QSU Seminars

The Quantitative Sciences Unit (QSU) hosts a forum to discuss research methods in medicine held on the first Tuesday of every month.  This seminar series is tailored to clinical investigators interested in research methods.

The Research Methods Seminar is an interactive and informal journal club-type format where topical papers in medical research, particularly relevant to faculty in the Department of Medicine, are discussed with an emphasis on the methods and/or study design.

QSU Research Methods Seminar

Location: 1070 Arastradero Road # 109

Time: 4-5pm first Tuesday of the month (unless otherwise noted)

 Refreshments served

Free parking

Upcoming Seminar


Date: Tuesday, November 6, 2018

Charles McCulloch, PhD

Professor and Head, Biostatistics, UCSF 


Title: “Improving the prediction of extreme clusters in multilevel data”


Predicted random effects are widely used to evaluate the performance of and rank clusters such as patients and hospitals using longitudinal and multilevel data.  For example, the Center for Medicare and Medicaid Services uses these approaches to estimate hospital-specific performance metrics such as mortality or re-admission rates.  Using predicted random effects from mixed regression models has been shown to outperform using standard regression to estimate hospital specific effects. However, predicted random effects are often used to identify extreme values such as poorly performing hospitals and their performance has not been systematically evaluated in this context.  After introducing the ideas of mixed models and traditional methods of best prediction, I motivate new methods to predict extreme clusters and evaluate their performance. I show that methods that assume distributions with heavier tails than the normal distribution can produce predicted values with smaller mean square or absolute error of prediction than standard best predicted values when interest focuses on extreme clusters.  The methods are illustrated using data on length of stay in hospitals.


We welcome your suggestions for the seminar series and are happy to answer any questions you may have - please contact Ni Deng (nideng@stanford.edu)

We look forward to seeing you at our next seminar!

Directions to get to 1070 Arastradero:

Driving: Enter the driveway and take the first left into the parking lot and park.  Enter through the front door - you will be in the lobby.  Take the door on the right and follow it toward the back of the building.  Room 109 will be on your left before you reach the glass door leading outside.

Shuttle: 1070 Arastradero is served by the 1050A Marguerite Shuttle which picks up at the Medical Center at the corner of Campus Drive and Welch Road.

