Oct19Virtual12:00pm - 1:00pm
Mauricio Sadinle, PhD
Assistant Professor, Department of Biostatistics
University of Washington, School of Public Health
Sequentially additive nonignorable missing data modelling using auxiliary marginal information
Abstract: We study a class of missingness mechanisms, referred to as sequentially additive nonignorable, for modelling multivariate data with item nonresponse. These mechanisms explicitly allow the probability of nonresponse for each variable to depend on the value of that variable, thereby representing nonignorable missingness mechanisms. These missing data models are identified by making use of auxiliary information on marginal distributions, such as marginal probabilities for multivariate categorical variables or moments for numeric variables. We prove identification results and illustrate the use of these mechanisms in an application.
Bio: Mauricio Sadinle, PhD is an Assistant Professor in the Department of Biostatistics at the University of Washington. Previously, he was a Postdoctoral Associate in the Department of Statistical Science at Duke University and the National Institute of Statistical Sciences, working under the mentoring of Jerry Reiter. He completed his PhD in the Department of Statistics at Carnegie Mellon University, where his advisor was Stephen E. Fienberg. Dr. Sadinle’s undergraduate studies are from the National University of Colombia in Bogota, where he majored in statistics. Dr. Sadinle’s methodological research mainly focuses on 1. Record linkage techniques to combine datafiles that contain information on overlapping sets of individuals but lack unique identifiers and 2. Nonignorable missing data modeling, and the usage of auxiliary information to identify nonignorable missing data mechanisms. Dr. Sadinle also has experience working with social network models for valued ties, capture-recapture models in the context of human rights violations, and set-valued classifiers that output sets of plausible labels for ambiguous sample points.
For more information about the Statistics Seminar Series go here .
Oct5Virtual12:00pm - 1:00pm
Stephanie Shipp, PhD
Deputy Director and Professor
Social and Decision Analytics Division, Biocomplexity Institute
University of Virginia
Abstract Title: ”Ethical Principles and Data Science - Repurposing Administrative & Opportunity Data”
The data revolution is changing the conduct of research as increasing amounts of internet-based and administrative data become accessible for use. At the same time, the new data landscape has created signiﬁcant tension around data privacy and conﬁdentiality. To bridge this gap, conversations about ethics, privacy, transparency, and reproducibility need to play a prominent role in both research partnerships and policymaking. At the research level, these conversations must be translated to action. We have created a comprehensive framework that forms the foundation to data science problem solving through defining rigorous, flexible, and iterative processes where learning at each stage informs the other stages. Embedded in this framework is close attention to ethics. The Institutional Review Board structure is well known in parts of academia and industry, but our public and local government partners are not always aware of these processes. The IRB framework could help them think about informed consent and privacy, as well as ethical considerations around the benefits and risks to individuals and communities under study. Through case studies, these principles are demonstrated.
Keywords: conﬁdentiality, ethics, trust but verify, data science framework
Brief Bio: Data scientists have the opportunity to use their skills to influence and improve society, especially vulnerable populations who need champions. Stephanie Shipp enthusiastically works with communities, policy makers and other data scientists who have also taken that challenge to heart.
For more information about the Statistics Seminar Series go here .
Sep28Virtual12:00pm - 1:00pm
Despina Kontos, PhD
Associate Professor of Radiology
Department of Radiology
Perelman School of Public Health
University of Pennsylvania
Bio: Dr. Despina Kontos, Ph.D., is an Associate Professor of Radiology and director of the Computational Biomarker Imaging Group (CBIG) in the Center for Biomedical Image Computing and Analytics (CBICA) at the Radiology Department of the University of Pennsylvania. Dr. Kontos received her C.Eng. Diploma in Computer Engineering and Informatics from the University of Patras in Greece and her MSc and Ph.D. degrees in Computer Science from Temple University in Philadelphia. She completed her postdoctoral training in radiologic physics and biostatistics at the University of Pennsylvania. Her research interests focus on investigating the role of quantitative imaging as a predictive biomarker for guiding personalized clinical decisions in cancer screening, prognosis, and treatment. She is leading several research studies, funded both by the NIH/NCI and private foundations, to incorporate novel quantitative multi-modality imaging measures of breast tumor and tissue composition into cancer risk prediction models.
Title: “Radiomic Biomarkers for Deciphering Tumor Heterogeneity”
Abstract - Breast cancer is a heterogeneous disease, with known inter-tumor and intra-tumor heterogeneity. Established histopathologic prognostic biomarkers generally acquired from a tumor biopsy may be limited by sampling variation. Radiomics is an emerging field with the potential to leverage the whole tumor via non-invasive sampling afforded by medical imaging to extract high throughput, quantitative features for personalized tumor characterization. Identifying imaging phenotypes via radiomics analysis and understanding their relationship with prognostic markers and patient outcomes can allow for a non-invasive assessment of tumor heterogeneity. In this study, we identified and independently validated intrinsic radiomic phenotypes of tumor heterogeneity for invasive breast cancer that have independent prognostic value when predicting 10-year recurrence. The independent and additional prognostic value of imaging heterogeneity phenotypes suggests that radiomic phenotypes can provide a non-invasive characterization of tumor heterogeneity to augment personalized prognosis and treatment.
