RANDOM EFFECTS MODELS IN A META-ANALYSIS OF THE ACCURACY OF DIAGNOSTIC TESTS WITHIN A GOLD STANDARD IN THE PRESENCE OF MISSING DATA

In evaluating the accuracy of diagnosis tests, it is common to apply two imperfect tests jointly or sequentially to a study population. In a recent meta-analysis of the accuracy of microsatellite instability testing (MSI) and traditional mutation analysis (MUT) in predicting germline mutations of the mismatch repair (MMR) genes, a Bayesian approach (Chen, Watson, and Parmigiani 2005) was proposed to handle missing data resulting from partial testing and the lack of a gold standard. In this paper, we demonstrate an improved estimation of the sensitivities and specificities of MSI and MUT by using a nonlinear mixed model and a Bayesian hierarchical model, both of which account for the heterogeneity across studies through study-specific random effects. The methods can be used to estimate the accuracy of two imperfect diagnostic tests in other meta-analyses when the prevalence of disease, the sensitivities and/or the specificities of diagnostic tests are heterogeneous among studies. Furthermore, simulation studies have demonstrated the importance of carefully selecting appropriate random effects on the estimation of diagnostic accuracy measurements in this scenario.

Evaluation of a Bayesian MCMC Random Effects Inference Methodology for Capture-Mark-Recapture Data

Monte Carlo simulation was used to evaluate properties of a simple Bayesian MCMC analysis of the random effects model for single group Cormack-Jolly-Seber capture-recapture data. The MCMC method is applied to the model via a logit link, so parameters p, S are on a logit scale, where logit(S) is assumed to have, and is generated from, a normal distribution with mean μ and variance σ2 . Marginal prior distributions on logit(p) and μ were independent normal with mean zero and standard deviation 1.75 for logit(p) and 100 for μ ; hence minimally informative. Marginal prior distribution on σ2 was placed on τ2=1/σ2 as a gamma distribution with α=β=0.001 . The study design has 432 points spread over 5 factors: occasions (t) , new releases per occasion (u), p, μ , and σ . At each design point 100 independent trials were completed (hence 43,200 trials in total), each with sample size n=10,000 from the parameter posterior distribution. At 128 of these design points comparisons are made to previously reported results from a method of moments procedure. We looked at properties of point and interval inference on μ , and σ based on the posterior mean, median, and mode and equal-tailed 95% credibility interval. Bayesian inference did very well for the parameter μ , but under the conditions used here, MCMC inference performance for σ was mixed: poor for sparse data (i.e., only 7 occasions) or σ=0 , but good when there were sufficient data and not small σ .

Dilution and redistribution effects of rapid 2-litre infusions of 0.9% (w/v) saline and 5% (w/v) dextrose on haematological parameters and serum biochemistry in normal subjects: a double-blind crossover study

Although hypoalbuminaemia after injury may result from increased vascular permeability, dilution secondary to crystalloid infusions may contribute significantly. In this double-blind crossover study, the effects of bolus infusions of crystalloids on serum albumin, haematocrit, serum and urinary biochemistry and bioelectrical impedance analysis were measured in healthy subjects. Ten male volunteers received 2-litre infusions of 0.9% (w/v) saline or 5% (w/v) dextrose over 1 h; infusions were carried out on separate occasions, in random order. Weight, haemoglobin, serum albumin, serum and urinary biochemistry and bioelectrical impedance were measured pre-infusion and hourly for 6 h. The serum albumin concentration fell in all subjects (20% after saline; 16% after dextrose) by more than could be explained by dilution alone. This fall lasted more than 6 h after saline infusion, but values had returned to baseline 1 h after the end of the dextrose infusion. Changes in haematocrit and haemoglobin were less pronounced (7.5% after saline; 6.5% after dextrose). Whereas all the water from dextrose was excreted by 2 h after completion of the infusion, only one-third of the sodium and water from the saline had been excreted by 6 h, explaining its persistent diluting effect. Impedances rose after dextrose and fell after saline (P<0.001). Subjects voided more urine (means 1663 and 563 ml respectively) of lower osmolality (means 129 and 630 mOsm/kg respectively) and sodium content (means 26 and 95 mmol respectively) after dextrose than after saline (P<0.001). While an excess water load is excreted rapidly, an excess sodium load is excreted very slowly, even in normal subjects, and causes persistent dilution of haematocrit and serum albumin. The greater than expected change in serum albumin concentration when compared with that of haemoglobin suggests that, while dilution is responsible for the latter, redistribution also has a role in the former. Changes in bioelectrical impedance may reflect the electrolyte content rather than the volume of the infusate, and may be unreliable for clinical purposes.

