Statistics and Probability Commons^™

Longitudinal Measurement And Hierarchical Classification Framework For The Prediction Of Alzheimer's Disease, Meiyan Huang, Wei Yang, Qianjin Feng, Wufan Chen, Michael Weiner, Paul Aisen, Ronald Petersen, Clifford R. Jack Jr., William Jagust, John Trojanowki, Arthur W. Toga, Laurel Beckett, Robert C. Green, Andrew Saykin, John Morris, Leslie M. Shaw, Jeffrey Kaye, Joseph Quinn, Lisa Silbert, Betty Lind, Raina Carter, Sara Dolen, Lon S. Schneider, Sonia Pawluczyk, Mauricio Beccera, Liberty Teodoro, Bryan Spann, James Brewer, Helen Vanderswag, Adam Fleisher, Charles D. Smith, Greg A. Jicha, Peter A. Hardy, Partha Sinha, Elizabeth Oates, Gary Conrad

Neurology Faculty Publications

Accurate prediction of Alzheimer’s disease (AD) is important for the early diagnosis and treatment of this condition. Mild cognitive impairment (MCI) is an early stage of AD. Therefore, patients with MCI who are at high risk of fully developing AD should be identified to accurately predict AD. However, the relationship between brain images and AD is difficult to construct because of the complex characteristics of neuroimaging data. To address this problem, we present a longitudinal measurement of MCI brain images and a hierarchical classification method for AD prediction. Longitudinal images obtained from individuals with MCI were investigated to acquire important …

Quantifying The Totality Of Treatment Effect With Multiple Event-Time Observations In The Presence Of A Terminal Event From A Comparative Clinical Study, Brian Claggett, Lu Tian, Haoda Fu, Scott D. Solomon, L. J. Wei

Harvard University Biostatistics Working Paper Series

To evaluate the totality of one treatment's benefit/risk profile relative to an alternative treatment via a longitudinal comparative clinical study, the timing and occurrence of multiple clinical events are typically collected during the patient's followup. These multiple observations reflect the patient's disease progression/burden over time. The standard practice is to create a composite endpoint from the multiple outcomes, the timing of the occurrence of the first clinical event, to evaluate the treatment via the standard survival analysis techniques. By ignoring all events after the composite outcome, this type of assessment may not be ideal. Various parametric or semi-parametric procedures have …

Real Data Is Messy... And Manageable, Beverly Wood, Carl Clark

Publications

Using real data in an introductory statistics course is a delicate balance between reality and manageability. The internet is awash with data that is useful for students to answer questions of interest to them but it is not always formatted as neatly as textbook data. The ASA's recently endorsed GAISE College Report 2016 points to the plausibility of considering multivariable thinking even if only at a rudimentary level. With both messy and multivariable data in mind, we present some activities/projects and sources for data to give introductory students the opportunity to engage with real data.

Quantifying Similarity In Reliability Surfaces Using The Probability Of Agreement, Nathaniel Stevens, C. M. Anderson-Cook

Mathematics

When separate populations exhibit similar reliability as a function of multiple explanatory variables, combining them into a single population is tempting. This can simplify future predictions and reduce uncertainty associated with estimation. However, combining these populations may introduce bias if the underlying relationships are in fact different. The probability of agreement formally and intuitively quantifies the similarity of estimated reliability surfaces across a two-factor input space. An example from the reliability literature demonstrates the utility of the approach when deciding whether to combine two populations or to keep them as distinct. New graphical summaries provide strategies for visualizing the results.

