Physical Sciences and Mathematics Commons^™

Generalized Matrix Decomposition Regression: Estimation And Inference For Two-Way Structured Data, Yue Wang, Ali Shojaie, Tim Randolph, Jing Ma

UW Biostatistics Working Paper Series

Analysis of two-way structured data, i.e., data with structures among both variables and samples, is becoming increasingly common in ecology, biology and neuro-science. Classical dimension-reduction tools, such as the singular value decomposition (SVD), may perform poorly for two-way structured data. The generalized matrix decomposition (GMD, Allen et al., 2014) extends the SVD to two-way structured data and thus constructs singular vectors that account for both structures. While the GMD is a useful dimension-reduction tool for exploratory analysis of two-way structured data, it is unsupervised and cannot be used to assess the association between such data and an outcome of interest. …

Statistical Inference For Networks Of High-Dimensional Point Processes, Xu Wang, Mladen Kolar, Ali Shojaie

UW Biostatistics Working Paper Series

Fueled in part by recent applications in neuroscience, high-dimensional Hawkes process have become a popular tool for modeling the network of interactions among multivariate point process data. While evaluating the uncertainty of the network estimates is critical in scientific applications, existing methodological and theoretical work have only focused on estimation. To bridge this gap, this paper proposes a high-dimensional statistical inference procedure with theoretical guarantees for multivariate Hawkes process. Key to this inference procedure is a new concentration inequality on the first- and second-order statistics for integrated stochastic processes, which summarizes the entire history of the process. We apply this …

Concentrations Of Criteria Pollutants In The Contiguous U.S., 1979 – 2015: Role Of Model Parsimony In Integrated Empirical Geographic Regression, Sun-Young Kim, Matthew Bechle, Steve Hankey, Elizabeth (Lianne) A. Sheppard, Adam A. Szpiro, Julian D. Marshall

UW Biostatistics Working Paper Series

BACKGROUND: National- or regional-scale prediction models that estimate individual-level air pollution concentrations commonly include hundreds of geographic variables. However, these many variables may not be necessary and parsimonious approach including small numbers of variables may achieve sufficient prediction ability. This parsimonious approach can also be applied to most criteria pollutants. This approach will be powerful when generating publicly available datasets of model predictions that support research in environmental health and other fields. OBJECTIVES: We aim to (1) build annual-average integrated empirical geographic (IEG) regression models for the contiguous U.S. for six criteria pollutants, for all years with regulatory monitoring data …

Robust Inference For The Stepped Wedge Design, James P. Hughes, Patrick J. Heagerty, Fan Xia, Yuqi Ren

UW Biostatistics Working Paper Series

Based on a permutation argument, we derive a closed form expression for an estimate of the treatment effect, along with its standard error, in a stepped wedge design. We show that these estimates are robust to misspecification of both the mean and covariance structure of the underlying data-generating mechanism, thereby providing a robust approach to inference for the treatment effect in stepped wedge designs. We use simulations to evaluate the type I error and power of the proposed estimate and to compare the performance of the proposed estimate to the optimal estimate when the correct model specification is known. The …

Using Multilevel Outcomes To Construct And Select Biomarker Combinations For Single-Level Prediction, Allison Meisner, Chirag R. Parikh, Kathleen F. Kerr

UW Biostatistics Working Paper Series

Biomarker studies may involve a multilevel outcome, such as no, mild, or severe disease. There is often interest in predicting one particular level of the outcome due to its clinical significance. The standard approach to constructing biomarker combinations in this context involves dichotomizing the outcome and using a binary logistic regression model. We assessed whether information can be usefully gained from instead using more sophisticated regression methods. Furthermore, it is often necessary to select among several candidate biomarker combinations. One strategy involves selecting a combination on the basis of its ability to predict the outcome level of interest. We propose …

Nonparametric Variable Importance Assessment Using Machine Learning Techniques, Brian D. Williamson, Peter B. Gilbert, Noah Simon, Marco Carone

UW Biostatistics Working Paper Series

In a regression setting, it is often of interest to quantify the importance of various features in predicting the response. Commonly, the variable importance measure used is determined by the regression technique employed. For this reason, practitioners often only resort to one of a few regression techniques for which a variable importance measure is naturally defined. Unfortunately, these regression techniques are often sub-optimal for predicting response. Additionally, because the variable importance measures native to different regression techniques generally have a different interpretation, comparisons across techniques can be difficult. In this work, we study a novel variable importance measure that can …

