A review on reliability models with covariates

Modern Engineering Asset Management (EAM) requires the accurate assessment of current and the prediction of future asset health condition. Suitable mathematical models that are capable of predicting Time-to-Failure (TTF) and the probability of failure in future time are essential. In traditional reliability models, the lifetime of assets is estimated using failure time data. However, in most real-life situations and industry applications, the lifetime of assets is influenced by different risk factors, which are called covariates. The fundamental notion in reliability theory is the failure time of a system and its covariates. These covariates change stochastically and may influence and/or indicate the failure time. Research shows that many statistical models have been developed to estimate the hazard of assets or individuals with covariates. An extensive amount of literature on hazard models with covariates (also termed covariate models), including theory and practical applications, has emerged. This paper is a state-of-the-art review of the existing literature on these covariate models in both the reliability and biomedical fields. One of the major purposes of this expository paper is to synthesise these models from both industrial reliability and biomedical fields and then contextually group them into non-parametric and semi-parametric models. Comments on their merits and limitations are also presented. Another main purpose of this paper is to comprehensively review and summarise the current research on the development of the covariate models so as to facilitate the application of more covariate modelling techniques into prognostics and asset health management.

Modeling crash types: New insights into the effects of covariates on crashes at rural intersections

Many studies focused on the development of crash prediction models have resulted in aggregate crash prediction models to quantify the safety effects of geometric, traffic, and environmental factors on the expected number of total, fatal, injury, and/or property damage crashes at specific locations. Crash prediction models focused on predicting different crash types, however, have rarely been developed. Crash type models are useful for at least three reasons. The first is motivated by the need to identify sites that are high risk with respect to specific crash types but that may not be revealed through crash totals. Second, countermeasures are likely to affect only a subset of all crashes—usually called target crashes—and so examination of crash types will lead to improved ability to identify effective countermeasures. Finally, there is a priori reason to believe that different crash types (e.g., rear-end, angle, etc.) are associated with road geometry, the environment, and traffic variables in different ways and as a result justify the estimation of individual predictive models. The objectives of this paper are to (1) demonstrate that different crash types are associated to predictor variables in different ways (as theorized) and (2) show that estimation of crash type models may lead to greater insights regarding crash occurrence and countermeasure effectiveness. This paper first describes the estimation results of crash prediction models for angle, head-on, rear-end, sideswipe (same direction and opposite direction), and pedestrian-involved crash types. Serving as a basis for comparison, a crash prediction model is estimated for total crashes. Based on 837 motor vehicle crashes collected on two-lane rural intersections in the state of Georgia, six prediction models are estimated resulting in two Poisson (P) models and four NB (NB) models. The analysis reveals that factors such as the annual average daily traffic, the presence of turning lanes, and the number of driveways have a positive association with each type of crash, whereas median widths and the presence of lighting are negatively associated. For the best fitting models covariates are related to crash types in different ways, suggesting that crash types are associated with different precrash conditions and that modeling total crash frequency may not be helpful for identifying specific countermeasures.

Handling incomplete data in survival analysis with multiple covariates

This paper studies the missing covariate problem which is often encountered in survival analysis. Three covariate imputation methods are employed in the study, and the effectiveness of each method is evaluated within the hazard prediction framework. Data from a typical engineering asset is used in the case study. Covariate values in some time steps are deliberately discarded to generate an incomplete covariate set. It is found that although the mean imputation method is simpler than others for solving missing covariate problems, the results calculated by it can differ largely from the real values of the missing covariates. This study also shows that in general, results obtained from the regression method are more accurate than those of the mean imputation method but at the cost of a higher computational expensive. Gaussian Mixture Model (GMM) method is found to be the most effective method within these three in terms of both computation efficiency and predication accuracy.