Tuning Parameter Selection In Cox Proportional Hazards Model With A Diverging Number Of Parameters, Andy Ni, Jianwen Cai

Memorial Sloan-Kettering Cancer Center, Dept. of Epidemiology & Biostatistics Working Paper Series

Regularized variable selection is a powerful tool for identifying the true regression model from a large number of candidates by applying penalties to the objective functions. The penalty functions typically involve a tuning parameter that control the complexity of the selected model. The ability of the regularized variable selection methods to identify the true model critically depends on the correct choice of the tuning parameter. In this study we develop a consistent tuning parameter selection method for regularized Cox's proportional hazards model with a diverging number of parameters. The tuning parameter is selected by minimizing the generalized information criterion. We …

Variance Prior Specification For A Basket Trial Design Using Bayesian Hierarchical Modeling, Kristen Cunanan, Alexia Iasonos, Ronglai Shen, Mithat Gonen

Memorial Sloan-Kettering Cancer Center, Dept. of Epidemiology & Biostatistics Working Paper Series

Background: In the era of targeted therapies, clinical trials in oncology are rapidly evolving, wherein patients from multiple diseases are now enrolled and treated according to their genomic mutation(s). In such trials, known as basket trials, the different disease cohorts form the different baskets for inference. Several approaches have been proposed in the literature to efficiently use information from all baskets while simultaneously screening to find individual baskets where the drug works. Most proposed methods are developed in a Bayesian paradigm that requires specifying a prior distribution for a variance parameter, which controls the degree to which information is shared …

Optimized Variable Selection Via Repeated Data Splitting, Marinela Capanu, Colin B. Begg, Mithat Gonen

Memorial Sloan-Kettering Cancer Center, Dept. of Epidemiology & Biostatistics Working Paper Series

We introduce a new variable selection procedure that repeatedly splits the data into two sets, one for estimation and one for validation, to obtain an empirically optimized threshold which is then used to screen for variables to include in the final model. Simulation results show that the proposed variable selection technique enjoys superior performance compared to candidate methods, being amongst those with the lowest inclusion of noisy predictors while having the highest power to detect the correct model and being unaffected by correlations among the predictors. We illustrate the methods by applying them to a cohort of patients undergoing hepatectomy …

Mediation Analysis For Censored Survival Data Under An Accelerated Failure Time Model, Isabel Fulcher, Eric J. Tchetgen Tchetgen, Paige Williams

Harvard University Biostatistics Working Paper Series

Recent advances in causal mediation analysis have formalized conditions for estimating direct and indirect effects in various contexts. These approaches have been extended to a number of models for survival outcomes including accelerated failure time (AFT) models which are widely used in a broad range of health applications given their intuitive interpretation. In this setting, it has been suggested that under standard assumptions, the “difference” and “product” methods produce equivalent estimates of the indirect effect of exposure on the survival outcome. We formally show that these two methods may produce substantially different estimates in the presence of censoring or truncation, …

Studying The Optimal Scheduling For Controlling Prostate Cancer Under Intermittent Androgen Suppression, Sunil K. Dhar, Hans R. Chaudhry, Bruce G. Bukiet, Zhiming Ji, Nan Gao, Thomas W. Findley

Harvard University Biostatistics Working Paper Series

This retrospective study shows that the majority of patients’ correlations between PSA and Testosterone during the on-treatment period is at least 0.90. Model-based duration calculations to control PSA levels during off-treatment are provided. There are two pairs of models. In one pair, the Generalized Linear Model and Mixed Model are both used to analyze the variability of PSA at the individual patient level by using the variable “Patient ID” as a repeated measure. In the second pair, Patient ID is not used as a repeated measure but additional baseline variables are included to analyze the variability of PSA.

The Pareto-G Extended Weibull Distribution, Oluwaseun Elizabeth Otunuga

Theses, Dissertations and Capstones

In this thesis, the Pareto family of extended Weibull distribution is introduced and discussed extensively. This family consists of the Pareto (Type I) Extended Weibull Distribution or PEW for short, and the Pareto (Type II) Extended Weibull Distribution or otherwise called the Lomax Extended Weibull Distribution or LEW for short. The numbers of the parameters of PEW or LEW depend on the number of parameters for the extended Weibull distribution and type of the Pareto distribution. Some properties of these distributions, such as the hazard rate function, the survival function, moments, skewness, kurtosis, mean deviation and entropies are discussed. The …

Gender Differences In Hiv Care Among Criminal Justice-Involved Persons: Baseline Data From The Care+ Corrections Study., Curt Beckwith, Breana Uhrig Castonguay, Claudia Trezza, Lauri Bazerman, Rudy Patrick, Alice Cates, Halli Olsen, Ann Kurth, Tao Liu, James Peterson, Irene Kuo

Epidemiology Faculty Publications

BACKGROUND: HIV-infected individuals recently released from incarceration have suboptimal linkage and engagement in community HIV care. We conducted a study to evaluate an information and communication technology intervention to increase linkage to community care among HIV-infected persons recently involved in the criminal justice (CJ) system. Baseline characteristics including risk behaviors and HIV care indicators are reported and stratified by gender.

