We consider a nonnegative random variable T representing the lifetime of a system. We discuss the residual lifetime $T_X=(T-X|T \gt X)$, where X denotes the random age of the system. We also discuss the mean residual life (MRL) of T at the random time X. It is shown that the MRL at random age (MRLR) is closely related to some well-known variability measures. In particular, we show that the MRLR can be considered a generalization of Gini’s mean difference (GMD). Under the proportional hazards model, we show that the MRLR gives the extended GMD and the extended cumulative residual entropy as special cases. Then, we provide a decomposition result indicating that the MRLR has a covariance representation. Some comparison results are also established for the MRLs of two systems at random ages.

Semiparametric models express a set of distributions of event times in terms of (a) a single parameter which varies across distributions and (b) a single function which does not vary across distributions and which has an unspecified form. These models appear to be attractive alternatives to parametric models of response times in psychometrics. However, our use of such models may require incorporating additional functions which do not vary across distributions and may require expressing the models in terms of the joint distribution of response class and response time.

Duration analyses in political science often model nonproportional hazards through interactions with analysis time. To facilitate their interpretation, methodologists have proposed methods to visualize time-varying coefficients or hazard ratios. While these techniques are a useful, initial postestimation step, I argue that they are insufficient to identify the overall impact of a time-varying effect and may lead to faulty inference when a coefficient changes its sign. I show how even significant changes of a coefficient’s sign do not imply that the overall effect is reversed over time. In order to enable a correct interpretation of time-varying effects in this context, researchers should visualize their results with survivor functions. I outline how survivor functions are calculated for models with time-varying effects and demonstrate the need for such a nuanced interpretation using the prominent finding of a time-varying effect of mediation on interstate conflict. The reanalysis of the data using the proposed visualization methods indicates that the conclusions of earlier mediation research are misleading. The example highlights how survivor functions are an essential tool to clarify the ambiguity inherent in time-varying coefficients in event history models.

Population aging is likely to lead to an increase in the number of people in need of assistance. It is well known that a large part of this assistance originates, and will continue to originate, from the network of relatives and friends. However, the effects of the provision of care on individuals’ employment trajectories when this care is combined with employment of varying intensity or with childcare responsibilities have rarely been examined. The present study used proportional hazards models with the General Social Survey, Cycles 20 and 21, to assess the impact of providing care to a partner, a parent or parent-in-law, another relative, or a non-relative on the risk of leaving employment. The analyses show that providing care to a parent or parent-in-law increases the probability of leaving employment only among women employed full-time and among men and women who have no children or only adult children.

This paper surveys some statistical models of survival data. Competing risks models are described; the unidentifiability of net decrements suggests a sceptical approach to the use of underlying single decrement tables. Approaches based on observations of complete lifetimes (with censoring) are surveyed including the Kaplan-Meier and Nelson-Aalen estimates. Regression models for lifetimes depending on covariates are discussed, in particular the Cox model and partial likelihood estimation.

In this paper we present methods and results for the estimation and modelling of the recovery intensity for Income Protection (IP) insurance claims, allowing for different causes of claim. We use UK data supplied by the Continuous Mortality Investigation relating to claims paid in the years 1975 to 2002, inclusive. Each claim is classified by one of 70 possible causes according to ICD8.

We group causes where appropriate, and then use the Cox model and generalised linear models to model the recovery intensity.

In two subsequent papers we complete our modelling of IP claim termination rates by discussing the modelling of the mortality of IP claimants.

There are two main reasons why it is useful to incorporate cause of sickness in the modelling of IP claim terminations:

(i) The cause of sickness will be known to the insurer for a claim in the course of payment. A reserve can be set more accurately for such a claim if a model of the termination rates appropriate for this cause is available.

(ii) Different causes of claim will become more or less significant over time. For example, tuberculosis may have been an important cause of sickness in the past, but is likely to be far less significant now; the swine flu pandemic starting in 2009 is likely to have a significant effect on observed aggregate claim termination rates, skewing them towards higher rates at shorter durations. Information about trends in morbidity, together with a model of termination rates by cause of claim, allows future aggregate claim termination rates to be predicted more accurately, reserves to be set at more appropriate levels and policies to be priced more accurately.

One of the covariates included in our models for recovery intensities is Calendar Year. Aggregate recovery intensities have been decreasing over the period considered, 1975 to 2002, and this is generally reflected in the models for recovery intensities by cause of sickness. However, when these intensities are projected for years beyond 2002, the results are not always plausible.

Ordinary survival models implicitly assume that all individuals in a group have the same risk of death. It may, however, be relevant to consider the group as heterogeneous, i.e. a mixture of individuals with different risks. For example, after an operation each individual may have constant hazard of death. If risk factors are not included, the group shows decreasing hazard. This offers two fundamentally different interpretations of the same data. For instance, Weibull distributions with shape parameter less than 1 can be generated as mixtures of constant individual hazards. In a proportional hazards model, neglect of a subset of the important covariates leads to biased estimates of the other regression coefficients. Different choices of distributions for the unobserved covariates are discussed, including binary, gamma, inverse Gaussian and positive stable distributions, which show both qualitative and quantitative differences. For instance, the heterogeneity distribution can be either identifiable or unidentifiable. Both mathematical and interpretational consequences of the choice of distribution are considered. Heterogeneity can be evaluated by the variance of the logarithm of the mixture distribution. Examples include occupational mortality, myocardial infarction and diabetes.

A device is repaired at failure. With probability p, it is returned to the ‘good-as-new' state (perfect repair), with probability 1 – p, it is returned to the functioning state, but it is only as good as a device of age equal to its age at failure (imperfect repair). Repair takes negligible time. We obtain the distribution Fp of the interval between successive good-as-new states in terms of the underlying life distribution F. We show that if F is in any of the life distribution classes IFR, DFR, IFRA, DFRA, NBU, NWU, DMRL, or IMRL, then Fp is in the same class. Finally, we obtain a number of monotonicity properties for various parameters and random variables of the stochastic process. The results obtained are of interest in the context of stochastic processes in general, as well as being useful in the particular imperfect repair model studied.

By assuming the proportionality of the intensity functions at each time point for a continuous-time non-homogeneous Markov process, strong ergodicity for the process is determined through strong ergodicity of a related discrete-time Markov process. For processes having proportional intensities, strong ergodicity implies having the limiting matrix L satisfy L · P(s, t) = L, where P(s, t) is the matrix of transition functions.

The usual definition of stochastic comparison of random vectors is extended to stochastic comparison of random processes. Conditions are stated under which {X (t), t ≧ 0} stochastically larger than {Y (t), t ≧ 0} implies that for increasing functionals f.

Applications are made to reliability problems, yielding stochastic comparisons for systems of independently operating machines assuming exponential failure and exponential repair. From these stochastic comparisons we may then deduce similar stochastic comparisons for functionals of practical importance in reliability applications, such as the total machine up-time, the first time that the number of functioning machines drops below a specified number, the total time during which at least a specified number of machines are functioning, etc.