Species interactions and random dispersal rather than habitat filtering drive community assembly during early plant succession

Theory on plant succession predicts a temporal increase in the complexity of spatial community structure and of competitive interactions: initially random occurrences of early colonising species shift towards spatially and competitively structured plant associations in later successional stages. Here we use long-term data on early plant succession in a German post mining area to disentangle the importance of random colonisation, habitat filtering, and competition on the temporal and spatial development of plant community structure. We used species co-occurrence analysis and a recently developed method for assessing competitive strength and hierarchies (transitive versus intransitive competitive orders) in multispecies communities. We found that species turnover decreased through time within interaction neighbourhoods, but increased through time outside interaction neighbourhoods. Successional change did not lead to modular community structure. After accounting for species richness effects, the strength of competitive interactions and the proportion of transitive competitive hierarchies increased through time. Although effects of habitat filtering were weak, random colonization and subsequent competitive interactions had strong effects on community structure. Because competitive strength and transitivity were poorly correlated with soil characteristics, there was little evidence for context dependent competitive strength associated with intransitive competitive hierarchies.

Identifying Individual and Group Effects in the Presence of Sorting: A Neighborhood Effects Application

Researchers have long recognized that the non-random sorting of individuals into groups generates correlation between individual and group attributes that is likely to bias naive estimates of both individual and group effects. This paper proposes a non-parametric strategy for identifying these effects in a model that allows for both individual and group unobservables, applying this strategy to the estimation of neighborhood effects on labor market outcomes. The first part of this strategy is guided by a robust feature of the equilibrium in the canonical vertical sorting model of Epple and Platt (1998), that there is a monotonic relationship between neighborhood housing prices and neighborhood quality. This implies that under certain conditions a non- parametric function of neighborhood housing prices serves as a suitable control function for the neighborhood unobservable in the labor market outcome regression. The second part of the proposed strategy uses aggregation to develop suitable instruments for both exogenous and endogenous group attributes. Instrumenting for each individual's observed neighborhood attributes with the average neighborhood attributes of a set of observationally identical individuals eliminates the portion of the variation in neighborhood attributes due to sorting on unobserved individual attributes. The neighborhood effects application is based on confidential microdata from the 1990 Decennial Census for the Boston MSA. The results imply that the direct effects of geographic proximity to jobs, neighborhood poverty rates, and average neighborhood education are substantially larger than the conditional correlations identified using OLS, although the net effect of neighborhood quality on labor market outcomes remains small. These findings are robust across a wide variety of specifications and robustness checks.

Temperature and precipitation effects on wheat yield across a European transect: a crop model ensemble analysis using impact response surfaces

This study explored the utility of the impact response surface (IRS) approach for investigating model ensemble crop yield responses under a large range of changes in climate. IRSs of spring and winter wheat Triticum aestivum yields were constructed from a 26-member ensemble of process-based crop simulation models for sites in Finland, Germany and Spain across a latitudinal transect. The sensitivity of modelled yield to systematic increments of changes in temperature (-2 to +9°C) and precipitation (-50 to +50%) was tested by modifying values of baseline (1981 to 2010) daily weather, with CO2 concentration fixed at 360 ppm. The IRS approach offers an effective method of portraying model behaviour under changing climate as well as advantages for analysing, comparing and presenting results from multi-model ensemble simulations. Though individual model behaviour occasionally departed markedly from the average, ensemble median responses across sites and crop varieties indicated that yields decline with higher temperatures and decreased precipitation and increase with higher precipitation. Across the uncertainty ranges defined for the IRSs, yields were more sensitive to temperature than precipitation changes at the Finnish site while sensitivities were mixed at the German and Spanish sites. Precipitation effects diminished under higher temperature changes. While the bivariate and multi-model characteristics of the analysis impose some limits to interpretation, the IRS approach nonetheless provides additional insights into sensitivities to inter-model and inter-annual variability. Taken together, these sensitivities may help to pinpoint processes such as heat stress, vernalisation or drought effects requiring refinement in future model development.

