Statistical Insights Research Articles

Markov transition models for binary repeated measures with ignorable and nonignorable missing values

Motivated by problems encountered in studying treatments for drug dependence, where repeated binary outcomes arise from monitoring biomarkers for recent drug use, this article discusses a statistical strategy using Markov transition model for analyzing incomplete binary longitudinal data. When the mechanism giving rise to missing data can be assumed to be ;ignorable', standard Markov transition models can be applied to observed data to draw likelihood-based inference on transition probabilities between outcome events. Illustration of this approach is provided using binary results from urine drug screening in a clinical trial of baclofen for cocaine dependence. When longitudinal data have ;nonignorable' missingness mechanisms, random-effects Markov transition models can be used to model the joint distribution of the binary data matrix and the matrix of missingness indicators. Categorizing missingness patterns into those for occasional or ;intermittent' missingness and those for monotonic missingness or ;missingness due to dropout', the random-effects Markov transition model was applied to a data set containing repeated breath samples analyzed for expired carbon monoxide levels among opioid-dependent, methadone-maintained cigarette smokers in a smoking cessation trial. Markov transition models provide a novel reconceptualization of treatment outcomes, offering both intuitive statistical values and relevant clinical insights.

The OECD FIRE database

Abstract Realistic modelling of fire scenarios is still difficult due to the scarcity of reliable data needed for deterministic and probabilistic fire safety analysis. Therefore, it has been recognized as highly important to establish a fire event database on an international level. In consequence, several member countries of the Nuclear Energy Agency of the OECD have decided in 2000 to establish the International Fire Data Exchange Project (OECD FIRE) to encourage multilateral co-operation in the collection and analysis of data related to fire events at nuclear power plants. This paper presents the OECD FIRE project objectives, work scope and current status of the OECD FIRE database after 3 years of operation as well as first preliminary statistical insights gained from the collected data.

Evolutionary Efficiency and Happiness

We model happiness as a measurement tool used to rank alternative actions. Evolution favors a happiness function that measures the individual’s success in relative terms. The optimal function is based on a time‐varying reference point—or performance benchmark—that is updated over time in a statistically optimal way in order to match the individual’s potential. Habits and peer comparisons arise as special cases of such an updating process. This updating also results in a volatile level of happiness that continuously reverts to its long‐term mean. Throughout, we draw a parallel with a problem of optimal incentives, which allows us to apply statistical insights from agency theory to the study of happiness.

How (Not) to Lie with Spatial Statistics

How (Not) to Lie with Spatial Statistics

Geometric Representation of High Dimension, Low Sample Size Data

SummaryHigh dimension, low sample size data are emerging in various areas of science. We find a common structure underlying many such data sets by using a non-standard type of asymptotics: the dimension tends to ∞ while the sample size is fixed. Our analysis shows a tendency for the data to lie deterministically at the vertices of a regular simplex. Essentially all the randomness in the data appears only as a random rotation of this simplex. This geometric representation is used to obtain several new statistical insights.

Novel target genes of the Wnt pathway and statistical insights into Wnt target promoter regulation

The Wnt pathway controls biological processes via the regulation of target gene expression. The expression of direct Wnt target genes, e.g. cyclin D1 and MYC, is activated by the transcription factor TCF, which binds to specific sequence motifs in the promoter. Indirect target genes are regulated via transcription regulators, which are targets of the Wnt pathway. As an example, MYC regulates the MYC interacting zinc finger protein-1 (MIZ-1), which is able to inhibit the expression of the indirect target p21WAF1. We intended to identify new Wnt target genes and to get a deeper insight into the regulatory mechanisms of Wnt target gene expression. For this we analyzed the differential expression pattern of Wnt-1 activated cells by microarray analysis. We identified 43 sequences including eight expressed sequence tags (ESTs), which showed increased transcript levels, and 104 sequences including 19 ESTs with decreased RNA levels. Northern blot and real-time quantitative PCR analysis of the differential expression levels of 15 genes confirmed the differential expression trends of eight candidate genes. When the Wnt pathway was regulated at the lower level of glycogen synthase kinase-3 beta (GSK-3 beta) or adenomatous polyposis coli (APC), we detected discrepant expression trends. We compared the number of binding sites of transcription factors in the genomic regions of all candidate target genes with the number of sites in control genes. We found that the genomic regions of the down-regulated genes include an increased number of putative MIZ-1 binding sites. Our study introduces several new Wnt target genes and provides indications that the specific gene expression pattern depends on the type of the activation trigger or the level of interference with the Wnt pathway. Furthermore, our data indicate that a high proportion of Wnt target genes are regulated by indirect mechanisms.

