Non-ignorable missing covariate data in survival analysis: a case-study of an International Breast Cancer Study Group trial

Summary.  Non-ignorable missing data, a serious problem in both clinical trials and observational studies, can lead to biased inferences. Quality-of-life measures have become increasingly popular in clinical trials. However, these measures are often incompletely observed, and investigators may suspect that missing quality-of-life data are likely to be non-ignorable. Although several recent references have addressed missing covariates in survival analysis, they all required the assumption that missingness is at random or that all covariates are discrete. We present a method for estimating the parameters in the Cox proportional hazards model when missing covariates may be non-ignorable and continuous or discrete. Our method is useful in reducing the bias and improving efficiency in the presence of missing data. The methodology clearly specifies assumptions about the missing data mechanism and, through sensitivity analysis, helps investigators to understand the potential effect of missing data on study results.     

‘I'm a good mother really!’ gendered parenting roles and responses to the disclosure of incest

The disclosure of incestuous abuse of a child constitutes such a departure from normative expectations of mothering that it necessitates a complete reappraisal and reconstruction of the mother's identity. This article draws on interview material with women in this situation in order to demonstrate how very traditional gendered parenting models were employed in repairing damaged identities. The ways in which such reliance on the protective and selfless aspects of mothering could contribute to the failure to recognise the abuse of their child are also discussed. In conclusion, motherhood as a source of power for women within the family is critically examined. Copyright © 2002 John Wiley & Sons, Ltd.     

A class of count models and a new consistent test for the Poisson distribution

We define a class of count distributions which includes the Poisson as well as many alternative count models. Then the empirical probability generating function is utilized to construct a test for the Poisson distribution, which is consistent against this class of alternatives. The limit distribution of the test statistic is derived in case of a general underlying distribution, and efficiency considerations are addressed. A simulation study indicates that the new test is comparable in performance to more complicated omnibus tests.     

Quadratic estimators of covariance components in a multivariate mixed linear model

It is known that the Henderson Method III (Biometrics 9:226–252, 1953) is of special interest for the mixed linear models because the estimators of the variance components are unaffected by the parameters of the fixed factor (or factors). This article deals with generalizations and minor extensions of the results obtained for the univariate linear models. A MANOVA mixed model is presented in a convenient form and the covariance components estimators are given on finite dimensional linear spaces. The results use both the usual parametric representations and the coordinate-free approach of Kruskal (Ann Math Statist 39:70–75, 1968) and Eaton (Ann Math Statist 41:528–538, 1970). The normal equations are generalized and it is given a necessary and sufficient condition for the existence of quadratic unbiased estimators for covariance components in the considered model.     

Wer das Stabilisieren vergisst: Warum eine schöne Abschlusspräsentation noch kein gutes Ende bedeutet

Many organizational change projects fail — despite a well thought out concept, professional project management and a binding implementation timetable. The reason why the expected success doesn’t materialize often lies with the missing maintenance. Once the project is officially completed and the client and project manager withdraw, the manner in which the organization will take up these changes is seldom systematically dealt with. Difficulties often arise when no discrete stabilizing measures are planned. We have derived four central action areas which will be clarified by way of practical examples. It will also be shown what is important to pay attention to in the stabilization phase so that a lasting success of the organizational change project can be ensured.     

Consumption and Saving Behavior of Older and Younger Baby Boomers in Korea

The 1998 Korean Survey of Family Income and Expenditures was used to examine the overall consumption and saving behavior of Korean baby boomers and compared the differences in consumption and saving behavior between older and younger boomers. The t -test results indicated that the younger boomers allocated a significantly higher percentage of their expenditures on food away from home, household appliances, transportation and communication than did the older boomers, whereas the older boomers spent higher amounts and allocated larger budget shares on their children's education than did the younger boomers. The results of Ordinary Least-Squares (OLS) regression analysis showed that, holding other factors constant, older boomers not only spent significantly more in the total consumption expenditures and education expenditures, but older boomers also saved significantly less than did younger boomers.     

Privatization of refuse removal and labor costs

We examine the labor-cost savings associated with privatization by comparing earnings and employment trends of public and private sector refuse workers. Findings suggest that high union earnings for workers in the public sector are a source of labor-cost savings in the refuse industry. Evidence on job changers does not indicate that earnings for this group of workers are a compensating differential. Metropolitan area employment findings suggest that municipalities are less likely to use union refuse workers in the public sector when a relatively small percentage of area residents belong to a union. The authors thank Jacqueline Agesa, Keith Bender, Maria Crawford, and Richard Perlman for valuable suggestions. Research assistance from Eric Blackburn is greatly appreciated.     

Choice of the primary analysis in longitudinal clinical trials

Missing data, and the bias they can cause, are an almost ever‐present concern in clinical trials. The last observation carried forward (LOCF) approach has been frequently utilized to handle missing data in clinical trials, and is often specified in conjunction with analysis of variance (LOCF ANOVA) for the primary analysis. Considerable advances in statistical methodology, and in our ability to implement these methods, have been made in recent years. Likelihood‐based, mixed‐effects model approaches implemented under the missing at random (MAR) framework are now easy to implement, and are commonly used to analyse clinical trial data. Furthermore, such approaches are more robust to the biases from missing data, and provide better control of Type I and Type II errors than LOCF ANOVA. Empirical research and analytic proof have demonstrated that the behaviour of LOCF is uncertain, and in many situations it has not been conservative. Using LOCF as a composite measure of safety, tolerability and efficacy can lead to erroneous conclusions regarding the effectiveness of a drug. This approach also violates the fundamental basis of statistics as it involves testing an outcome that is not a physical parameter of the population, but rather a quantity that can be influenced by investigator behaviour, trial design, etc. Practice should shift away from using LOCF ANOVA as the primary analysis and focus on likelihood‐based, mixed‐effects model approaches developed under the MAR framework, with missing not at random methods used to assess robustness of the primary analysis. Copyright © 2004 John Wiley & Sons, Ltd.