Past Seminars


  • October 2018
  • Jonathan Chen, MD, PhD
  • Wisdom of the Crowd or Tyranny of the Mob? Discovering and Distributing the Latent Knowledge Embedded in Clinical Data
  • Auguest 2018
  • Drew Levy, PhD
  • Managing Uncertainty: Evidence Quality and Personalized Health Decision-making
  • May 2018
  • Mithat Gönen, PhD
  • Analysis of Drug Development Portfolios
  • May 2018
  • Dheeraj Raju, PhD
  • An Evaluation of the US Army Nurse Corps Patient CaringTouch System: From Project Inception to Final Report
  • March 2018
  • Shouhao Zhou, PhD
  • Prediction of Longitudinal Biomarker Distributions Using Bayesian Semiparametric Beta Regression
  • April 2018
  • Pranav Samir Rajpurkar, PhD candidate
  • An introduction to deep learning and convolutional neural networks and presentation of CheXNet: Radiologist-level pneumonia detection on chest X-rays
  • March 2018
  • Zihuai He, PhD
  • A semi-supervised approach for Predicting cell type/tissue specific functional consequences of non-coding variation using massively parallel reporter assays
  • March 2018
  • Himel Mallick, PhD
  • The Human Microbiome and Public Health
  • February 2018
  • Summer Han, PhD
  • Novel statistical methods for identifying interactions between genes and environmental exposures for complex diseases
  • January 2018
  • Jeffrey Blume, PhD
  • Second-Generation p-values
  • December, 2017
  • Andrew Gentles, PhD
  • Clinically-relevant interaction networks in the lung cancer tumor microenvironment
  • November, 2017
  • Kristen Cunanan, PhD
  • A Bayesian basket trial design and its statistical properties
  • October, 2017
  • Chiara Sabatii, PhD
  • Testing hypotheses on a tree: new error rates and controlling strategies
  • May, 2017
  • Joseph Rigdon, PhD
  • Improved Development and Validation of Clinical Risk Scores for Benefit and Harm from Intensive Blood Pressure Treatment: A Machine Learning Approach
  • April 2017
  • Trevor Hastie, PhD
  • Statistical Learning with Sparsity
  • Februrary, 2017
  • Susan Holmes, PhD
  • Reproducible research: the challenges of the human microbiome
  • January, 2017
  • Anna Liza Antonio, MS, candidate of DrPH
  • A Hierarchical Bayes Model for Patient Preferences Elicited through Discrete Choice Experiments
  • December, 2016
  • John Whitte, PhD
  • Methods for evaluating the shared genetic basis of common diseases 
  • November 2016
  • Jonathan Taylor, PhD
  • “Selective inference in linear regression”
  • October 2016
  • Michael Baiocchi, PhD
  • "The IMP: Interference Manipulating Permutation"
  • May 2016
  • Andre Valdez, PhD 
  • Using Accelerometry data to study patterns of sleep and performance in blood and marrow transplantation patients
  • April 2016
  • Susan Weber, PhD
  • "Stanford Center for Clinical Informatics Services Overview" 
  • March 2016
  • Andrew Spieker, PhD
  • "Accounting for endogenous medication use when estimating natural biomarker associations using observational data"
  • February 2016
  • Philip Lavori, PhD
  • "A tale of two studies, and 'Stump-the-Chump, the Sequel!' "
  • January 2016
  • Eric Bair, PhD
  • "Random forest importance scores: significance testing and conditional importance"
  • January 2016
  • Andrew Gentles, PhD
  • "The prognostic landscape of genes and infiltrating immune cells across human cancers"
  • December 2015
  • James Dai, PhD
  • "Constraints and orthogonality:  powerful tests for gene-treatment and gene-environment interactions"
  • November 2015
  • Lu Tian, PhD
  • "Exact Inference on the Restricted Mean Survival Time" 
  • October 2015
  • Joseph Rigdon
  • "Exact Confidence Intervals in the Presence of Interference" 
  • April 2015
  • Helena Kraemer
  • “Why Most Published Research Findings are False” Why?
  • March 2015
  • Rita Popat
  • Reliability and Validity Studies: Design and Analytic Approaches
  • Slides for Dr. Rita Popat's presentation can be found here.
  • Benedict Anchang
  • Modeling Cellular Decision Making Processes using Nested Effects Models with Application to Optimizing Drug Combinations
  • Mei-Chiung Shih
  • Innovative Designs of Point of Care Comparative Effectiveness Trials
  • Shanshan Li
  • Recurrent Event Data Analysis With Intermittently Observed Time-Varying Covariates
  • February 2015
  • Armin Schwartzman
  • Image Comparison Problems in Biomedicine
  • Summer Han
  • Statistical methods for genetic associations, gene-environment interactions, and population-level cancer screening
  • Linda Valeri
  • Understanding Racial/Ethnic Disparities In Cancer Survival – A Counterfactual Approach For Conceptual And Missing Data Issues
  • December 2014
  • Phil Lavori
  • If You Need A Statistician To Analyze Your Data, You Should Do A Better Experiment
  • November 2014
  • Hua Tang
  • Challenges and Opportunities in Complex Trait Genetics in Minority Populations
  • October 2014
  • Joseph Rigdon
  • Causal Inference for Binary Data with Interference
  • Nandita Mitra
  • Assessing the sensitivity of cost-effectiveness measures to unmeasured confounding
  • May 2014
  • Amrita Ray
  • Accounting for family relationship errors in genetic studies: focusing on linkage analysis
  • April 2014
  • Sanjay Basu
  • Interpreting Mathematical Simulation Models Applied To Comparative And Cost-Effectiveness Analyses For Population Health Problems
  • March 2014
  • Josephine Asafu-Adjei 
  • Matching and Covariate Adjustment in Discrimination and Bayesian Variable Selection
  • February 2014
    Haley Hedlin
  • An Introduction to Time Series Analysis with Examples from Neuroscience
  • Bruce Swihart
  • Sleep Hypnograms, Insurance Claims, & Hand Movement After Stroke: Big Data and Potentially Many Weak, Predictive Signals
  • Benjamin Goldstein
  • Improving Clinical Decision Making: Three Models for Collaborative Academic Biostatistics
  • Yan Ma
    Improved Statistical Methods for Evidence-based Medicine: Meta-analysis for Multiple Outcomes