Historical Perspectives

This month’s Historical Perspectives is in a slightly different format, but should be of considerable interest to readers of NeoReviews. Professor Nicholas Wald is an epidemiologist who has evaluated and written about a variety of topics, but he is best known for his work on screening methods, particularly in the area of maternal serum alpha-fetoprotein (AFP) and fetal abnormalities. Because of his continuing heavy commitments and to minimize the disruption of his schedule, I posed a number of questions to him, to which he responded with detailed answers. Before moving to his responses, some background information seems in order. Shortly before publication of the article highlighted this month, evaluation of amniotic fluid in cases of open spina bifida and other neural tube defects had shown increased levels of AFP. (1) Somewhat later, it was shown that abdominal wall defects, as well as neural tube defects, could contribute to increased levels of amniotic fluid AFP. (2) In his responses, Professor Wald documents the sequence of events that followed, noting that Professor David Brock in Edinburgh described the association of increased levels of maternal serum alpha-fetoprotein (MSAFP) with fetal anencephaly, which triggered the idea that this also might be true of open spina bifida. Thus, MSAFP was used as a screening tool for the detection of neural tube and abdominal wall defects. Elevated MSAFP levels were noted in other “at-risk” pregnancies (3) and twin pregnancies. (4) To provide better standardization, Professor Wald introduced the concept of “multiples of the median” (MoM) at different gestational ages. Approximately a decade passed before the association was made between low MoMs for AFP and chromosomal defects, particularly Down syndrome. Soon thereafter, several other serum markers for Down syndrome were described, and by linking them to maternal age, screening became increasingly valuable. In the early 1980s, Dr …

A multiplicative statistical model predicts the size distribution of unruptured intracranial aneurysms

A statistical model for characterizing the erratic nature of aneurysm evolution is developed and tested. This model is based upon a multiplicative hypothesis, whereby it is theorized that the progressive changes in the size of a given aneurysm are determined by random multipliers. Such a model would predict that within a large population of aneurysms, a lognormal histogram for aneurysm sizes would occur (J.e. the logarithms of aneurysm size would have a normal distribution). When applied to previously published clinical data of unruptured aneurysms by Crompton (7966) and McCormick et al. (7970), the model is found to adequately describe both sets of data. The methods introduced in this paper illustrate the utility of incorporating statistical and clinical insights with fundamental biometry for studying the complex phenomena of aneurysm growth and rupture. [Neural Res 1998; 20: 421–426]

STATISTICAL ISSUES IN DISINFESTATION RESEARCH

Food and Agriculture Organization international initiatives are moving the focus from quarantine procedures based around the probit 9 concept to a broader examination of the risk that pests will survive through to establishment in the importing country. Thorough critical evaluation is needed to ensure that this new focus provides an improved scientific basis for disinfestation research. Irrespective of changes which may occur in the regulatory environment, experimentation should be multi-factor, with produce and insects examined in parallel, in order to find optimum treatment conditions quickly and reliably. Analyses should accurately model the experimental design, the insect response and other aspects of the underlying biology. Sampling error calculations must take account of failure to detect pests on infested sampled fruit. There are large potential returns from effective use of statistical insights and from the use of modern statistical software tools.

