Bayesian generalized linear models for meta-analysis of diagnostic tests

With the recognition of the importance of evidence-based medicine, there is an emerging need for methods to systematically synthesize available data. Specifically, methods to provide accurate estimates of test characteristics for diagnostic tests are needed to help physicians make better clinical decisions. To provide more flexible approaches for meta-analysis of diagnostic tests, we developed three Bayesian generalized linear models. Two of these models, a bivariate normal and a binomial model, analyzed pairs of sensitivity and specificity values while incorporating the correlation between these two outcome variables. Noninformative independent uniform priors were used for the variance of sensitivity, specificity and correlation. We also applied an inverse Wishart prior to check the sensitivity of the results. The third model was a multinomial model where the test results were modeled as multinomial random variables. All three models can include specific imaging techniques as covariates in order to compare performance. Vague normal priors were assigned to the coefficients of the covariates. The computations were carried out using the 'Bayesian inference using Gibbs sampling' implementation of Markov chain Monte Carlo techniques. We investigated the properties of the three proposed models through extensive simulation studies. We also applied these models to a previously published meta-analysis dataset on cervical cancer as well as to an unpublished melanoma dataset. In general, our findings show that the point estimates of sensitivity and specificity were consistent among Bayesian and frequentist bivariate normal and binomial models. However, in the simulation studies, the estimates of the correlation coefficient from Bayesian bivariate models are not as good as those obtained from frequentist estimation regardless of which prior distribution was used for the covariance matrix. The Bayesian multinomial model consistently underestimated the sensitivity and specificity regardless of the sample size and correlation coefficient. In conclusion, the Bayesian bivariate binomial model provides the most flexible framework for future applications because of its following strengths: (1) it facilitates direct comparison between different tests; (2) it captures the variability in both sensitivity and specificity simultaneously as well as the intercorrelation between the two; and (3) it can be directly applied to sparse data without ad hoc correction. ^

Assessing the improved discriminatory power of a new biomarker in prognostic models

Although the area under the receiver operating characteristic (AUC) is the most popular measure of the performance of prediction models, it has limitations, especially when it is used to evaluate the added discrimination of a new biomarker in the model. Pencina et al. (2008) proposed two indices, the net reclassification improvement (NRI) and integrated discrimination improvement (IDI), to supplement the improvement in the AUC (IAUC). Their NRI and IDI are based on binary outcomes in case-control settings, which do not involve time-to-event outcome. However, many disease outcomes are time-dependent and the onset time can be censored. Measuring discrimination potential of a prognostic marker without considering time to event can lead to biased estimates. In this dissertation, we have extended the NRI and IDI to survival analysis settings and derived the corresponding sample estimators and asymptotic tests. Simulation studies were conducted to compare the performance of the time-dependent NRI and IDI with Pencina’s NRI and IDI. For illustration, we have applied the proposed method to a breast cancer study.^ Key words: Prognostic model, Discrimination, Time-dependent NRI and IDI ^

Assessment of the effect on statistical power of regression model misspecification by using techniques of mathematical statistics and simulation study

Objectives. This paper seeks to assess the effect on statistical power of regression model misspecification in a variety of situations. ^ Methods and results. The effect of misspecification in regression can be approximated by evaluating the correlation between the correct specification and the misspecification of the outcome variable (Harris 2010).In this paper, three misspecified models (linear, categorical and fractional polynomial) were considered. In the first section, the mathematical method of calculating the correlation between correct and misspecified models with simple mathematical forms was derived and demonstrated. In the second section, data from the National Health and Nutrition Examination Survey (NHANES 2007-2008) were used to examine such correlations. Our study shows that comparing to linear or categorical models, the fractional polynomial models, with the higher correlations, provided a better approximation of the true relationship, which was illustrated by LOESS regression. In the third section, we present the results of simulation studies that demonstrate overall misspecification in regression can produce marked decreases in power with small sample sizes. However, the categorical model had greatest power, ranging from 0.877 to 0.936 depending on sample size and outcome variable used. The power of fractional polynomial model was close to that of linear model, which ranged from 0.69 to 0.83, and appeared to be affected by the increased degrees of freedom of this model.^ Conclusion. Correlations between alternative model specifications can be used to provide a good approximation of the effect on statistical power of misspecification when the sample size is large. When model specifications have known simple mathematical forms, such correlations can be calculated mathematically. Actual public health data from NHANES 2007-2008 were used as examples to demonstrate the situations with unknown or complex correct model specification. Simulation of power for misspecified models confirmed the results based on correlation methods but also illustrated the effect of model degrees of freedom on power.^