METHODS: We recruited HIV-infected individuals in the District of Columbia jail and persons with a recent history of incarceration through community and street outreach. Participants completed a baseline computer-assisted personal interview regarding HIV care and antiretroviral treatment (ART) adherence, substance …

Use Of National Standards To Monitor Hiv Care And Treatment In A High Prevalence City-Washington, Dc., Amanda D Castel, Arpi Terzian, Rachel Hart, Nabil Rayeed, Mariah M Kalmin, Heather Young, Alan E Greenberg

Epidemiology Faculty Publications

We sought to benchmark the quality of HIV care being received by persons living with HIV in care in Washington, DC and identify individual-level and structural-level differences. Data from the DC Cohort, an observational HIV cohort of persons receiving outpatient care in DC, were used to estimate the Institute of Medicine (IOM) and Department of Health and Human Services (HHS) quality of care measures. Differences in care by demographics and clinic type were assessed using χ2 tests and multivariable regression models. Among 8,047 participants, by HHS standards, 69% of participants were retained in care (RIC), 95% were prescribed antiretroviral therapy …

"It Makes You Feel Like Someone Cares" Acceptability Of A Financial Incentive Intervention For Hiv Viral Suppression In The Hptn 065 (Tlc-Plus) Study., Elizabeth Greene, Allison Pack, Jill Stanton, Victoria Shelus, Elizabeth E Tolley, Jamilah Taylor, Wafaa M El Sadr, Bernard M Branson, Jason Leider, Natella Rakhmanina, Theresa Gamble

Pediatrics Faculty Publications

BACKGROUND: HPTN 065 (TLC-Plus) evaluated the feasibility and effectiveness of providing quarterly $70 gift card financial incentives to HIV-infected patients on antiretroviral therapy (ART) to encourage ART adherence and viral suppression, and represents the largest study to-date of a financial incentive intervention for HIV viral suppression. A post-trial qualitative substudy was undertaken to examine acceptability of the financial incentives among those receiving and implementing the intervention.

METHODS: Between July and October 2013, semi-structured interviews were conducted with 72 patients and 12 investigators from 14 sites; three focus groups were conducted with 12 staff from 10 sites. Qualitative data collection elicited …

Binary Classification On Past Due Of Service Accounts Using Logistic Regression And Decision Tree, Yan Wang, Jennifer L. Priestley

Published and Grey Literature from PhD Candidates

This paper aims at predicting businesses’ past due in service accounts as well as determining the variables that impact the likelihood of repayment. Two binary classification approaches, logistic regression and the decision tree, were conducted and compared. Both approaches have very good performances with respect to the accuracy. However, the decision tree only uses 10 predictors and reaches an accuracy of 96.69% on the validation set while logistic regression includes 14 predictors and reaches an accuracy of 94.58%. Due to the large concern of false negatives in financial industry, the decision tree technique is a better option than logistic regression …

Conventional Early Infant Diagnosis In Lesotho From Specimen Collection To Results Usage To Manage Patients: Where Are The Bottlenecks?, Appolinaire Tiam, Michelle M Gill, Heather J Hoffman, Anthony Isavwa, Mafusi Mokone, Matokelo Foso, Jeffrey T Safrit, Lynne M Mofenson, Thorkild Tylleskär, Laura Guay

Epidemiology Faculty Publications

INTRODUCTION: Early infant diagnosis is an important step in identifying children infected with HIV during the perinatal period or in utero. Multiple factors contribute to delayed antiretroviral treatment initiation for HIV-infected children, including delays in the early infant HIV diagnosis cascade.

