crc_review.md

CHAPMAN & HALL/CRC STATISTICS BOOK PROPOSAL

TITLE AND AUTHOR/EDITOR(S)

1. Provisional title of your book.

   • The Hitchhiker's Guide to the tlverse: A Targeted Learning Practitioner's Handbook

2. Names, titles, affiliations, and email addresses for all authors/editors

   • Mark van der Laan, UC Berkeley, laan@berkeley.edu
   • Alan Hubbard, UC Berkeley, hubbard@berkeley.edu
   • Jeremy Coyle, UC Berkeley, jeremy.coyle@gmail.com
   • Nima Hejazi, UC Berkeley, nhejazi@berkeley.edu
   • Ivana Malenica, UC Berkeley, imalenica@berkeley.edu
   • Rachael Phillips, UC Berkeley, rachaelvphillips@berkeley.edu

CONTENTS

3. Please include a complete table of contents including chapter and section headings.

   • The Roadmap for Targeted Learning

     - Learning Objectives
     - Introduction
     - The Roadmap
     - Summary of the Roadmap
     - Causal Target Parameters
     

   • Welcome to the tlverse

     - Learning Objectives
     - What is the `tlverse`?
     - `tlverse` Components
     - Installation
     

   • Example Datasets and Case Studies

     - WASH Benefits Example Dataset
     - International Stroke Trial Example Dataset
     - Veterans' Administration Lung Cancer Trial Dataset
     

   • Cross-validation

     - Learning Objectives?
     - Why Split Your Sample?
     - Cross-validating arbitrary models
     - Exercises
     

   • Super (Machine) Learning

     - Learning Objectives
     - Motivation
     - Introduction
     - `sl3` "Microwave Dinner" Implementation
     - Cross-validated Super Learner
     - Variable Importance Measures with `sl3`
     - Exercises
     - Concluding Remarks
     - Appendix
     

   • The TMLE Framework

     - Learning Objectives
     - Introduction
     - Easy-Bake Example: `tmle3` for ATE
     - `tmle3` Components
     - Fitting `tmle3` with Multiple Parameters
     - Exercises
     - Summary
     

   • Optimal Individualized Treatments Regimes

     - Learning Objectives
     - Introduction to Optimal Individualized Interventions
     - Data Structure and Notation
     - Defining the Causal Effect of an Optimal Individualized Intervention
     - Interpreting the Causal Effect of Optimal Individualized Interventions
     - Evaluating the Causal Effect of an OIT with Binary Treatment
     - Evaluating the Causal Effect of an Optimal ITR with Categorical
          Treatment
     - Extensions to Causal Effect of an OIT
     - Variable Importance Analysis with OIT
     - Real-World Data and `tmle3mopttx`
     - Exercises
     

   • Stochastic Treatment Regimes

     - Learning Objectives
     - Introduction to Stochastic Interventions
     - Data Structure and Notation
     - Defining the Causal Effect of a Stochastic Intervention
     - Interpreting the Causal Effect of a Stochastic Intervention
     - Evaluating the Causal Effect of a Stochastic Intervention
     - Extensions: Variable Importance Analysis with Stochastic Interventions
     - Exercises
     

   • Appendix: A Primer on the R6 Class System

     - Classes, Fields, and Methods
     - Object Oriented Programming: Python and R
     

SUBJECT/AUDIENCE

4. Please describe in detail the subject of your book. Why will this book be important, who will find it useful, and what is new? What background will you assume and can you name any books that will suffice?

The Hitchhiker's Guide to the tlverse: A Targeted Learning Practitioner's Handbook is an open-source and fully-reproducible handbook for applying the Targeted Learning framework and statistical methodology in practice using the tlverse software ecosystem (https://github.com/tlverse). The materials have been developed, tested, and refined across a series of short courses focused on the applications of Targeted Learning, all delivered at major academic conferences.

The contents of this handbook are meant to serve as a reference guide for applying the Targeted Learning methodology in research settings, throughout the biomedical, health, and social sciences. The book will empower all manner of researchers to use this state-of-the-art methodology in rigorously developing statistical answers to scientific questions grounded in causal inference. Each section introduces a set of distinct causal questions, each motivated by a set of case studies, alongside statistical methodology and software for assessing the causal claims of interest.

Necessary background for engaging with the handbook includes intermediate knowledge of statistical concepts and methodology (e.g., null hypotheses, hypothesis testing, confidence intervals), a basic working knowledge of statistical estimation and machine learning (e.g., what is a linear model?, what are prediction algorithms?), and some background in R programming. The handbook is intended to be self-contained -- as such, much background in causal inference and Targeted Learning are included. Introductory reference material will be provided in an appendix.

