Statistical Thinking

This blog is devoted to statistical thinking and its impact on science and everyday life. Emphasis is given to maximizing the use of information, avoiding statistical pitfalls, describing problems caused by the frequentist approach to statistical inference, describing advantages of Bayesian and likelihood methods, and discussing intended and unintended differences between statistics and data science. I’ll also cover regression modeling strategies, clinical trials, drug evaluation, medical diagnosis, and decision making.

Frank Harrell

Frank Harrell

Professor of Biostatistics

Vanderbilt University

Biography

Frank Harrell is a Professor of Biostatistics in the School of Medicine at Vanderbilt Unversity. His research interests include statistical modeling, predictive models and model validation, Bayesian statistics, Bayesian clinical trial design, clinical trial design, analysis, and reporting, statistical computing, statistical graphics, reproducible research, drug development, medical decision making and diagnostic research, health services research, cardiology, COVID-19 therapeutic clinical trial design, and teaching.

Interests

  • Statistical modeling
  • Bayesian statistics and study design
  • Predictive modeling and model validation
  • Biomedical research
  • Medical decision making
  • Medical diagnostic research
  • Statistical computing, graphics, and reproducible research methods
  • Cardiology
  • Clinical trial design, analysis, and reporting
  • COVID-19 therapeutic clinical trials
  • Health services research
  • Drug development

Education

  • PhD in Biostatistics, 1979

    University of North Carolina

  • BS in Mathematics, 1973

    University of Alabama in Birmingham

Skills

Biostatistics

R

Biomedical Research

Teaching

This is a free web course in introductory and intermediate biostatistics. Details are on the course web page.

Regression Modeling Strategies

I teach the BIOS7330 Regression Modeling Strategies course in the Biostatistics Graduate Program at Vanderbilt University in the spring semester. The course web page is here. I teach a 4-day virtual version of this course each May. Registration information for the short course may be found here.

Master of Science in Clinical Investigation Biostatistics II

I co-teach this course at Vanderbilt each February for postdoctoral medical and surgical fellows and junior faculty in the MSCI program.

Recent Posts

If You Like the Wilcoxon Test You Must Like the Proportional Odds Model

Since the Wilcoxon test is a special case of the proportional odds (PO) model, if one likes the Wilcoxon test, one must like the PO model. This is made more convincing by showing examples of how one may accurately compute the Wilcoxon statistic from the PO model’s odds ratio.
Last updated on 2021-04-13 7 min read

Implementation of the PATH Statement

The recent PATH (Predictive Approaches to Treatment effect Heterogeneity) Statement outlines principles, criteria, and key considerations for applying predictive approaches to clinical trials to provide patient-centered evidence in support of decision making. Here challenges in implementing the PATH Statement are addressed with the GUSTO-I trial as a case study.
Last updated on 2020-11-28 18 min read

Violation of Proportional Odds is Not Fatal

Many researchers worry about violations of the proportional hazards assumption when comparing treatments in a randomized study. Besides the fact that this frequently makes them turn to a much worse approach, the harm done by violations of the proportional odds assumption usually do not prevent the proportional odds model from providing a reasonable treatment effect assessment.
Last updated on 2021-04-13 12 min read

Unadjusted Odds Ratios are Conditional

This article discusses issues with unadjusted effect ratios such as odds ratios and hazard ratios, showing a simple example of non-generalizability of unadjusted odds ratios.
Last updated on 2020-09-18 9 min read

RCT Analyses With Covariate Adjustment

This article summarizes arguments for the claim that the primary analysis of treatment effect in a RCT should be with adjustment for baseline covariates. It reiterates some findings and statements from classic papers, with illustration on the GUSTO-I trial.
Last updated on 2020-09-15 15 min read
See all posts

Projects

Recent & Upcoming Talks

Sequential Bayesian Designs for Rapid Learning in COVID-19 Clinical Trials

Continuous learning from data and computation of probabilities that are directly applicable to decision making in the face of …
2020-11-25 00:00
Code Slides Video

Musings on Statistical Models vs. Machine Learning in Health Research

Video
Health researchers and practicing clinicians are with increasing frequency hearing about machine learning (ML) and artificial …
2020-11-23 00:00
Slides Video

Controlling α vs. Probability of a Decision Error

This presentation clarifies what type I assertion probability α protects against, by making a clear distinction between how often we …
2020-11-12 00:00
Slides

Ways to Make Randomized Clinical Trials More Statistically Efficient

This presentation covers several ways to make clinical trials more efficient and to reduce the chance of ending with an equivocal …
2020-10-23 00:00
Slides Video

Controversies in Predictive Modeling, Machine Learning, and Validation

This talk covers a variety of controversial and/or current issues related to statistical modeling and prediction research. Some of the …
2020-09-25 00:00
Code Slides Video
See all talks

Topics

accuracy-score backward-probability bayes big-data bioinformatics biomarker bootstrap change-scores classification collaboration conditioning covid-19 data-reduction data-science decision-making design diagnosis dichotomization drug-evaluation EHR endpoints evidence forward-probability generalizability graphics hypothesis-testing inductive-reasoning inference judgment likelihood logic machine-learning measurement medical-literature medicine metrics multiplicity observational p-value personalized-medicine posterior precision prediction principles prior probability r RCT regression reporting reproducible responder-analysis sample-size sensitivity sequential specificity subgroup survival-analysis teaching validation variability 2017 2018 2019 2020 2021

Contact

Datamethods

datamethods.org is a discussion site where data methodologists meet each other and subject matter experts including clinical trialists and clinical researchers. Its development is documented here. Datamethods is provided by the Department of Biostatistics, Vanderbilt University School of Medicine.

I have written some short articles on the site, listed below.