Combining Biomarkers By Maximizing The True Positive Rate For A Fixed False Positive Rate, Allison Meisner, Marco Carone, Margaret Pepe, Kathleen F. Kerr

UW Biostatistics Working Paper Series

Biomarkers abound in many areas of clinical research, and often investigators are interested in combining them for diagnosis, prognosis and screening. In many applications, the true positive rate for a biomarker combination at a prespecified, clinically acceptable false positive rate is the most relevant measure of predictive capacity. We propose a distribution-free method for constructing biomarker combinations by maximizing the true positive rate while constraining the false positive rate. Theoretical results demonstrate good operating characteristics for the resulting combination. In simulations, the biomarker combination provided by our method demonstrated improved operating characteristics in a variety of scenarios when compared with …

Developing Biomarker Combinations In Multicenter Studies Via Direct Maximization And Penalization, Allison Meisner, Chirag R. Parikh, Kathleen F. Kerr

UW Biostatistics Working Paper Series

When biomarker studies involve patients at multiple centers and the goal is to develop biomarker combinations for diagnosis, prognosis, or screening, we consider evaluating the predictive capacity of a given combination with the center-adjusted AUC (aAUC), a summary of conditional performance. Rather than using a general method to construct the biomarker combination, such as logistic regression, we propose estimating the combination by directly maximizing the aAUC. Furthermore, it may be desirable to have a biomarker combination with similar predictive capacity across centers. To that end, we allow for penalization of the variability in center-specific performance. We demonstrate good asymptotic properties …

Evaluation Of Multiple Interventions Using A Stepped Wedge Design, Vivian H. Lyons, Lingyu Li, James Hughes, Ali Rowhani-Rahbar

UW Biostatistics Working Paper Series

Background: Stepped wedge cluster randomized trials are a class of unidirectional crossover studies that have historically been limited to evaluating a single intervention. This design is especially suitable for pragmatic trials where the study feasibility can be improved with a phased introduction of the intervention. We examined variations of stepped wedge designs that would support evaluation of multiple interventions. Methods: We propose four different design variants for implementing a stepped wedge trial with two interventions: concurrent design, supplementation, replacement, and factorial designs. Analyses were conducted comparing the precision of the estimated intervention effects for the different designs. Results: Concurrent, …

Biomarker Combinations For Diagnosis And Prognosis In Multicenter Studies: Principles And Methods, Allison Meisner, Chirag R. Parikh, Kathleen F. Kerr

UW Biostatistics Working Paper Series

Many investigators are interested in combining biomarkers to predict an outcome of interest or detect underlying disease. This endeavor is complicated by the fact that many biomarker studies involve data from multiple centers. Depending upon the relationship between center, the biomarkers, and the target of prediction, care must be taken when constructing and evaluating combinations of biomarkers. We introduce a taxonomy to describe the role of center and consider how a biomarker combination should be constructed and evaluated. We show that ignoring center, which is frequently done by clinical researchers, is often not appropriate. The limited statistical literature proposes using …

Adaptive Non-Inferiority Margins Under Observable Non-Constancy, Brett S. Hanscom, Deborah J. Donnell, Brian D. Williamson, Jim Hughes

UW Biostatistics Working Paper Series

A central assumption in the design and conduct of non-inferiority trials is that the active-control therapy will have the same degree of effectiveness in the planned non-inferiority trial as it had in the prior placebo-controlled trials used to define the non-inferiority margin. This is referred to as the `constancy' assumption. If the constancy assumption fails, the chosen non-inferiority margin is not valid and the study runs the risk of approving an inferior product or failing to approve a beneficial product. The constancy assumption cannot be validated in a trial without a placebo arm, and it is unlikely ever to be …

Predicting Future Years Of Life, Health, And Functional Ability: A Healthy Life Calculator For Older Adults, Paula Diehr, Michael Diehr, Alice M. Arnold, Laura Yee, Michelle C. Odden, Calvin H. Hirsch, Stephen Thielke, Bruce Psaty, W Craig Johnson, Jorge Kizer, Anne B. Newman

UW Biostatistics Working Paper Series

Introduction

Planning for the future would be easier if we knew how long we will live and, more importantly, how many years we will be healthy and able to enjoy it. There are few well-documented aids for predicting our future health. We attempted to meet this need for persons 65 years of age and older.