December 2013
Steven Goodman
What do P-values mean? Probably not what you think

November 2013
Nigam Shah
Generating Practice-based Evidence from Electronic Health Records

September 2013
David Rogosa
Statistical Methods for Longitudinal Research

May 2013
Aya Mitani
Multiple Imputation in Practice -- Approaches for handling categorical and interaction variables

April 2013
Kristin Sainani
Writing about Biostatistics

March 2013
Iryna V. Lobach
Analysis of Gene-Environment Interactions with Measurement Errors in Environmental Exposures

February 2013
Ying Lu
Statistical Designs for Phase I Cancer Clinical Trials

December 2012
Sergio Bacallado
An Introduction to Bayesian Analysis Using Case Studies in Medical Research

November 2012
Kristin Sainani
Introduction to Propensity Scores

October 2012
John Ioannidis
Genetic Prediction Models: Practice, Metrics and a Discovery Extension

May 2012 
Ben Goldstein
Predicting Acute Sudden Cardiac Death using Electronic Health Records
April 2012 
Sepideh Modrek
An Application of Instrumental Variables: Maternal Education as a Driver for Eliminating Female Circumcision
March 2012
Hui Wang
Applications of Targeted MLE Based Variable Importance Measurement in Dimension Reduction with Gene Expression Data
February 2012
Mike Baiocchi
Estimating the Effectiveness of Intensity of Care on Rates of Death for Premature Infants
January 2012
Raúl Aguilar
Things You Can Do When You Have Missing Covariates
December 2011
David Shilane
Comparative Effectiveness Research in Cardiology with Messy Data
November 2011
Ben Goldstein
Prediction in Medical Studies: What, Why & How
October 2011
Jane Paik
Using Regression Models to Analyze Randomized Trials: Robustness of Survival Models to Misspecification
September 2011
David Rehkopf
Applying Machine Learning Algorithm to Answer Questions from Observational Data: Essential Complement or Dangerous Tool?
June 2011
Susan Gruber
Targeted Maximum Likelihood Estimation for Causal Inference
May 2011
Gunnar Carlsson
Topological Data Analysis for Biology
April 2011
Maria E Montez-Rath
Methods for Handling Survey Data
March 2011
Mark Cullen, Manisha Desai, Jessica Kubo
Modeling the Hazard of Injury as a Function of Experience Among Hourly Aluminum Manufacturing Workers
February 2011
Jose Montoya

Could a Recently Found Virus, XMRV, Cause Chronic Fatigue Syndrome?
January 2011
Manisha Desai
An Introduction to Missing Data and Imputation Methods
December 2010
Wolfgang Winkelmayer
Propensity Scores
November 2010
Jay Bhattacharya
Does Swan-Ganz Catheterization Increase Mortality in the ICU? An Instrumental Variables Bounding Approach
October 2010
Tim Assimes

Epidemiological Issues in Contemporary Human Genetic Studies

September 2010
Doug Owens
Where Angels Dare not Tread: Development of a Guideline for Screening Mammography in 40 to 49 year Old Women