Slope estimation in a parametric measurement error model: A simulation study

In regression analysis, covariate measurement error occurs in many applications. The error-prone covariates are often referred to as latent variables. In this proposed study, we extended the study of Chan et al. (2008) on recovering latent slope in a simple regression model to that in a multiple regression model. We presented an approach that applied the Monte Carlo method in the Bayesian framework to the parametric regression model with the measurement error in an explanatory variable. The proposed estimator applied the conditional expectation of latent slope given the observed outcome and surrogate variables in the multiple regression models. A simulation study was presented showing that the method produces estimator that is efficient in the multiple regression model, especially when the measurement error variance of surrogate variable is large.^

Detecting genetic and nutritional lung cancer risk factors related to folate metabolism using Bayesian generalized linear models

Complex diseases, such as cancer, are caused by various genetic and environmental factors, and their interactions. Joint analysis of these factors and their interactions would increase the power to detect risk factors but is statistically. Bayesian generalized linear models using student-t prior distributions on coefficients, is a novel method to simultaneously analyze genetic factors, environmental factors, and interactions. I performed simulation studies using three different disease models and demonstrated that the variable selection performance of Bayesian generalized linear models is comparable to that of Bayesian stochastic search variable selection, an improved method for variable selection when compared to standard methods. I further evaluated the variable selection performance of Bayesian generalized linear models using different numbers of candidate covariates and different sample sizes, and provided a guideline for required sample size to achieve a high power of variable selection using Bayesian generalize linear models, considering different scales of number of candidate covariates. ^ Polymorphisms in folate metabolism genes and nutritional factors have been previously associated with lung cancer risk. In this study, I simultaneously analyzed 115 tag SNPs in folate metabolism genes, 14 nutritional factors, and all possible genetic-nutritional interactions from 1239 lung cancer cases and 1692 controls using Bayesian generalized linear models stratified by never, former, and current smoking status. SNPs in MTRR were significantly associated with lung cancer risk across never, former, and current smokers. In never smokers, three SNPs in TYMS and three gene-nutrient interactions, including an interaction between SHMT1 and vitamin B12, an interaction between MTRR and total fat intake, and an interaction between MTR and alcohol use, were also identified as associated with lung cancer risk. These lung cancer risk factors are worthy of further investigation.^