METHODS: We conducted a retrospective study to evaluate early infant diagnosis turnaround times in Lesotho. Trained staff reviewed records of HIV-exposed infants (aged-6-8 weeks) who received an HIV test during 2011. Study sites were drawn from Highlands, Foothills and Lowlands regions of Lesotho. Central laboratory database data were linked to facility and laboratory register information. Turnaround time geometric means …

A Review Of The Multiple-Sample Tests For The Continuous-Data Type, Dewi Rahardja

Journal of Modern Applied Statistical Methods

For continuous data, various statistical hypotheses testing methods have been extensively discussed in the literature. In this article a review is provided of the multiple-sample continuous-data testing methods. It includes traditional methods, such as the two-sample t-test, Welch ANOVA test, etc., as well as newly-developed ones, such as the various Multiple Comparison Procedure (MCP). A roadmap is provided in a figure or diagram format as to which methods are available in the literature. Additionally, the implementation of these methods in popular statistical software packages such as SAS is also presented. This review will be helpful to determine which continuous-data testing …

Factor Analysis By Limited Scales: Which Factors To Analyze?, Stan Lipovetsky

Journal of Modern Applied Statistical Methods

Factor Analysis (FA) and Principal Component Analysis (PCA) are well-known main tools of the multivariate statistics for data analysis, reduction, and visualization. Commonly, the analysis and interpretation of their solutions is performed for each of several main eigenvectors with variances explaining a big part of the total variability in data. The recommendation is to determine if all the main vectors are really needed in the analysis, or some of them should be skipped if they correspond to the absence of the analyzing features. A simple criterion for identifying redundant vectors of loadings is their negative correlation with the vector of …

Prediction Of Percent Change In Linear Regression By Correlated Variables, Stan Lipovetsky

Journal of Modern Applied Statistical Methods

Multiple linear regression can be applied for predicting an individual value of dependent variable y by the given values of independent variables x. But it is not immediately clear how to estimate percent change in y due to changes in predictors, especially when those are correlated. This work considers several approaches to this problem, including its formulation via predictors adjusted by their correlation structure. Ordinary least squares regression is used, together with Shapley value regression and another model based on solving some system of differential equations. Numerical estimations performed for a real marketing research data demonstrate meaningful results. The considered …

Conventional Early Infant Diagnosis In Lesotho From Specimen Collection To Results Usage To Manage Patients: Where Are The Bottlenecks?, Appolinaire Tiam, Michelle M Gill, Heather J Hoffman, Anthony Isavwa, Mafusi Mokone, Matokelo Foso, Jeffrey T Safrit, Lynne M Mofenson, Thorkild Tylleskär, Laura Guay

Epidemiology Faculty Publications

Introduction

Early infant diagnosis is an important step in identifying children infected with HIV during the perinatal period or in utero. Multiple factors contribute to delayed antiretroviral treatment initiation for HIV-infected children, including delays in the early infant HIV diagnosis cascade.

Methods

We conducted a retrospective study to evaluate early infant diagnosis turnaround times in Lesotho. Trained staff reviewed records of HIV-exposed infants (aged-6-8 weeks) who received an HIV test during 2011. Study sites were drawn from Highlands, Foothills and Lowlands regions of Lesotho. Central laboratory database data were linked to facility and laboratory register information. Turnaround time geometric means …

Perennial Warm-Season Grasses For Producing Biofuel And Enhancing Soil Properties: An Alternative To Corn Residue Removal, Humberto Blanco-Canqui, Robert B. Mitchell, Virginia L. Jin, Marty R. Schmer, Kent M. Eskridge

Department of Statistics: Faculty Publications

Removal of corn (Zea mays L.) residues at high rates for biofuel and other off-farm uses may negatively impact soil and the environment in the long term. Biomass removal from perennial warm-season grasses (WSGs) grown in marginally-productive lands could be an alternative to corn residue removal as biofuel feedstocks while controlling water and wind erosion, sequestering carbon (C), cycling water and nutrients, and enhancing other soil ecosystem services. We compared wind and water erosion potential, soil compaction, soil hydraulic properties, soil organic C (SOC), and soil fertility between biomass removal from WSGs and corn residue removal from rainfed no-till …