Intelligent reliability analysis with incomplete covariates

The reliability analysis is crucial to reducing unexpected down time, severe failures and ever tightened maintenance budget of engineering assets. Hazard based reliability methods are of particular interest as hazard reflects the current health status of engineering assets and their imminent failure risks. Most existing hazard models were constructed using the statistical methods. However, these methods were established largely based on two assumptions: one is the assumption of baseline failure distributions being accurate to the population concerned and the other is the assumption of effects of covariates on hazards. These two assumptions may be difficult to achieve and therefore compromise the effectiveness of hazard models in the application. To address this issue, a non-linear hazard modelling approach is developed in this research using neural networks (NNs), resulting in neural network hazard models (NNHMs), to deal with limitations due to the two assumptions for statistical models. With the success of failure prevention effort, less failure history becomes available for reliability analysis. Involving condition data or covariates is a natural solution to this challenge. A critical issue for involving covariates in reliability analysis is that complete and consistent covariate data are often unavailable in reality due to inconsistent measuring frequencies of multiple covariates, sensor failure, and sparse intrusive measurements. This problem has not been studied adequately in current reliability applications. This research thus investigates such incomplete covariates problem in reliability analysis. Typical approaches to handling incomplete covariates have been studied to investigate their performance and effects on the reliability analysis results. Since these existing approaches could underestimate the variance in regressions and introduce extra uncertainties to reliability analysis, the developed NNHMs are extended to include handling incomplete covariates as an integral part. The extended versions of NNHMs have been validated using simulated bearing data and real data from a liquefied natural gas pump. The results demonstrate the new approach outperforms the typical incomplete covariates handling approaches. Another problem in reliability analysis is that future covariates of engineering assets are generally unavailable. In existing practices for multi-step reliability analysis, historical covariates were used to estimate the future covariates. Covariates of engineering assets, however, are often subject to substantial fluctuation due to the influence of both engineering degradation and changes in environmental settings. The commonly used covariate extrapolation methods thus would not be suitable because of the error accumulation and uncertainty propagation. To overcome this difficulty, instead of directly extrapolating covariate values, projection of covariate states is conducted in this research. The estimated covariate states and unknown covariate values in future running steps of assets constitute an incomplete covariate set which is then analysed by the extended NNHMs. A new assessment function is also proposed to evaluate risks of underestimated and overestimated reliability analysis results. A case study using field data from a paper and pulp mill has been conducted and it demonstrates that this new multi-step reliability analysis procedure is able to generate more accurate analysis results.

Shrinkage empirical likelihood estimator in longitudinal analysis with time-dependent covariates-application to modeling the health of Filipino children

The method of generalized estimating equations (GEE) is a popular tool for analysing longitudinal (panel) data. Often, the covariates collected are time-dependent in nature, for example, age, relapse status, monthly income. When using GEE to analyse longitudinal data with time-dependent covariates, crucial assumptions about the covariates are necessary for valid inferences to be drawn. When those assumptions do not hold or cannot be verified, Pepe and Anderson (1994, Communications in Statistics, Simulations and Computation 23, 939–951) advocated using an independence working correlation assumption in the GEE model as a robust approach. However, using GEE with the independence correlation assumption may lead to significant efficiency loss (Fitzmaurice, 1995, Biometrics 51, 309–317). In this article, we propose a method that extracts additional information from the estimating equations that are excluded by the independence assumption. The method always includes the estimating equations under the independence assumption and the contribution from the remaining estimating equations is weighted according to the likelihood of each equation being a consistent estimating equation and the information it carries. We apply the method to a longitudinal study of the health of a group of Filipino children.

The impact of pain on the quality of life of people with multiple sclerosis: A community survey

The purpose of this study was to examine the impact of pain on functioning across multiple quality of life (QOL) domains among individuals with multiple sclerosis (MS). A total of 219 people were recruited from a regional MS society membership database to serve as the community-based study sample. All participants completed a questionnaire containing items about their demographic and clinical characteristics, validated measures of QOL and MS-related disability, and a question on whether or not they had experienced clinically significant pain in the preceding 2 weeks. Respondents who reported pain then completed an in-person structured pain interview assessing pain characteristics (intensity, quality, location, extent, and duration). Comparisons between participants with and without MS-related pain demonstrated that pain prevalence and intensity were strongly correlated with QOL: physical health, psychological health, level of independence, and global QOL were more likely to be impaired among people with MS when pain was present, and the extent of impairment was associated with the intensity of pain. Moreover, these relationships remained significant even after statistically controlling for multiple demographic and clinical covariates associated with self-reported QOL. These findings suggest that for people with MS, pain is an important source of distress and disability beyond that caused by neurologic impairments.