When “Where” is More Important Than “When”: Birthplace and Birthdate Effects On The Achievement Of Sporting Expertise

In this study, we assessed whether contextual factors related to where or when an athlete is born influence their likelihood of playing professional sport. The birthplace and birth month of all American players in the National Hockey League, National Basketball Association, Major League Baseball, and Professional Golfer's Association, and all Canadian players in the National Hockey League were collected from official websites. Monte Carlo simulations were used to verify if the birthplace of these professional athletes deviated in any systematic way from the official census population distribution, and chi-square analyses were conducted to determine whether the players' birth months were evenly distributed throughout the year. Results showed a birthplace bias towards smaller cities, with professional athletes being over-represented in cities of less than 500,000 and under-represented in cities of 500,000 and over. A birth month/relative age effect (in the form of a distinct bias towards elite athletes being relatively older than their peers) was found for hockey and baseball but not for basketball and golf. Comparative analyses suggested that contextual factors associated with place of birth contribute more influentially to the achievement of an elite level of sport performance than does relative age and that these factors are essentially independent in their influences on expertise development.

Finite mixture regression model with random effects: application to neonatal hospital length of stay

A two-component mixture regression model that allows simultaneously for heterogeneity and dependency among observations is proposed. By specifying random effects explicitly in the linear predictor of the mixture probability and the mixture components, parameter estimation is achieved by maximising the corresponding best linear unbiased prediction type log-likelihood. Approximate residual maximum likelihood estimates are obtained via an EM algorithm in the manner of generalised linear mixed model (GLMM). The method can be extended to a g-component mixture regression model with the component density from the exponential family, leading to the development of the class of finite mixture GLMM. For illustration, the method is applied to analyse neonatal length of stay (LOS). It is shown that identification of pertinent factors that influence hospital LOS can provide important information for health care planning and resource allocation. (C) 2002 Elsevier Science B.V. All rights reserved.

A Mixture model with random-effects components for clustering correlated gene-expression profiles

Motivation: The clustering of gene profiles across some experimental conditions of interest contributes significantly to the elucidation of unknown gene function, the validation of gene discoveries and the interpretation of biological processes. However, this clustering problem is not straightforward as the profiles of the genes are not all independently distributed and the expression levels may have been obtained from an experimental design involving replicated arrays. Ignoring the dependence between the gene profiles and the structure of the replicated data can result in important sources of variability in the experiments being overlooked in the analysis, with the consequent possibility of misleading inferences being made. We propose a random-effects model that provides a unified approach to the clustering of genes with correlated expression levels measured in a wide variety of experimental situations. Our model is an extension of the normal mixture model to account for the correlations between the gene profiles and to enable covariate information to be incorporated into the clustering process. Hence the model is applicable to longitudinal studies with or without replication, for example, time-course experiments by using time as a covariate, and to cross-sectional experiments by using categorical covariates to represent the different experimental classes. Results: We show that our random-effects model can be fitted by maximum likelihood via the EM algorithm for which the E(expectation) and M(maximization) steps can be implemented in closed form. Hence our model can be fitted deterministically without the need for time-consuming Monte Carlo approximations. The effectiveness of our model-based procedure for the clustering of correlated gene profiles is demonstrated on three real datasets, representing typical microarray experimental designs, covering time-course, repeated-measurement and cross-sectional data. In these examples, relevant clusters of the genes are obtained, which are supported by existing gene-function annotation. A synthetic dataset is considered too.

Nutrient deficiencies in greater yam and their effects on leaf nutrient concentrations

This paper describes the first systematic study of nutritional deficiencies of greater yam (Dioscorea alata). Yam plants (cv. 'Mahoa'a') were propagated from tuber discs and grown in nutrient solution, with nutrients supplied following a modified programmed nutrient-addition method. After an establishment period of four weeks, deficiencies of nitrogen (N), phosphorus (P), potassium (K), calcium (Ca), magnesium (Mg), sulfur (S), iron (Fe), boron (B), manganese (Mn), copper (Cu), zinc (Zn), and molybdenum (Mo) were induced by omitting the relevant nutrient from the solution. Foliar symptoms were recorded photographically. Notably, deficiencies of the mobile macronutrients failed to induce senescence of oldest leaves, while vine growth and younger leaves were affected. Leaf blades of the main stem were sampled in sequence and analyzed chemically, providing the distribution of each nutrient from youngest to oldest leaves in both adequately supplied and deficient plants. The nutrient-concentration profiles, together with the visible symptoms, indicated that little remobilization of mobile macronutrients had occurred. For both macro- and micronutrients, young leaves gave the best separation of nutrient concentrations between well-nourished and deficient plants.