For more information about the Statistics Seminar Series, click here .
Sep14Virtual12:00pm - 1:00pm
Jianqing Fan, PhD, Professor of Statistics
Frederick L. Moore ’18 Professor of Finance
Title: Communication—Efficient Accurate Statistical Estimation
Abstract: When the data are stored in a distributed manner, direct application of traditional statistical inference procedures is often prohibitive due to communication cost and privacy concerns. This paper develops and investigates two Communication-Efficient Accurate Statistical Estimators (CEASE), implemented through iterative algorithms for distributed optimization. In each iteration, node machines carry out computation in parallel and communicates with the central processor, which then broadcasts aggregated gradient vector to node machines for new updates. The algorithms adapt to the similarity among loss functions on node machines, and converge rapidly when each node machine has large enough sample size. Moreover, they do not require good initialization and enjoy linear converge guarantees under general conditions. The contraction rate of optimization errors is derived explicitly, with dependence on the local sample size unveiled. In addition, the improved statistical accuracy per iteration is derived. By regarding the proposed method as a multi-step statistical estimator, we show that statistical efficiency can be achieved infinite steps in typical statistical applications. In addition, we give the conditions under which one-step CEASE estimator is statistically efficient. Extensive numerical experiments on both synthetic and real data validate the theoretical results and demonstrate the superior performance of our algorithms.
(Joint work with Yongyi Guo and Kaizheng Wang)
For more information about the Statistics Seminar Series, click here
Apr612:00pm - 1:00pm
Daniel Almirall, PhD, Associate Professor, Co-Director, Data Science for Dynamic Intervention Decision-making Laboratory (d3lab), Survey Research Center, Institute for Social Research; Department of Statistics, College of Literature Sciences and the Arts, University of Michigan
Bio: Daniel Almirallis Associate Professor in the Institute for Social Research and the Department of Statistics at the University of Michigan. He is a methodologist and statistician who develops methods to form evidence-based adaptive interventions. Adaptive interventions can be used to inform individualized intervention guidelines for the on-going management of chronic illnesses or disorders such as drug abuse, depression, anxiety, autism, obesity, or HIV/AIDS. More recently, Dr. Almirall has been developing methods to form adaptive implementation interventions, to inform how best to tailor sequences of organizational-level strategies to improve the implementation of evidence-based practices. His work includes the development of approaches related to the design, execution, and analysis of sequential multiple assignment randomized trials (SMARTs) which can be used to build adaptive interventions, and of clustered SMARTs to build adaptive implementation interventions. He is particularly interested in applications in child and adolescent mental health research.
Abstract to be posted
Mar93:00pm - 4:00pm
Sumithra J. Mandrekar, PhD, Professor of Biostatistics and Oncology, Mayo Clinic
Clinical Trial Designs for Personalized Medicine in Oncology
Clinical trial design strategies have evolved as a means to accelerate the drug development process so that the right therapies can be delivered to the right patients. Basket, umbrella, SMART and adaptive enrichment strategies represent a class of novel designs for testing targeted therapeutics, and individualizing treatment in oncology. Umbrella trials include a central infrastructure for screening and identification of patients, and focus on a single tumor type or histology with multiple sub trials, each testing a targeted therapy within a molecularly defined subset. Basket trial designs offer the possibility to include multiple molecularly defined subpopulations, often across histology or tumor types, but included in one cohesive design to evaluate the targeted therapy in question. Adaptive enrichment designs offer the potential to enrich for patients with a particular molecular feature that is predictive of benefit for the test treatment based on accumulating evidence from the trial. A SMART design strategy is appropriate when comparing adaptive interventions, which are defined by an individual’s sequence of treatment decisions and guided by intermediate outcomes, such as response to therapy. This talk will focus on the fundamentals of these design strategies, the underlying statistical framework, the logistical barriers of implementation, and, ultimately, the interpretation of the trial results, using some case-studies including the National Cancer Institute’s precision medicine initiative trials.
Dr. Mandrekar received her interdisciplinary Ph.D. in Biostatistics, Psychology, Internal Medicine and Biomedical Engineering from the Ohio State University, Columbus OH in December 2002, and joined Mayo Clinic as a research associate in January 2003. She is currently Professor of Biostatistics and Oncology at the Mayo Clinic, Rochester MN, and Section Head for the Cancer Center Statistics within Mayo Clinic Department of Health Sciences Research. She is the Group Statistician for the Alliance for Clinical Trials in Oncology, which is one of the 4 NCI-funded national clinical trials networks for the conduct of phase II and III clinical trials in adult cancer. Dr. Mandrekar also holds an adjunct Professor of Biostatistics appointment at the School of Public Health, University of Minnesota, and at the University of Gainesville, FL.