THE NATURAL AND ORTHOGONAL INTERACTION (NOIA) MODELS FOR QUANTITATIVE TRAITS (QTs) AND COMPLEX DISEASES

My dissertation focuses on developing methods for gene-gene/environment interactions and imprinting effect detections for human complex diseases and quantitative traits. It includes three sections: (1) generalizing the Natural and Orthogonal interaction (NOIA) model for the coding technique originally developed for gene-gene (GxG) interaction and also to reduced models; (2) developing a novel statistical approach that allows for modeling gene-environment (GxE) interactions influencing disease risk, and (3) developing a statistical approach for modeling genetic variants displaying parent-of-origin effects (POEs), such as imprinting. In the past decade, genetic researchers have identified a large number of causal variants for human genetic diseases and traits by single-locus analysis, and interaction has now become a hot topic in the effort to search for the complex network between multiple genes or environmental exposures contributing to the outcome. Epistasis, also known as gene-gene interaction is the departure from additive genetic effects from several genes to a trait, which means that the same alleles of one gene could display different genetic effects under different genetic backgrounds. In this study, we propose to implement the NOIA model for association studies along with interaction for human complex traits and diseases. We compare the performance of the new statistical models we developed and the usual functional model by both simulation study and real data analysis. Both simulation and real data analysis revealed higher power of the NOIA GxG interaction model for detecting both main genetic effects and interaction effects. Through application on a melanoma dataset, we confirmed the previously identified significant regions for melanoma risk at 15q13.1, 16q24.3 and 9p21.3. We also identified potential interactions with these significant regions that contribute to melanoma risk. Based on the NOIA model, we developed a novel statistical approach that allows us to model effects from a genetic factor and binary environmental exposure that are jointly influencing disease risk. Both simulation and real data analyses revealed higher power of the NOIA model for detecting both main genetic effects and interaction effects for both quantitative and binary traits. We also found that estimates of the parameters from logistic regression for binary traits are no longer statistically uncorrelated under the alternative model when there is an association. Applying our novel approach to a lung cancer dataset, we confirmed four SNPs in 5p15 and 15q25 region to be significantly associated with lung cancer risk in Caucasians population: rs2736100, rs402710, rs16969968 and rs8034191. We also validated that rs16969968 and rs8034191 in 15q25 region are significantly interacting with smoking in Caucasian population. Our approach identified the potential interactions of SNP rs2256543 in 6p21 with smoking on contributing to lung cancer risk. Genetic imprinting is the most well-known cause for parent-of-origin effect (POE) whereby a gene is differentially expressed depending on the parental origin of the same alleles. Genetic imprinting affects several human disorders, including diabetes, breast cancer, alcoholism, and obesity. This phenomenon has been shown to be important for normal embryonic development in mammals. Traditional association approaches ignore this important genetic phenomenon. In this study, we propose a NOIA framework for a single locus association study that estimates both main allelic effects and POEs. We develop statistical (Stat-POE) and functional (Func-POE) models, and demonstrate conditions for orthogonality of the Stat-POE model. We conducted simulations for both quantitative and qualitative traits to evaluate the performance of the statistical and functional models with different levels of POEs. Our results showed that the newly proposed Stat-POE model, which ensures orthogonality of variance components if Hardy-Weinberg Equilibrium (HWE) or equal minor and major allele frequencies is satisfied, had greater power for detecting the main allelic additive effect than a Func-POE model, which codes according to allelic substitutions, for both quantitative and qualitative traits. The power for detecting the POE was the same for the Stat-POE and Func-POE models under HWE for quantitative traits.

DISPRO datasets for validation of coastal hydrodynamic models

Appropriate field data are required to check the reliability of hydrodynamic models simulating the dispersion of soluble substances in the marine environment. This study deals with the collection of physical measurements and soluble tracer data intended specifically for this kind of validation. The intensity of currents as well as the complexity of topography and tides around the Cap de La Hague in the center of the English Channel makes it one of the most difficult areas to represent in terms of hydrodynamics and dispersion. Controlled releases of tritium - in the form of HTO - are carried out in this area by the AREVA-NC plant, providing an excellent soluble tracer. A total of 14 493 measurements were acquired to track dispersion in the hours and days following a release. These data, supplementing previously gathered data and physical measurements (bathymetry, water-surface levels, Eulerian and Lagrangian current studies) allow us to test dispersion models from the hour following release to periods of several years which are not accessible with dye experiments. The dispersion characteristics are described and methods are proposed for comparing models against measurements. An application is proposed for a 2 dimensions high-resolution numerical model. It shows how an extensive dataset can be used to build, calibrate and validate several aspects of the model in a highly dynamic and macrotidal area: tidal cycle timing, tidal amplitude, fixed-point current data, hodographs. This study presents results concerning the model's ability to reproduce residual Lagrangian currents, along with a comparison between simulation and high-frequency measurements of tracer dispersion. Physical and tracer data are available from the SISMER database of IFREMER (www.ifremer.fr/sismer/catal). This tool for validation of models in macro-tidal seas is intended to be an open and evolving resource, which could provide a benchmark for dispersion model validation.