When "where'' is more important than "when": Birthplace and birthdate effects on the achievement of sporting expertise

In this study, we assessed whether contextual factors related to where or when an athlete is born influence their likelihood of playing professional sport. The birthplace and birth month of all American players in the National Hockey League, National Basketball Association, Major League Baseball, and Professional Golfer's Association, and all Canadian players in the National Hockey League were collected from official websites. Monte Carlo simulations were used to verify if the birthplace of these professional athletes deviated in any systematic way from the official census population distribution, and chi-square analyses were conducted to determine whether the players' birth months were evenly distributed throughout the year. Results showed a birthplace bias towards smaller cities, with professional athletes being over-represented in cities of less than 500,000 and under-represented in cities of 500,000 and over. A birth month/relative age effect (in the form of a distinct bias towards elite athletes being relatively older than their peers) was found for hockey and baseball but not for basketball and golf. Comparative analyses suggested that contextual factors associated with place of birth contribute more influentially to the achievement of an elite level of sport performance than does relative age and that these factors are essentially independent in their influences on expertise development.

Personal uses and perceived social and economic effects of advertising in Bulgaria and Romania

Purpose - The objective of this paper is to uncover the underlying dimensions of, and examine the similarities and differences in, personal uses of advertising, perceived socio-economic effects of advertising, and consumer beliefs and attitudes toward advertising in Bulgaria and Romania. Moreover, it aims to identify the relative importance of the predictors of attitudes toward advertising in the two countries. Design/methodology/approach - The paper draws upon findings of previous research and theoretical developments by Bauer and Greyser, Sandage and Leckenby, and Pollay and Mittal. The study uses a stratified random sample of 947 face-to-face interviews with adult respondents from major urban areas in Bulgaria (507) and Romania (440). Variables are measured on multi-item scales as a typical application of the reflective indicator model. Findings - Results show that there are significant differences between Romanian and Bulgarian respondents in terms of their attitudes toward advertising. Romanians are more positive about advertising as an institution than the instruments of advertising. Romanians seem to accept the role of advertising in a free market economy, but have less confidence in advertising claims and techniques. Bulgarian respondents seem more sceptical toward advertising in general and are less enthusiastic about embracing the role of advertising as an institution. Moreover, Bulgarians are highly negative towards the instruments advertising uses to convey its messages to consumers. Research limitations/implications - The research findings reflect the views of urban dwellers and may not be generalisable to the wider population of the two countries. Interviewer bias was reduced by eliminating verbal or non-verbal cues to the respondents, and by the use of stratified random sampling. Practical implications - The paper suggests that the regulatory role of codes of advertising practice and industry regulating bodies should be enhanced, and their ability to protect consumers enforced. Marketing campaigns should be more inclusive to involve diverse social groups and reflect generally-accepted social norms. Originality/value - This study reveals that, while general attitudes toward advertising may be similar, attitudes toward the institution and instruments of advertising may differ even in countries with geographic proximity and low cultural distance. © Emerald Group Publishing Limited.

Location and network effects on innovation success:evidence for UK, German and Irish manufacturing plants

Recent developments in the new economic geography and the literature on regional innovation systems have emphasised the potentially important role of networking and the characteristics of firms' local operating environment in shaping their innovative activity. Modeling UK, German and Irish plants' investments in R&D, technology transfer and networking, and their effect on the extent and success of plants' innovation activities, casts some doubt on the importance of both of these relationships. In particular, our analysis provides no support for the contention that firms or plants in the UK, Ireland or Germany with more strongly developed external links (collaborative networks or technology transfer) develop greater innovation intensity. However, although inter-firm links also have no effect on the commercial success of plants' innovation activity, intra-group links are important in terms of achieving commercial success. We also find evidence that R&D, technology transfer and networking inputs are substitutes rather than complements in the innovation process, and that there are systematic sectoral and regional influences in the efficiency with which such inputs are translated into innovation outputs. © 2001 Elsevier Science B.V.

Statnote 9: The one-way analysis of variance (random effects model)

There is an alternative model of the 1-way ANOVA called the 'random effects' model or ‘nested’ design in which the objective is not to test specific effects but to estimate the degree of variation of a particular measurement and to compare different sources of variation that influence the measurement in space and/or time. The most important statistics from a random effects model are the components of variance which estimate the variance associated with each of the sources of variation influencing a measurement. The nested design is particularly useful in preliminary experiments designed to estimate different sources of variation and in the planning of appropriate sampling strategies.