Book Review: Covid by Numbers

Most people have their own personal stories to tell about the COVID-19 pandemic. The first encounter with the new virus, the experience of lockdowns (or lack hereof), (not) getting vaccinated, etc. We all have our own unique view on life during the pandemic. However, at the core of the pandemic was data. Statistics, numbers, graphs, and tables, and more importantly, discussions about data. Should we use a logarithmic scale? How should we make sense of the reproduction number? Did numbers on cases, hospital admissions, and deaths reflect true numbers? Etc.

We will see a lot of books in the years to come on the pandemic, also from a data perspective. The new book, Covid By Numbers: Making Sense of the Pandemic with Data, written by David Spiegelhalter and Anthony Masters, delivers on what the title promises. It makes sense of the pandemic with data. Most of the questions you can ask related to the first years of the pandemic, where data is relevant, are covered by this book. If you need a primer on the pandemic and only feel like reading one book, this might be the one.

The book is about the pandemic. I have lived through the pandemic. The book is about data. I breath data. The book focuses on the United Kingdom. I live in the United Kingdom. Ironically, for those exact reasons, the book is not for me. While there is information that is new to me, it is more of a trip down memory lane than anything else. And I am not sure I really need that at this point in time. However, if you didn’t bother to follow any of the data discussions when they took place during the pandemic, or if you are not a fan of statistics but would like to learn more about the pandemic, this book is for you.

There are some good general points in the book. Data cannot – nor should it – speak for itself. Data can help us answer questions, but they often ask even more questions than they answer. The book does a good job in discussing the strenghts and limitations of data, including how to deal with conflicting evidence. Accordingly, the authors are clearly familiar with the key issues and debates when using data to shed light on the pandemic, and it shows.

The chapters are very brief. There are 27 chapters in total (+ a postscript). Most of these chapters are structured as questions, such as “What is the risk from new variants?” and “How many people have been infected with SARS-CoV-2?”. Each chapter is then build up around a series of questions that are briefly touched upon. Chapter 9, for example, is titled “What happened in hospitals?”, and answers the questions “How many people have gone to hospital with Covid-19?”, “How did the NHS cope?”, “What happened to people in hospital with Covid-19?”, “What happened to people after discharge?”, and “How many caught Covid-19 in hospital?”.

The problem with this approach is that there are simply too many questions and too little space to do all the questions justice. Accordingly, the book comes of as being a collection of encyclopedia entries, FAQs, blog posts and textbook material, rather than a coherent book that will stand the test of time. Or, to give it a positive spin, this is a book that is very easy to read, especially if you are looking for some bedtime reading.

There is another problem with a timely book. For how long will the answers to the questions be relevant? Consider questions such as “How ill do people get with Covid-19?”, “What symptoms do people experience?”, “How effective are the vaccines?”, “What protection does a single dose provide?”, “What is the average risk of dying, for people who get infected with Covid-19?”. The answers to these questions are shaped by when the authors finished editing the book (June 2021). The further away we get from the publication date, the less relevant it is. The problem is that, when the chapters and sections are so brief, they mostly provide a few key graphs and some interpretations with no general message.

For example, consider Chapter 11, “How many people have died from Covid-19?”. It is interesting to deal with the different approaches to answer this question, including the numerous estimation strategies. However, the chapter only briefly touches upon a series of questions before moving on to the next chapter. Much more attention and detail to some of these questions would have made the book a lot more interesting.

My recommendation is to pick up another book by David Spiegelhalter, namely The Art of Statistics: How to Learn from Data. This book does a much better job at introducing a series of topics related to data and statistics that can easily be applied to the COVID-19 pandemic. In other words, instead of reading Covid by Numbers, go read The Art of Statistics (if you haven’t already done so).

If you care about data, the pandemic and in particular from a British perspective, and any of the above points have not scared you away, then this book is for you. Ironically, I believe the people who should read this book are exactly the people who are the least likely to pick it up (and the people that would never read this post).