Impact Of Menthol Smoking On Nicotine Dependence For Diverse Racial/Ethnic Groups Of Daily Smokers, Julia N. Soulakova, Ryan R. Danczak

Department of Statistics: Faculty Publications

Introduction: The aims of this study were to evaluate whether menthol smoking and race/ethnicity are associated with nicotine dependence in daily smokers. Methods: The study used two subsamples of U.S. daily smokers who responded to the 2010–2011 Tobacco Use Supplement to the Current Population Survey. The larger subsample consisted of 18,849 non-Hispanic White (NHW), non-Hispanic Black (NHB), and Hispanic (HISP) smokers. The smaller subsample consisted of 1112 non-Hispanic American Indian/Alaska Native (AIAN), non-Hispanic Asian (ASIAN), non-Hispanic Hawaiian/Pacific Islander (HPI), and non-Hispanic Multiracial (MULT) smokers. Results: For larger (smaller) groups the rates were 45% (33%) for heavy smoking (16+ cig/day), 59% …

Evaluating Current Practices In Shelf Life Estimation, Robert Capen, David Christopher, Patrick Forenzo, Kim Huynh-Ba, David Leblond, Oscar Liu, John O'Neill, Nate Patterson, Michelle Quinlan, Radhika Rajagopalan, James Schwenke, Walter W. Stroup

Department of Statistics: Faculty Publications

The current International Council for Harmonisation of Technical Requirements for Pharmaceuticals for Human Use (ICH) methods for determining the supported shelf life of a drug product, described in ICH guidance documents Q1A and Q1E, are evaluated in this paper. To support this evaluation, an industry data set is used which is comprised of 26 individual stability batches of a common drug product where most batches are measured over a 24 month storage period. Using randomly sampled sets of 3 or 6 batches from the industry data set, the current ICH methods are assessed from three perspectives. First, the distributional properties …

Generalized Confidence Intervals Compatible With The Min Test For Simultaneous Comparisons Of One Subpopulation To Several Other Subpopulations, Julia N. Soulakova

Department of Statistics: Faculty Publications

A problem where one subpopulation is compared to several other subpopulations in terms of means with the goal of estimating the smallest difference between the means commonly arises in biology, medicine, and many other scientific fields. A generalization of Strassburger, Bretz and Hochberg (2004) approach for two comparisons is presented for cases with three and more comparisons. The method allows constructing an interval-estimator for the smallest mean difference, which is compatible with the Min test. An application to a fluency-disorder study is illustrated. Simulations confirmed adequate probability coverage for normally distributed outcomes for a number of designs.

Increasing Genomic-Enabled Prediction Accuracy By Modeling Genotype X Environment Interactions In Kansas Wheat, Diego Jarquin, Cristiano Lemas Da Silva, R. Chris Gaynor, Jesse Poland, Allan Fritz, Reka Howard, Sarah Battenfield, José Crossa

Department of Statistics: Faculty Publications

Wheat (Triticum aestivum L.) breeding programs test experimental lines in multiple locations over multiple years to get an accurate assessment of grain yield and yield stability. Selections in early generations of the breeding pipeline are based on information from only one or few locations and thus materials are advanced with little knowledge of the genotype × environment interaction (G × E) effects. Later, large trials are conducted in several locations to assess the performance of more advanced lines across environments. Genomic selection (GS) models that include G × E covariates allow us to borrow information not only from related …

Application Of Response Surface Methods To Determine Conditions For Optimal Genomic Prediction, Reka Howard, Alicia L. Carriquiry, William D. Beavis

Department of Statistics: Faculty Publications

An epistatic genetic architecture can have a significant impact on prediction accuracies of genomic prediction (GP) methods. Machine learning methods predict traits comprised of epistatic genetic architectures more accurately than statistical methods based on additive mixed linear models. The differences between these types of GP methods suggest a diagnostic for revealing genetic architectures underlying traits of interest. In addition to genetic architecture, the performance of GP methods may be influenced by the sample size of the training population, the number of QTL, and the proportion of phenotypic variability due to genotypic variability (heritability). Possible values for these factors and the …