The NOURISH randomised control trial : positive feeding practices and food preferences in early childhood - a primary prevention program for childhood obesity

Background Primary prevention of childhood overweight is an international priority. In Australia 20-25% of 2-8 year olds are already overweight. These children are at substantially increased the risk of becoming overweight adults, with attendant increased risk of morbidity and mortality. Early feeding practices determine infant exposure to food (type, amount, frequency) and include responses (eg coercion) to infant feeding behaviour (eg. food refusal). There is correlational evidence linking parenting style and early feeding practices to child eating behaviour and weight status. A focus on early feeding is consistent with the national focus on early childhood as the foundation for life-long health and well being. The NOURISH trial aims to implement and evaluate a community-based intervention to promote early feeding practices that will foster healthy food preferences and intake and preserve the innate capacity to self-regulate food intake in young children. Methods/Design This randomised controlled trial (RCT) aims to recruit 820 first-time mothers and their healthy term infants. A consecutive sample of eligible mothers will be approached postnatally at major maternity hospitals in Brisbane and Adelaide. Initial consent will be for re-contact for full enrolment when the infants are 4-7 months old. Individual mother- infant dyads will be randomised to usual care or the intervention. The intervention will provide anticipatory guidance via two modules of six fortnightly parent education and peer support group sessions, each followed by six months of regular maintenance contact. The modules will commence when the infants are aged 4-7 and 13-16 months to coincide with establishment of solid feeding, and autonomy and independence, respectively. Outcome measures will be assessed at baseline, with follow up at nine and 18 months. These will include infant intake (type and amount of foods), food preferences, feeding behaviour and growth and self-reported maternal feeding practices and parenting practices and efficacy. Covariates will include sociodemographics, infant feeding mode and temperament, maternal weight status and weight concern and child care exposure. Discussion Despite the strong rationale to focus on parents’ early feeding practices as a key determinant of child food preferences, intake and self-regulatory capacity, prospective longitudinal and intervention studies are rare. This trial will be amongst to provide Level II evidence regarding the impact of an intervention (commencing prior to age 12 months) on children’s eating patterns and behaviours. Trial Registration: ACTRN12608000056392

Incomplete pregnancy and risk of ovarian cancer : results from two Australian case-control studies and systematic review

Although full-term pregnancies reduce the risk of ovarian cancer, it has not been conclusively established whether incomplete pregnancies also influence risk. We investigated the relationship between a history of incomplete pregnancy and incident epithelial ovarian cancer among over 4,500 women who participated in two large Australian population-based case-control studies in 1990-1993 and 2002-2005. They provided responses to detailed questions about their reproductive histories and other personal factors. Summary odds ratios (OR) and confidence intervals (CI) derived from each study using the same covariates were aggregated. We found no significant associations between the number of incomplete pregnancies and ovarian cancer, for parous (OR = 0.98, 95% CI: 0.89, 1.08) or nulliparous (OR = 1.06, 95% CI: 0.75, 1.48) women, nor for the number of spontaneous or induced abortions and ovarian cancer for parous women (OR = 0.95, 95% CI 0.82, 1.09; OR = 1.08, 95% CI: 0.86, 1.36) or nulliparous women (OR = 1.2, 95% CI: 0.6, 2.4; OR = 0.8, 95% CI: 0.47, 1.38), respectively. A systematic review of 37 previous studies of the topic confirmed our findings that a history of incomplete pregnancy does not influence a woman’s risk of epithelial ovarian cancer.

A fully Bayesian approach to inference for Coxian phase-type distributions with covariate dependent mean

Phase-type distributions represent the time to absorption for a finite state Markov chain in continuous time, generalising the exponential distribution and providing a flexible and useful modelling tool. We present a new reversible jump Markov chain Monte Carlo scheme for performing a fully Bayesian analysis of the popular Coxian subclass of phase-type models; the convenient Coxian representation involves fewer parameters than a more general phase-type model. The key novelty of our approach is that we model covariate dependence in the mean whilst using the Coxian phase-type model as a very general residual distribution. Such incorporation of covariates into the model has not previously been attempted in the Bayesian literature. A further novelty is that we also propose a reversible jump scheme for investigating structural changes to the model brought about by the introduction of Erlang phases. Our approach addresses more questions of inference than previous Bayesian treatments of this model and is automatic in nature. We analyse an example dataset comprising lengths of hospital stays of a sample of patients collected from two Australian hospitals to produce a model for a patient's expected length of stay which incorporates the effects of several covariates. This leads to interesting conclusions about what contributes to length of hospital stay with implications for hospital planning. We compare our results with an alternative classical analysis of these data.