Time of outage restoration analysis in distribution systems

This paper presents an approach that uses statistical techniques to analyze and interpret time of outage restoration (TOR) data in distribution systems with respect to several selected factors. The factors an classified into three main categories: (1) time; (2) consequence; and (3) environment. Each category is further broken down into groups and classes for more detailed analysis. Both statistical, quantitative results and field engineering insights are correlated and discussed. The approach taken in this paper can be easily modified for other utilities' TOR analysis. These analyses can provide a better understanding as to how the restoration time is affected by different influential factors. The methodology and general discussion in the paper will be useful for future research and development of the TOR issues in distribution systems, such as model development for customer service improvement, distribution operations, and planning.

A LONGITUDINAL STUDY OF THE COMPETITIVE POSITIONS AND ENTRY PATHS OF EUROPEAN FIRMS IN THE U.S. PHARMACEUTICAL MARKET

This paper combines the statistical insights of dynamic strategic group analysis with the qualitative richness of historical analysis to explore the modes of entry, expansion paths, and competitive postures of European firms in the U.S. pharmaceutical market. Patterns of entry and market development over a 20-year period are analyzed. The roles of strategic assets and competencies in determining both the entry strategy and the final competitive posture of these firms in the U.S. market are discussed.

A Simulation Study of Fishway Design: an Example of Simulation in Environmental Problem Solving

The paper argues that modelling provides a convenient and cost-effective means of characterizing and predicting the behaviour of environmental systems disturbed by man. A series of modelling criteria are suggested that aim at compelling modelling to work with current data, explicitly represent the current understanding of a system, include considerations of variability, and address predictive uncertainty. Model which do so provide a rigorous framework for the comparison of alternative courses of action. As an example of the methodology, the problems associated with fishway design, highlighting the required statistical analysis and insights, are discussed. Comparison of the modelling results to those gained from operational experience with real fishways were used to validate model results and ensure model credibility. The models were then used to examine the effect of design on interactions between pools in fishway as a means of selecting the designs which minimized migration stress.

Statistical Methods and Insights for Protein and DNA Sequences

Note: Department of Mathematics, Stanford University, California 94305. Reference GR-BUCHER-ARTICLE-1991-002 Record created on 2007-12-17, modified on 2017-05-12

Statistical insights. IV. Variation as a key to research design.

Statistical insights. IV. Variation as a key to research design.

Currency and the Economy in Mid-Tudor England

T tHE study of English economic history in the early modern period has undergone a profound change during the last few decades for, as Prof. Fisher has remarked, statistical data have come increasingly to replace literary sources as the stuff upon which our interpretations of the past are based. Through statistical analysis insights into important problems, such as the geographical distribution of wealth, have become possible for the first time; conflicting and misleading figures-nowhere, perhaps, more apparent than in the broad field of government revenue-have been corrected; old indexes have yielded place to new; and almost everywhere it has become fashionable to underpin arguments with statistics of some kind. If the numerate historian has not yet provided a panacea-partly because there is seldom agreement as to what his statistics actually mean, and partly because he has not yet counted long enough or in sufficient fields-it still seems possible that in some areas and within carefully defined chronological limits sufficient has already been done to make substantial reinterpretation feasible. The appearance of Prof. Gould's book' marks this point in respect of currency and the economy in mid-Tudor England and it is some measure of the achievements of earlier generations of economic historians that he can attempt his task in under 200 pages. Broadly speaking, this study falls into two parts: the second, dealing with the exchange rate and the export trade, stemming from and being partly dependent upon the first which is devoted to an examination of the determinants of mint supply, the output of the mints, and the size of the circulating medium. In both parts, but particularly in the second, Prof. Gould is able to base himself firmly on earlier statistical compilations such as those of Prof. Carus-Wilson and Miss Coleman on exports, G. C. Brooke and Miss Stokes on coin output, and Dr van der Wee on exchange rates.2 Even in these areas, however, in the last resort his conclusions are the outcome of his own calculations, and in addition he presents a body of information-on the size of the circulating medium during the debasement, and the volume of cloth and wool exports, I 544/5-I 560/I -which he himself has derived first hand from manuscript sources. While Prof. Gould's book is