Methods

Data came from the Cardiovascular Health Study, a large longitudinal study of older adults that began in 1990. Years of life (YOL) were defined by measuring time to death. Years of healthy life (YHL) were defined by an annual question about self-rated health, and …

Confidence Intervals For Heritability Via Haseman-Elston Regression, Tamar Sofer

UW Biostatistics Working Paper Series

Heritability is the proportion of phenotypic variance in a population that is attributable to individual genotypes. Heritability is considered an important measure in both evolutionary biology and in medicine, and is routinely estimated and reported in genetic epidemiology studies. In population-based genome-wide association studies (GWAS), mixed models are used to estimate variance components, from which a heritability estimate is obtained. The estimated heritability is the proportion of the model's total variance that is due to the genetic relatedness matrix (kinship measured from genotypes). Current practice is to use bootstrapping, which is slow, or normal asymptotic approximation to estimate the precision …

A Powerful Statistical Framework For Generalization Testing In Gwas, With Application To The Hchs/Sol, Tamar Sofer, Ruth Heller, Marina Bogomolov, Christy L. Avery, Mariaelisa Graff, Kari E. North, Alex Reiner, Timothy A. Thornton, Kenneth Rice, Yoav Benjamini, Cathy C. Laurie, Kathleen F. Kerr

UW Biostatistics Working Paper Series

In GWAS, “generalization” is the replication of genotype-phenotype association in a population with different ancestry than the population in which it was first identified. The standard for reporting findings from a GWAS requires a two-stage design, in which discovered associations are replicated in an independent follow-up study. Current practices for declaring generalizations rely on testing associations while controlling the Family Wise Error Rate (FWER) in the discovery study, then separately controlling error measures in the follow-up study. While this approach limits false generalizations, we show that it does not guarantee control over the FWER or False Discovery Rate (FDR) of …

Methods For Dealing With Death And Missing Data, And For Standardizing Different Health Variables In Longitudinal Datasets: The Cardiovascular Health Study, Paula Diehr

UW Biostatistics Working Paper Series

Longitudinal studies of older adults usually need to account for deaths and missing data. The study databases often include multiple health-related variables, whose trends over time are hard to compare because they were measured on different scales. Here we present a unified approach to these three problems that was developed and used in the Cardiovascular Health Study. Data were first transformed to a new scale that had integer/ratio properties, and on which “dead” logically takes the value zero. Missing data were then imputed on this new scale, using each person’s own data over time. Imputation could thus be informed by …

Recommendation To Use Exact P-Values In Biomarker Discovery Research, Margaret Sullivan Pepe, Matthew F. Buas, Christopher I. Li, Garnet L. Anderson

UW Biostatistics Working Paper Series

Background: In biomarker discovery studies, markers are ranked for validation using P-values. Standard P-value calculations use normal approximations that may not be valid for small P-values and small sample sizes common in discovery research.

Methods: We compared exact P-values, valid by definition, with normal and logit-normal approximations in a simulated study of 40 cases and 160 controls. The key measure of biomarker performance was sensitivity at 90% specificity. Data for 3000 uninformative markers and 30 true markers were generated randomly, with 10 replications of the simulation. We also analyzed real data on 2371 antibody array markers …

Models For Hsv Shedding Must Account For Two Levels Of Overdispersion, Amalia Magaret

UW Biostatistics Working Paper Series

We have frequently implemented crossover studies to evaluate new therapeutic interventions for genital herpes simplex virus infection. The outcome measured to assess the efficacy of interventions on herpes disease severity is the viral shedding rate, defined as the frequency of detection of HSV on the genital skin and mucosa. We performed a simulation study to ascertain whether our standard model, which we have used previously, was appropriately considering all the necessary features of the shedding data to provide correct inference. We simulated shedding data under our standard, validated assumptions and assessed the ability of 5 different models to reproduce the …

Meta-Analysis Of Genome-Wide Association Studies With Correlated Individuals: Application To The Hispanic Community Health Study/Study Of Latinos (Hchs/Sol), Tamar Sofer, John R. Shaffer, Misa Graff, Qibin Qi, Adrienne M. Stilp, Stephanie M. Gogarten, Kari E. North, Carmen R. Isasi, Cathy C. Laurie, Adam A. Szpiro

UW Biostatistics Working Paper Series

Investigators often meta-analyze multiple genome-wide association studies (GWASs) to increase the power to detect associations of single nucleotide polymorphisms (SNPs) with a trait. Meta-analysis is also performed within a single cohort that is stratified by, e.g., sex or ancestry group. Having correlated individuals among the strata may complicate meta-analyses, limit power, and inflate Type 1 error. For example, in the Hispanic Community Health Study/Study of Latinos (HCHS/SOL), sources of correlation include genetic relatedness, shared household, and shared community. We propose a novel mixed-effect model for meta-analysis, “MetaCor", which accounts for correlation between stratum-specific effect estimates. Simulations show that MetaCor controls …