Trans-Ancestry Fine Mapping And Molecular Assays Identify Regulatory Variants At The Angptl8 Hdl-C Gwas Locus, Maren E. Cannon, Qing Duan, Ying Wu, Monica Zeynalzadeh, Zheng Xu, Antti J. Kangas, Pasi Soininen, Mika Ala-Korpela, Mete Civelek, Aldons J. Lusis, Johanna Kuusisto, Francis S. Collins, Michael Boehnke, Hua Tang, Markku Laakso, Yun Li, Karen L. Mohlke

Department of Statistics: Faculty Publications

Recent genome-wide association studies (GWAS) have identified variants associated with highdensity lipoprotein cholesterol (HDL-C) located in or near the ANGPTL8 gene. Given the extensive sharing of GWAS loci across populations, we hypothesized that at least one shared variant at this locus affects HDL-C. The HDL-C–associated variants are coincident with expression quantitative trait loci for ANGPTL8 and DOCK6 in subcutaneous adipose tissue; however, only ANGPTL8 expression levels are associated with HDL-C levels. We identified a 400-bp promoter region of ANGPTL8 and enhancer regions within 5 kb that contribute to regulating expression in liver and adipose. To identify variants functionally responsible for …

A Bayes Interpretation Of Stacking For M-Complete And M-Open Settings, Tri Le, Bertrand S. Clarke

Department of Statistics: Faculty Publications

In M-open problems where no true model can be conceptualized, it is common to back off from modeling and merely seek good prediction. Even in M-complete problems, taking a predictive approach can be very useful. Stacking is a model averaging procedure that gives a composite predictor by combining individual predictors from a list of models using weights that optimize a cross validation criterion. We show that the stacking weights also asymptotically minimize a posterior expected loss. Hence we formally provide a Bayesian justification for cross-validation. Often the weights are constrained to be positive and sum to one. For greater generality, …

Analyzing Electricity Use Of Low Income Weatherization Program Participants Using Propensity Score Analysis And A Hierarchical Linear Growth Model, Ksenia Polson

Electronic Theses and Dissertations

This evaluation utilized propensity score matching methods and a longitudinal hierarchical linear growth model to determine the effect of residential energy efficiency upgrade(s) on household electricity use for the low-income community over the course of a year in the City and County of Denver, Colorado. Propensity score analysis with risk set matching was performed at each month under analysis applying nearest neighbor and nearest neighbor with caliper approaches by balancing covariates across the treatment and control groups. Following the completion of propensity score analysis, the data were aggregated to form a data set that was used in a hierarchical linear …

An Evaluation Of Critical Realignment Theory: Comparing Bayesian And Frequentist Approaches, Tara A. Rhodes

Electronic Theses and Dissertations

Prior to this study, critical realignment theory, which presupposes eras of substantial and sustained swings in American political party dominance, had only been evaluated using the classical, frequentist approach to modeling. However, potential for more information concerning these electoral phenomena exists given a shift in the design and approach to realigning elections. This study sought to explore those options through one particular alternative to the classical approach to statistics--in this particular case, the Bayesian approach to statistics. Bayesian methods differ from the frequentist approach in three main ways: the treatment of probability, the treatment of parameters, and the treatment of …

Consumers' Perceptions Of Voluntary And Involuntary Deconsumption: An Exploratory Sequential Scale Development Study, Kranti K. Dugar

Electronic Theses and Dissertations

This exploratory sequential mixed methods study of scale development was conducted among baby boomers in the United States to render conceptual clarity to the concepts of voluntary and involuntary deconsumption, to explore deconsumption behavior under the tenets of the attribution theory of motivation, and to examine the components, structures, uses, and measurement properties of scales of voluntary and involuntary deconsumption. It was also an attempt to reiterate the importance of the baby boomer segment(s) for marketing practitioners based on growth, economic viability, and the power of influence, and to establish a deep understanding of the deconsumption processes, which could enable …