Medem Channel profiles based on stand-alone model simulation and corrections with nudging

State-of-the-art process-based models have shown to be applicable to the simulation and prediction of coastal morphodynamics. On annual to decadal temporal scales, these models may show limitations in reproducing complex natural morphological evolution patterns, such as the movement of bars and tidal channels, e.g. the observed decadal migration of the Medem Channel in the Elbe Estuary, German Bight. Here a morphodynamic model is shown to simulate the hydrodynamics and sediment budgets of the domain to some extent, but fails to adequately reproduce the pronounced channel migration, due to the insufficient implementation of bank erosion processes. In order to allow for long-term simulations of the domain, a nudging method has been introduced to update the model-predicted bathymetries with observations. The model-predicted bathymetry is nudged towards true states in annual time steps. Sensitivity analysis of a user-defined correlation length scale, for the definition of the background error covariance matrix during the nudging procedure, suggests that the optimal error correlation length is similar to the grid cell size, here 80-90 m. Additionally, spatially heterogeneous correlation lengths produce more realistic channel depths than do spatially homogeneous correlation lengths. Consecutive application of the nudging method compensates for the (stand-alone) model prediction errors and corrects the channel migration pattern, with a Brier skill score of 0.78. The proposed nudging method in this study serves as an analytical approach to update model predictions towards a predefined 'true' state for the spatiotemporal interpolation of incomplete morphological data in long-term simulations.

Winner-Take-All Contention of Innovation under Globalization: A Simulation Analysis and East Asia's Empirics

This paper sets out to examine how innovation enhances export competitiveness: The proposition that export volume becomes enhanced as more productivity-enhancing innovation is captured by the exporting economy is the focus of this study. From a Schumpeterian perspective, innovation can be characterized by continuous creation and subsequent diffusion of newer technologies on the basis of the exporters' existing capital stock. Then we highlight the theoretical possibility that concentration of innovative activities in a small group of "winner" economies would lead to larger shares of "winner" economies' exports of innovation-active commodities than those commodities for which technology involved is already mature. The world's export data corroborates this theoretical prediction overall, and a focus upon East Asia has revealed the region's increasing resort to technology-intensive commodity sectors, which has presumably been enabled through attracting technology-bearing inward foreign direct investment. Considering the overall gains from innovation, acceleration of full "cycle" of innovation and imitation might be a desirable option.

Financial Policies and Dynamic Game Simulation in Poland and Hungary

Recently, steady economic growth rates have been kept in Poland and Hungary. Money supplies are growing rather rapidly in these economies. In large, exchange rates have trends of depreciation. Then, exports and prices show the steady growth rates. It can be thought that per capita GDPs are in the same level and development stages are similar in these two countries. It is assumed that these two economies have the same export market and export goods are competing in it. If one country has an expansion of monetary policy, price increase and interest rate decrease. Then, exchange rate decrease. Exports and GDP will increase through this phenomenon. At the same time, this expanded monetary policy affects another country through the trade. This mutual relationship between two countries can be expressed by the Nash-equilibrium in the Game theory. In this paper, macro-econometric models of Polish and Hungarian economies are built and the Nash- equilibrium is introduced into them.

Geographical simulation analysis for logistics enhancement in Asia

This paper presents a simulation of the reduction of several components in trade cost for Asia and examines its impact on the economy. Our simulation model based on the new economic geography embraces seven sectors, including manufacturing and non-manufacturing sectors, and 1,715 regions in 18 countries/economies in Asia, in addition to the two economies of the US and the European Union. The geographical course of transactions among regions is modeled as determined based on firms’ modal choice. The model also includes estimates of some border cost measures such as tariff rates, non-tariff barriers, other border clearance costs, transshipment costs and so on. Our simulation analysis for Asia includes several scenarios involving the improvement/development of routes and the reduction of the above-mentioned border cost. We have shown that the contribution of physical and non-physical infrastructure improvements conducted together is larger than the sum of the contribution by each when conducted independently.