Historical Prediction Modeling Approach For Estimating Long-Term Concentrations Of Pm In Cohort Studies Before The 1999 Implementation Of Widespread Monitoring, Sun-Young Kim, Casey Olives, Lianne Sheppard, Paul D. Sampson, Timothy V. Larson, Joel Kaufman

UW Biostatistics Working Paper Series

Introduction: Recent cohort studies use exposure prediction models to estimate the association between long-term residential concentrations of PM_2.5 and health. Because these prediction models rely on PM_2.5 monitoring data, predictions for times before extensive spatial monitoring present a challenge to understanding long-term exposure effects. The Environmental Protection Agency (EPA) Federal Reference Method (FRM) network for PM_2.5 was established in 1999. We evaluated a novel statistical approach to produce high quality exposure predictions from 1980-2010 for epidemiological applications.

Methods: We developed spatio-temporal prediction models using geographic predictors and annual average PM_2.5 data from 1999 through 2010 from …

Stochastic Optimization Via Forward Slice, Bob A. Salim, Lurdes Y. T. Inoue

UW Biostatistics Working Paper Series

Optimization consists of maximizing or minimizing a real-valued objective function. In many problems, the objective function may not yield closed-form solutions. Over many decades, optimization methods, both deterministic and stochastic, have been developed to provide solutions to these problems. However, some common limitations of these methods are the sensitivity to the initial value and that often current methods only find a local (non-global) extremum. In this article, we propose an alternative stochastic optimization method, which we call "Forward Slice", and assess its performance relative to available optimization methods.

Testing Gene-Environment Interactions In The Presence Of Measurement Error, Chongzhi Di, Li Hsu, Charles Kooperberg, Alex Reiner, Ross Prentice

UW Biostatistics Working Paper Series

Complex diseases result from an interplay between genetic and environmental risk factors, and it is of great interest to study the gene-environment interaction (GxE) to understand the etiology of complex diseases. Recent developments in genetics field allows one to study GxE systematically. However, one difficulty with GxE arises from the fact that environmental exposures are often measured with error. In this paper, we focus on testing GxE when the environmental exposure E is subject to measurement error. Surprisingly, contrast to the well-established results that the naive test ignoring measurement error is valid in testing the main effects, we find that …

Personalized Evaluation Of Biomarker Value: A Cost-Benefit Perspective, Ying Huang, Eric Laber

UW Biostatistics Working Paper Series

For a patient who is facing a treatment decision, the added value of information provided by a biomarker depends on the individual patient’s expected response to treatment with and without the biomarker, as well as his/her tolerance of disease and treatment harm. However, individualized estimators of the value of a biomarker are lacking. We propose a new graphical tool named the subject-specific expected benefit curve for quantifying the personalized value of a biomarker in aiding a treatment decision. We develop semiparametric estimators for two general settings: i) when biomarker data are available from a randomized trial; and ii) when biomarker …

Nonparametric Identifiability Of Finite Mixture Models With Covariates For Estimating Error Rate Without A Gold Standard, Zheyu Wang, Xiao-Hua Zhou

UW Biostatistics Working Paper Series

Finite mixture models provide a flexible framework to study unobserved entities and have arisen in many statistical applications. The flexibility of these models in adapting various complicated structures makes it crucial to establish model identifiability when applying them in practice to ensure study validity and interpretation. However, researches to establish the identifiability of finite mixture model are limited and are usually restricted to a few specific model configurations. Conditions for model identifiability in the general case have not been established. In this paper, we provide conditions for both local identifiability and global identifiability of a finite mixture model. The former …

Efficiently Identifying Failures Using Quantitative Tests, Matrix-Pooling And The Em-Algorithm, Brett Hanscom, Susanne May, Jim Hughes

UW Biostatistics Working Paper Series

Pooled-testing methods can greatly reduce the number of tests needed to identify failures in a collection of samples. Existing methodology has focused primarily on binary tests, but there is a clear need for improved efficiency when using expensive quantitative tests, such as tests for HIV viral load in resource-limited settings. We propose a matrix-pooling method which, based on pooled-test results, uses the EM algorithm to identify individual samples most likely to be failures. Two hundred datasets for each of a wide range of failure prevalence were simulated to test the method. When the measurement of interest was normally distributed, at …