Dr. Mandrekar’s primary research interests include adaptive dose-finding early phase trial designs, designs for predictive biomarker validation, and general clinical trial methodology related to conduct of clinical trials and identification of alternative Phase II cancer clinical trial endpoints. Dr. Mandrekar has co-authored over 140 original papers; several book chapters and editorials; and has given numerous lectures, invited presentations and workshops on these topics.
Her primary collaborative areas are lung cancer and leukemia. She is the faculty statistician for the national adjuvant lung cancer trial, ALCHEMIST, an approximately 8000 patient trial, which is part of the NCI precision medicine initiative. She was the primary statistician on the Phase III C10603 trial that led to the FDA approval of Midostaurin for AML patients with FLT3 mutations.
Dr. Mandrekar is a voting member of the NCI thoracic malignancies steering committee, past president of the society for clinical trials (2018-2019), voting member on the clinical trials transformation initiative on master protocols, member of the ASCO mCODE executive council, voting member of the international RECIST working group and the biostatistics editor for the Journal of Thoracic Oncology.
She has a national and international presence as an expert in clinical trial design as evidenced by her bibliography and membership on various national and international committees.
Dec93:00pm - 4:00pm
Ludovic Trinquart PhD, Assistant Professor of Biostatistics, Boston University School of Public Health
“Counterfactual mediation analysis with illness-death model for right-censored surrogate and clinical outcomes”
We introduce a counterfactual-based mediation analysis for surrogate outcome evaluation when both the surrogate and clinical endpoint are subject to right-censoring. We use a multistate model for risk prediction to account for both direct transitions towards the clinical endpoint and transitions through the surrogate endpoint. We use the counterfactual framework to define natural direct and indirect effects and the proportion of the treatment effect on the clinical endpoint mediated by the surrogate endpoint. We define these quantities for the cumulative risk and restricted mean survival time. We illustrate our approach using 18-year follow-up data from the SPCG-4 randomized controlled trial of radical prostatectomy for prostate cancer. We assess time to metastasis as a potential surrogate outcome for all-cause mortality.
Oct73:00pm - 4:00pm
Fan Li, PhD, Associate Professor, Department of Statistical Science, Duke University
“Introducing the Overlap Weights in Causal Inference”
Covariate balance is crucial for confounding adjustment in causal studies with observational data. We propose a unified framework —the balancing weights– to balance the weighted distributions of the covariates between treatment groups. These weights incorporate the propensity score to weight each group to an analyst-selected target population, and include several commonly used weighting schemes such as inverse-probability weight and trimming as special cases. We derive the large-sample results on nonparametric estimation based on these weights. We further propose a new weighting scheme, the overlap weights, in which each unit’s weight is proportional to the probability of that unit being assigned to the opposite group. The overlap weights are bounded, and minimize the asymptotic variance of the weighted average treatment effect among the class of balancing weights. We prove a small-sample exact balance property of the overlap weights. We apply the method the Framingham Heart Study to evaluate the effect of statins on health outcomes. Extension to multiple treatments will also be discussed.
Sep303:00pm - 4:00pm
Prasanta Pal, PhD, Research Project Director, U. Mass. Medical School, Adjunct Assistant Professor of Behavioral and Social Sciences, School of Public Health, Brown University
Bio: I trained as a Physicist at Indian Institute of Technology, Kharagpur, India and Yale University School of Engineering, New Haven, CT. Starting with computer simulation on super-conducting quantum computer at Yale , my PhD work was focused on large scale computer simulations of soft matter particle systems. I briefly worked as a postdoc at Yale MRI center to study blood flow dynamics in human heart. Then I head started the technological platform of the startup company “Mind sciences” to build mobile applications like “Craving to Quit” for addiction treatment through mindfulness interventions. In parallel I built high-density real-time EEG neuro-feedback systems to provide mental-training to different population groups. This has been part of large scale clinical trials and demonstrated in commercial program like CBS 60 minutes. Currently I’m developing tools for big-data, 4D visualization of human mind and its quantification using AI and computational geometry.
Abstract: “Chasing the Mind”
Background: the quest to understand the mind is age-old and unresolved. Although the “Mind” is central to nearly all human activity, the quest for a complete scientific understanding of “how the human mind works” remains unresolved. In science, much progress has been made to understand the nature and function of the mind using various techniques. Useful models of mental and emotional function have emerged, with an increasingly detailed picture of the brain-body connection and its many roles in health and survival.