Is FTA/EPA effective for a developing country to attract FDI? : simulation analysis based on an extended knowledge-capital model

To prepare an answer to the question of how a developing country can attract FDI, this paper explored the factors and policies that may help bring FDI into a developing country by utilizing an extended version of the knowledge-capital model. With a special focus on the effects of FTAs/EPAs between market countries and developing countries, simulations with the model revealed the following: (1) Although FTA/EPA generally ends to increase FDI to a developing country, the possibility of improving welfare through increased demand for skilled and unskilled labor becomes higher as the size of the country declines; (2) Because the additional implementation of cost-saving policies to reduce firm-type/trade-link specific fixed costs ends to depreciate the price of skilled labor by saving its input, a developing country, which is extremely scarce in skilled labor, is better off avoiding the additional option; (3) If a country hopes to enjoy larger welfare gains with EPA, efforts to increase skilled labor in the country, such as investing in education, may be beneficial.

Behavioral characteristics of applied general equilibrium models with an Armington-Krugman-Melitz encompassing module

This paper explore how simulation results change with different choice of trade specification, and the strength of preference for traded variety by economic agent differs, utilizing two types of three-region, three-sector AGE model that includes the Armington-Krugman-Melitz Encompassing module based on Dixon and Rimmer (2012). Simulation experiments reveal that: (1) the Melitz-type specification does not always enhance effectiveness of a certain policy change more than the one obtained with the Krugman-type, especially when economic agents' preference for traded variety is not so strong; (2) there are likely to be points where the volumes of effects obtained with the Melitz-type exceed the ones with the Krugman-type; and (3) the preference of the producers, those who are in the sectors that exhibit increasing returns to scale, for traded variety might be the engine of explosive effects as suggested by Fujita, et al. (2000).

Simulation analysis of the EU ELV/RoHS directives based on an applied general equilibrium model with Melitz-type trade specification

This paper explores the potential usefulness of an AGE model with the Melitz-type trade specification to assess economic effects of technical regulations, taking the case of the EU ELV/RoHS directives as an example. Simulation experiments reveal that: (1) raising the fixed exporting cost to make sales in the EU market brings results that exports of the targeted commodities (motor vehicles and parts for ELV and electronic equipment for RoHS) to the EU from outside regions/countries expand while the domestic trade in the EU shrinks when the importer's preference for variety (PfV) is not strong; (2) if the PfV is not strong, policy changes that may bring reduction in the number of firms enable survived producers with high productivity to expand production to be large-scale mass producers fully enjoying the fruit of economies of scale; and (3) When the strength of the importer's PfV is changed from zero to unity, there is the value that totally changes simulation results and their interpretations.

Prediction of bendig load capacity of timber beams by finite element method simulation of knots and grain deviation

A finite element model was used to simulate timberbeams with defects and predict their maximum load in bending. Taking into account the elastoplastic constitutive law of timber, the prediction of fracture load gives information about the mechanisms of timber failure, particularly with regard to the influence of knots, and their local graindeviation, on the fracture. A finite element model was constructed using the ANSYS element Plane42 in a plane stress 2D-analysis, which equates thickness to the width of the section to create a mesh which is as uniform as possible. Three sub-models reproduced the bending test according to UNE EN 408: i) timber with holes caused by knots; ii) timber with adherent knots which have structural continuity with the rest of the beam material; iii) timber with knots but with only partial contact between knot and beam which was artificially simulated by means of contact springs between the two materials. The model was validated using ten 45 × 145 × 3000 mm beams of Pinus sylvestris L. which presented knots and graindeviation. The fracture stress data obtained was compared with the results of numerical simulations, resulting in an adjustment error less of than 9.7%