The explicit hazard model – part 1: theoretical development

Modern Engineering Asset Management (EAM) requires the accurate assessment of current and the prediction of future asset health condition. Appropriate mathematical models that are capable of estimating times to failures and the probability of failures in the future are essential in EAM. In most real-life situations, the lifetime of an engineering asset is influenced and/or indicated by different factors that are termed as covariates. Hazard prediction with covariates is an elemental notion in the reliability theory to estimate the tendency of an engineering asset failing instantaneously beyond the current time assumed that it has already survived up to the current time. A number of statistical covariate-based hazard models have been developed. However, none of them has explicitly incorporated both external and internal covariates into one model. This paper introduces a novel covariate-based hazard model to address this concern. This model is named as Explicit Hazard Model (EHM). Both the semi-parametric and non-parametric forms of this model are presented in the paper. The major purpose of this paper is to illustrate the theoretical development of EHM. Due to page limitation, a case study with the reliability field data is presented in the applications part of this study.

The Explicit Hazard Model – Part 2 : Applications

Hazard and reliability prediction of an engineering asset is one of the significant fields of research in Engineering Asset Health Management (EAHM). In real-life situations where an engineering asset operates under dynamic operational and environmental conditions, the lifetime of an engineering asset can be influenced and/or indicated by different factors that are termed as covariates. The Explicit Hazard Model (EHM) as a covariate-based hazard model is a new approach for hazard prediction which explicitly incorporates both internal and external covariates into one model. EHM is an appropriate model to use in the analysis of lifetime data in presence of both internal and external covariates in the reliability field. This paper presents applications of the methodology which is introduced and illustrated in the theory part of this study. In this paper, the semi-parametric EHM is applied to a case study so as to predict the hazard and reliability of resistance elements on a Resistance Corrosion Sensor Board (RCSB).

Elicitator : an expert elicitation tool for regression in ecology

Expert elicitation is the process of retrieving and quantifying expert knowledge in a particular domain. Such information is of particular value when the empirical data is expensive, limited, or unreliable. This paper describes a new software tool, called Elicitator, which assists in quantifying expert knowledge in a form suitable for use as a prior model in Bayesian regression. Potential environmental domains for applying this elicitation tool include habitat modeling, assessing detectability or eradication, ecological condition assessments, risk analysis, and quantifying inputs to complex models of ecological processes. The tool has been developed to be user-friendly, extensible, and facilitate consistent and repeatable elicitation of expert knowledge across these various domains. We demonstrate its application to elicitation for logistic regression in a geographically based ecological context. The underlying statistical methodology is also novel, utilizing an indirect elicitation approach to target expert knowledge on a case-by-case basis. For several elicitation sites (or cases), experts are asked simply to quantify their estimated ecological response (e.g. probability of presence), and its range of plausible values, after inspecting (habitat) covariates via GIS.

Motivation, development and validation of a new spectral greenness index : a spectral dimension related to foliage projective cover

A method is presented for the development of a regional Landsat-5 Thematic Mapper (TM) and Landsat-7 Enhanced Thematic Mapper plus (ETM+) spectral greenness index, coherent with a six-dimensional index set, based on a single ETM+ spectral image of a reference landscape. The first three indices of the set are determined by a polar transformation of the first three principal components of the reference image and relate to scene brightness, percent foliage projective cover (FPC) and water related features. The remaining three principal components, of diminishing significance with respect to the reference image, complete the set. The reference landscape, a 2200 km2 area containing a mix of cattle pasture, native woodland and forest, is located near Injune in South East Queensland, Australia. The indices developed from the reference image were tested using TM spectral images from 19 regionally dispersed areas in Queensland, representative of dissimilar landscapes containing woody vegetation ranging from tall closed forest to low open woodland. Examples of image transformations and two-dimensional feature space plots are used to demonstrate image interpretations related to the first three indices. Coherent, sensible, interpretations of landscape features in images composed of the first three indices can be made in terms of brightness (red), foliage cover (green) and water (blue). A limited comparison is made with similar existing indices. The proposed greenness index was found to be very strongly related to FPC and insensitive to smoke. A novel Bayesian, bounded space, modelling method, was used to validate the greenness index as a good predictor of FPC. Airborne LiDAR (Light Detection and Ranging) estimates of FPC along transects of the 19 sites provided the training and validation data. Other spectral indices from the set were found to be useful as model covariates that could improve FPC predictions. They act to adjust the greenness/FPC relationship to suit different spectral backgrounds. The inclusion of an external meteorological covariate showed that further improvements to regional-scale predictions of FPC could be gained over those based on spectral indices alone.