest, Ghoudi and Rivest (1995) to allow the marginal hazard for failure times to follow a stratified Cox (1972) model. The method is computationally simple and under mild regularity conditions produces a consistent, asymptotically normal estimator.

In a long term clinical trial to evaluate a new treatment, quite often each study subject may experience a number of 'failures' that correspond to repeated occurrences of the same type of event or events of entirely different natures during his/her follow-up period. To obtain efficient inference procedures for the therapeutic effect over time, it is desirable to utilize those multiple event times in the analysis. In this article, we review some useful procedures for analysing different kinds of multivariate failure time data. Specifically, we discuss the two-sample problems and the general regression problems with various survival models. We also give some recommendations of appropriate procedures for each type of multiple event data structure for practical usage.