Opportunity: Mind-body research reveals inner order, while technology opens frontiers for analysis. More recently, mystics of the east and clinicians of the west have come to recognize the common platform through which they seek to promote mental well being. Increasingly, the exploration of one’s inner self (e.g., through inner query and mental-training) can complement external tools like thermometers and drug regimens. The old notion of body-based medicine has evolved into mind-body medicine. Clinically, successful programs like MBSR (Mindful Based Stress Reduction) developed by Jon Kabat Zinn are geared towards harmonizing the physical and mental dimensions of human existence. Importantly, the inner order and wholeness perceived by mystics is gaining solid scientific footing and paving the way for a revolutionary new age of medicine.
On the other hand, the rapidly evolving knowledge-base of machine-learning, data science and technology are collapsing the boundaries between traditional academic disciplines. As a result, commercially available devices like MUSE for meditation and EEG sleep monitors are device incarnations to play with “Mind dynamics”.
Challenge: Reductionist frameworks miss hidden order in the mind’s complexity. The human brain is by far the most sophisticated computing-device known to science. High quality data measured from the human brain via EEG like sensors can effectively be utilized to create the cartography of the human mind.
Although simplistic reductionist models have very successfully permeated the domains of physical sciences, it is often limited in scope to explain the human mind. As an alternative, we introduce a more holistic approach to pave the cartographic building blocks of the human mind using high dimensional data collected through EEG sensors attached to the human brain.
Our work: New imaging and computational techniques to visualize the mind’s wholeness. As data scientists and technologists, our work is to build tools to visualize and investigate the human mind with as much scientific and mathematical rigor as complex fields like rocket science. The holy grail is to map out the deeper-most thought patterns in the human mind on our favorite mobile device. What has been inner can be unfolded to the outer.
In this work, we describe our unique journey to understand the human mind starting with a simple noisy one-dimensional time series plot of a small EEG data segment and, from there, building models with infinite-dimensional function space to characterize and visualize the experientially active human mind. We demonstrate that the inner journey of the self is also reciprocally, a number crunching exercise. With these tools, when fully developed, we can transform soulless-selfies to the speckles of our blissful inner-selves.
Sep233:00pm - 4:00pm
Lisa LaVange, PhD
bio: Lisa LaVange, PhD, is Professor and Associate Chair of the Department of Biostatistics in the Gillings School of Global Public Health at the University of North Carolina at Chapel Hill. She is also director of the department’s Collaborative Studies Coordinating Center (CSCC), overseeing faculty, staff, and students involved in large-scale clinical trials and epidemiological studies coordinated by the center. From 2011 to 2017, Dr. LaVange was director of the Office of Biostatistics in the United States Food and Drug Administration (FDA) Center for Drug Evaluation and Research (CDER). There, she oversaw more than 200 statisticians and other staff members involved in the development and application of statistical methodology for drug regulation. She was a leader in developing and assessing the effectiveness and appropriateness of innovative statistical methods intended to accelerate the process from drug discovery to clinical trials to FDA approval and patients’ benefit, with a particular focus on rare diseases. Prior to her government and academic experience, she spent 16 years in non-profit research and 10 years in the pharmaceutical industry. Dr. LaVange is an elected fellow of the American Statistical Association (ASA) and was the 2018 ASA President.
Abstract: Real-world Data, Statistical Innovation, and Evidence Generation
As a result of the 21st Century Cures Act and the Prescription Drug User’s Fee VI Act, both passed into law in 2016, the FDA launched several important initiatives. The first involves real-world evidence generation from real-world data in both the pre- and post-market settings. This initiative reflects the desire to use existing health care data and data from other non-standard sources in lieu of or to supplement data collected as part of a research study to support product approval or expansion to new indications. Patient advocacy groups and medical researchers alike view the rich data streams that are becoming increasingly available as too big to ignore in terms of providing insight into a therapy’s safety and effectiveness. The second initiative reflects a desire to use innovative or otherwise non-traditional trial designs to generate substantial evidence in support of regulatory submissions. This initiative was intended to counter the belief that FDA reviewers were not often accepting of novel or complex trial designs, as shown by the lack of available guidance on the topic or examples that could be pointed to as successes. The third initiative reflects a desire to incorporate pharmacometric modeling into a drug development program to reduce the amount of data required to be collected in new clinical studies. The goal is to more effectively leverage what is already known about a drug, or able to be predicted with a high level of confidence, in planning the later stages of clinical development or expansion to new indications. In this seminar, I will review these initiatives, based on my involvement while at FDA in their development and roll-out, and discuss their impact on clinical research today. I will also discuss some recent initiatives at ASA that dove-tail nicely with the FDA initiatives and relate to opportunities for statisticians to take a leadership role in both finding and advocating for appropriate use of innovative solutions to problems in public health research.