5. Will your book be primarily a reference/monograph or textbook? If a textbook, for which courses will it be the primary text and at what level is the course taught? Will you include exercises sets and supply a solutions manual?

The book is intended as a reference, aiming to support researchers in using the Targeted Learning framework to develop rigorous answers to questions rooted in causal inference and arising in a vast array of application areas. Principally, the handbook aims to provide a gentle but thorough introduction to the Targeted Learning methodology, demonstrating its use in case studies rather than delving into theoretical detail at the level of existing reference texts (see below) and relevant primary literature. Secondarily, the handbook also serves as long-form documentation of the rich set of tools available in the tlverse software ecosystem. While not intended as a textbook, this work has been developed and refined through teaching a series of workshops; to aid in this, a limited set of exercises were developed (solutions to exercises will be made available online and possibly in an appendix).

6. What related books are available and what advantages will your book have?

   • 2011, Targeted Learning: Causal Inference for Observational and Experimental Data (Mark van der Laan and Sherri Rose)
   • 2018, Targeted Learning in Data Science: Causal Inference for Complex Longitudinal Studies (Mark van der Laan and Sherri Rose)

Th work is focus not on providing in-depth technical descriptions or in communicating the most recent developments in state-of-the-art statistical methodology, in stark contrast to the two existing books on Targeted Learning. Instead, the goal is to convey both the underlying unifying concepts and key details of the broad set of state-of-the-art Targeted Learning techniques in a manner that is both clear and complete, without burdening the reader with extraneous technical details. We aim for the book to serve as a comprehensive reference and handbook for researchers -- methodologists and domain specialists alike -- who wish to employ the central statistical tools of the Targeted Learning framework efficiently. The material covered is a mix of key ideas in causal inference, machine learning, and non/semi-parametric statistical theory, all presented in the context of developing solutions to complex and carefully considered real-world data analysis questions. Throughout the text, all aspects of the methodology are demonstrated using the accompanying free and open source tlverse software ecosystem (https://github.com/tlverse).

PRODUCTION

7. Approximately how many printed pages will your book contain? Are color figures essential to your book? If so, about how many would have to be in color? Color printing is still very expensive and color figures will increase the price so black and white should be used unless color is essential.

   • Approximate page count: 200-250 pages.
   • Color figures are not essential to our book.

8. When would you hope to be able to submit the final draft of the book to us? Will you use LaTeX, bookdown, or Word? We will supply a style file for LaTeX authors and request an unformatted file from Word authors.

   • A final draft of the book will be available by 01 January 2021.
   • Our book has been prepared using the bookdown R package.

REVIEWS

9. Please give the names and e-mail addresses of four people who would be qualified to give an opinion on your proposed book.

   • Ashley Naimi, University of Pittsburgh
   • Linda Valeri, Columbia University
   • Constantin Frangakis, Johns Hopkins University
   • Michael Hudgens, University of North Carolina
   • Peter Gilbert, Fred Hutchinson Cancer Research Center

MARKETING

10. If your book is aimed at a professional/research market, are there key societies outside of statistics to which it should be marketed? (All major statistical societies will be covered.)

Society for Epidemiologic Research

11. Please list up to six key features of your proposed book that we can use in bulleted form.

    • Learn to leverage the powerful Targeted Learning framework and methodology without a mathematical deep-dive
    • Introduces both classical and cutting-edge topics in causal inference and non/semi-parametric statistics
    • Explores predictive and prescriptive data analytics with the tlverse software ecosystem
    • See Targeted Learning in action by engaging with real-world case studies and hands-on examples

12. Please list up to six key words or phrases that people interested in this topic may use to search Amazon or the web. Do not repeat words in the title as these will already be found.

    • Causal inference, machine learning, statistical computing, biostatistics, data science, predictive/prescriptive analytics

13. Please select the three most important markets for your book. Other categories are available including education, psychology, and economics so please mention other important disciplines.

Selected market sections in bold below.

STA01A-Statistics-Statistics for Life Sciences

STA02A-Statistics-Introductory Statistics & General References

STA04A-Statistics-Probability Theory & Applications

STA06A-Statistics-SPC/Reliability/Quality Control

STA07A-Statistics-Statistical Theory & Methods

STA07J-Statistics-Statistical Learning & Data Mining

STA08A-Statistics-Computational Statistics

STA09A-Statistics-Statistics for Business, Finance & Economics

STA10A-Statistics-Statistics for Engineering and Physical Science

STA11A-Statistics-Biostatistics and Epidemiology

STA12A-Statistics-Statistics for the Social and Behavioral Sciences

STA13A-Statistics-Environmental Statistics

STA14A-Statistics-Statistical Genetics & Bioinformatics