A Non-parametric Frailty Model for Temporally Clustered Multivariate Failure Times

Abstract A model is introduced here for multivariate failure time data arising from heterogenous populations. In particular, we consider a situation in which the failure times of individual subjects are often temporally clustered, so that many failures occur during a relatively short age interval. The clustering is modelled by assuming that the subjects can be divided into ‘internally homogenous’ latent classes, each such class being then described by a time‐dependent frailty profile function. As an example, we reanalysed the dental caries data presented earlier in Härkänen et al. [Scand. J. Statist. 27 (2000) 577], as it turned out that our earlier model could not adequately describe the observed clustering.     

Stochastic complexity and the mdl principle

A Search for the stochastic complexity of the observed data, as the greates lower bound with which the data can be encoded, represents a global maximum likelihood principle, which permits comparison of models regardless of the number of parameters in them. For important special classes, such as the guassian and the multinomial models, formulas for the stochastic complexity give new and powerful model selection criteria,while in the general case approximations can be computed with the MDL principle. Once a model is found with which the stochastic complexity is reached, there is nothing further to learn from the data with the proposed models. The basic notions are reviewed and numerical examples are given.     

Cash versus Child Care Services in Finland

In 1985 the Finnish Parliament passed a law which stipulated that all children under age three were to be guaranteed a day care place as from the beginning of the 1990s. The law was made possible by a political compromise in which the agrarian Centre Party won the backing it needed to push through a system of state subsidies for the home care of children. As an alternative to a day care place, parents were now given the option of taking a child home care allowance and using that allowance either for purposes of looking after their child themselves or for paying for a private place. Measured in terms of the number of users, the child home care allowance was a hugely successful innovation. Most parents of small children have used the allowance for at least some period of time. This was due above all to the size of the allowance compared with other social benefits. However, following cutbacks in allowance expenditure of more than 20 per cent from 1995, the use of home care allowances declined at almost the same rate as the allowances were reduced. This brought significant short-term savings to the Government and to local authorities, but in the longer term other costs have been rising. There has even been a sharp, unexpected decline in the birth rate. The case of Finland goes to show that, even in a country where wage-earning motherhood has become firmly established, income transfers through family policy can have a very significant influence on the numbers opting temporarily for homemaking.     

Interaction between employees' physical load factors and early support at the workplace

Physical workload is a continuous problem, even in modern workplaces. The purpose of the survey was to determine the effect of support on employees' physical load factors at workplaces. Training, guidance and support were the main focus areas of the early support intervention, which aimed to enable supervisors to find weak signals of impaired ergonomics.The survey was carried out in the form of a controlled longitudinal study, and the material was gathered via a questionnaire in both 2008 and 2010 from two co-operative trade groups. The final sample was 301 intervention subjects and 235 control subjects, and the response rate was 45% in both groups. We applied factor analysis to reduce the number of items. The physical load factors' sum score consisted of six items. We used logistic regression in the statistical analysis.Encouragement to improve processes at the workplace increased the probability of positive change (i.e. decrease) in physical load factors. The same applied to working pace, if individuals could control it themselves. In contrast, workload and the support of supervisors had a reversed impact on workers' physical load factors. Focusing on promoting workers' ergonomics is still important in workplaces when aiming to decrease physical load factors.     

Early Warning Indicators for Challenges in Underground Coal Storage

Early warning or leading indicators are discussed for unexpected incidences in case of large‐scale underground coal storage at a power plant. The experience is compared with above‐ground stockpiles for which established procedures are available but where access for prevention and mitigation are much easier. It is suggested that while the explicit organization, procedures, and the general safety systems aim to provide the targeted levels of performance for the storage, representing new technology without much precedence elsewhere in the world, the extensive experience and tacit knowledge from above‐ground open and closed storage systems can help to prepare for and to prevent unwanted incidents in the underground storage. This kind of experience has been also found useful for developing the leading or early warning indicators for underground storage. Examples are given on observed autoignition and freezing of coal in the storage silos, and on occupational hazards. Selection of the leading indicators needs to consider the specific features of the unique underground facility.     

Caries on Permanent Teeth: A Non-parametric Bayesian Analysis

Most earlier epidemiological investigations of dental caries have been based on cross-sectional data. Subject-specific information of dental caries in the past, and the duration of exposure of each tooth to the oral environment, are obviously important factors also influencing the presence of dental caries in the future. This has led us to consider multivariate survival models in which the information about the tooth eruption and failure times are combined to assess caries risk. A non-parametric Bayesian intensity model is presented, reflecting, on the one hand, the within subject and between subject sources of variability, and a corresponding split of variability when considering the 28 permanent teeth. We analyse a data set consisting of the dental history of 240 boys, where the observations are based on predetermined dental examinations taking place approximately once every year. Markov chain Monte Carlo integration techniques are applied in the numerical work.     

Effects of early support intervention on workplace ergonomics--a two-year followup study

The purpose of the controlled longitudinal study was to determine the effect of a tailored early support intervention method on workers' workplace ergonomics. The main areas of the early support intervention were training, guidance and support for supervisors in finding weak signals of impaired ergonomics. Supervisors were also trained to bring up these weak signals in discussion with employees and to make necessary changes at the workplace. The data consisted of 301 intervention subjects and 235 control subjects working in the field of commerce. The questionnaires were carried out in 2008 and in 2010, and the response rates among both groups were 45%. We used multivariate repeated measures analysis of variance (MANOVA) to test the difference in the groups at two points of time. The main result was that in the areas of work environment, the interaction between group and time was statistically significant (p=0.0004). The work environment improved in the intervention group, but deteriorated in the control. Working methods improved due to the interventions, but physical load factors increased over time in both groups. According to the study, tailored early support intervention has a generally beneficial impact on workers' workplace ergonomics in the areas of work methods, work environment and accident factors.     

Target Levels—Tools for Prevention

Although occupational exposure limits are sought to establish health-based standards, they do not always give a sufficient basis for planning an indoor air climate that is good and comfortable for the occupants in industrial work rooms. This paper considers methodologies by which the desired level, i.e., target level, of air quality in industrial settings can be defined, taking into account feasibility issues. Risk assessment based on health criteria is compared with risk-assessment based on "Best Available Technology" (BAT). Because health-based risk estimates at low concentration regions are rather inaccurate, the technology-based approach is emphasized. The technological approach is based on information on the prevailing concentrations in industrial work environments and the benchmark air quality attained with the best achievable technology. The prevailing contaminant concentrations are obtained from a contaminant exposure databank, and the benchmark air quality by field measurements in industrial work rooms equipped with advanced ventilation and production technology. As an example, the target level assessment has been applied to formaldehyde, total inorganic dust and hexavalent chromium, which are common contaminants in work room air.