A Joint Model For Multistate Disease Processes And Random Informative Observation Times, With Applications To Electronic Medical Records Data, Jane M. Lange, Rebecca A. Hubbard, Lurdes Y. T. Inoue, Vladimir Minin

UW Biostatistics Working Paper Series

Multistate models are used to characterize individuals' natural histories through diseases with discrete states. Observational data resources based on electronic medical records pose new opportunities for studying such diseases. However, these data consist of observations of the process at discrete sampling times, which may either be pre-scheduled and non-informative, or symptom-driven and informative about an individual's underlying disease status. We have developed a novel joint observation and disease transition model for this setting. The disease process is modeled according to a latent continuous time Markov chain; and the observation process, according to a Markov-modulated Poisson process with observation rates that …

Change Point Testing In Logistic Regression Models With Interaction Term, Youyi Fong, Chongzhi Di, Sallie Permar

UW Biostatistics Working Paper Series

The threshold effect takes place in situations where the relationship between an outcome variable and a predictor variable changes as the predictor value crosses a certain threshold/change point. Threshold effects are often plausible in a complex biological system, especially in defining immune responses that are protective against infections such as HIV-1, which motivates the current work. We study two hypothesis testing problems in change point models. We first compare three different approaches to obtaining a p-value for the maximum of scores test in a logistic regression model with change point variable as a main effect. Next, we study the testing …

Multi-State Models For Natural History Of Disease, Amy Laird, Rebecca A. Hubbard, Lurdes Y. T. Inoue

UW Biostatistics Working Paper Series

Longitudinal studies are a useful tool for investigating the course of chronic diseases. Many chronic diseases can be characterized by a set of health states. We can improve our understanding of the natural history of the disease by modeling the sequence of visited health states and the duration in each state. However, in most applications, subjects are observed only intermittently. This observation scheme creates a major modeling challenge: the transition times are not known exactly, and in some cases the path through the health states is not known.

In this manuscript we review existing approaches for modeling multi-state longitudinal data. …

Issues Related To Combining Multiple Speciated Pm2.5 Data Sources In Spatio-Temporal Exposure Models For Epidemiology: The Npact Case Study, Sun-Young Kim, Lianne Sheppard, Timothy V. Larson, Joel Kaufman, Sverre Vedal

UW Biostatistics Working Paper Series

Background: Regulatory monitoring data have been the most common exposure data resource in studies of the association between long-term PM_2.5 components and health. However, data collected for regulatory purposes may not be compatible with epidemiological study.

Objectives: We aimed to explore three important features of the PM_2.5 component monitoring data obtained from multiple sources to combine all available data for developing spatio-temporal prediction models in the National Particle Component and Toxicity (NPACT) study.

Methods: The NPACT monitoring data were collected in an extensive monitoring campaign targeting cohort participants. The regulatory monitoring data were obtained from the Chemical Speciation …

Prediction Of Fine Particulate Matter Chemical Components For The Multi-Ethnic Study Of Atherosclerosis Cohort: A Comparison Of Two Modeling Approaches, Sun-Young Kim, Lianne Sheppard, Silas Bergen, Adam A. Szpiro, Paul D. Sampson, Joel Kaufman, Sverre Vedal

UW Biostatistics Working Paper Series

Recent epidemiological cohort studies of the health effects of PM_2.5 have developed exposure estimates from advanced exposure prediction models. Such models represent spatial variability across participant residential locations. However, few cohort studies have developed exposure predictions for PM_2.5 components. We used two exposure modeling approaches to obtain long-term average predicted concentrations for four PM_2.5 components: sulfur, silicon, and elemental and organic carbon (EC and OC). The models were specifically developed for the Multi-Ethnic Study of Atherosclerosis (MESA) cohort as a part of the National Particle Component and Toxicity (NPACT) study. The spatio-temporal model used 2-week average measurements …

Characterizing Expected Benefits Of Biomarkers In Treatment Selection, Ying Huang, Eric Laber, Holly Janes

UW Biostatistics Working Paper Series

Biomarkers associated with the treatment response heterogeneity hold potential for treatment selection. In practice, the decision regarding whether to adopt a treatment selection marker depends on the effect of treatment selection on the rate of targeted disease as well as additional cost associated with the treatment. We propose an expected benefit measure that incorporates both aspects to quantify a biomarker's treatment selection capacity. This measure extends an existing decision-theoretic framework, to account for the fact that optimal treatment absent marker information varies with the cost of treatment. In addition, we establish upper and lower bounds for the performance of a …