Chapman and Hall/CRC
362 pages | 101 B/W Illus.
Self-Controlled Case Series Studies: A Modelling Guide with R provides the first comprehensive account of the self-controlled case series (SCCS) method, a statistical technique for investigating associations between outcome events and time-varying exposures. The method only requires information from individuals who have experienced the event of interest, and automatically controls for multiplicative time-invariant confounders, even when these are unmeasured or unknown. It is increasingly being used in epidemiology, most frequently to study the safety of vaccines and pharmaceutical drugs.
Key features of the book include:
The book is aimed at a broad range of readers, including epidemiologists and medical statisticians who wish to use the SCCS method, and also researchers with an interest in statistical methodology. The three authors have been closely involved with the inception, development, popularisation and programming of the SCCS method.
"As an epidemiologist who has worked for some 40 years in the field of vaccine safety I have first-hand experience of the huge impact that the innovative self –controlled case series method developed in the mid-1990s by Professor Paddy Farrington has had in this field. The SCCS method enabled researchers such as me to rapidly conduct robust studies to assess whether there was an increased risk of a particular adverse event after a vaccine without the need to use the more laborious and bias-prone case control and cohort method methods. The SCCS has wider application than just studies of vaccine safety and is now an essential component of the methodological tool kit available to epidemiologists interested in assessing risks, particularly those that are short term and potentially related to an intervention such as vaccination, environmental exposure such as an infection or life style activity such as strong exercise. This well-written book should be on the shelves of all public health institutions and required reading for any aspiring statistician."
—Professor Elizabeth Miller, Public Health England
"'Self-controlled Case Series Studies: A Modelling Guide with R’ is essential reading for anyone who has ever used or is contemplating using the SCCS design in their research. This informative book provides a clear and concise commentary on the SCCS method, its initial use in vaccine safety studies and its eventual foray into medication safety studies. The text delves into the importance of the assumptions of the method and critically, describes what you can do when those assumptions are not met. The inclusion of elegant R code throughout and description of the SSCS package provides everything researchers need to know to implement their own studies, check their assumptions and interpret their results. This book will help to solidify researchers understanding of the SCCS method and will surely motivate further methodological developments to generate robust evidence of the safety of medicines and vaccines in the real world."
—Nicole Pratt, Associate Professor, University of South Australia
"The self-controlled case series has emerged as a key methodology for studying the effects of healthcare interventions. Professor Farrington and his colleagues introduced the self-controlled case series and have led the way in developing an impressive array of extensions. The overall literature around the self-controlled case series has exploded in recent years and this important and timely book pulls it all together in an effective and clear manner. The book includes extensive practical examples with R code and would be an ideal text for a master’s or doctoral-level course in a statistics or epidemiology program. A particular strength of the book is its relentless focus on model assumptions and model checking. It certainly belongs on the shelf (or beside the keyboard) of every analyst conducting observational studies in healthcare."
—David Madigan, Columbia University
"The self-controlled case series method is an important approach to overcoming between-person confounding that is increasingly widely used to assess the effects of health related exposures. This book provides an invaluable practical guide to implementing the approach in R. As well as detailed guidance, the book covers the assumptions and limitations of the approach, and will thus help ensure the validity of resulting analyses. I thoroughly recommend it to anyone planning to use the case series approach."
—Liam Smeeth, Professor of Clinical Epidemiology, London School of Hygiene and Tropical Medicine
"The Self-controlled case series method is an increasingly popular analysis method in modern epidemiological research. This approach is particularly useful when time invariant confounding is difficult to capture as is very often the case in research using electronic health records. This book is written by the team that invented and pioneered the method and it is a comprehensive guide to its use. The book is aimed at both statisticians and epidemiologists. All the essential concepts are presented at a level understandable to the non-statistician while further statistical detail is provided in optional starred sections that can be skipped without loss of continuity. The authors cover both the theoretical underpinnings of the method and also the practical aspects of conducting an analysis in R using an accompanying software package. They also detail a range of methods to check that the method is valid for an individual study and some extensions to the method that can deal with some instances where the standard analysis is not valid. In short, this is an essential reference for any epidemiologist interested in using the self-controlled case series method."
—Adrian Root, Academic General Practitioner, London School of Hygiene & Tropical Medicine
"The self-controlled case series (SCCS) is one of the most important self-controlled designs in observational research, but has often been misunderstood and even misapplied. This book, written by the team that has led the development of the SCCS from the beginning, is an essential guide to all aspects related to this method. It leaves no stone unturned when it comes to the current state of the art of the SCCS method, and even includes developments that have not been published elsewhere. For epidemiologists, the book describes the rationale behind the SCCS, and the requirements that need to be met for appropriate use of the method. For statisticians, every detail of the statistics behind the SCCS is included (luckily in clearly marked sections that are optional for those seeking a higher-level understanding of the method). I especially appreciate the fact that each bit of theory is accompanied by several real-world examples, complete with functional R code."
—Dr. Martijn Schuemie, Population-Level Estimation Workgroup leader, Observational Health Data Science and Informatics (OHDSI)
"This is a much-anticipated first book on the self-controlled case series (SCCS) method. Written by the originator of the method and its main developers, it covers the basic principles of the design before going on to show how it can be adapted to assess associations in a wide range of case-only studies. The book provides rigorous detail of the method, its likelihood and properties, whilst frequently applying it to data based on real examples including vaccine and drug safety, as well as environmental exposures. Of particular value is the accompanying R code, which will allow readers to fit and plot a wide range of SCCS models, many of which were previously unavailable in standard statistical software. This is the SCCS "user manual" and will provide new insights to those familiar with the method as well as a starting point for those wishing to learn about it and apply it in their field."
—Professor Nick Andrews, Public Health England
"The authors have written the first authoritative book on SCCS studies that will be useful to a wide spectrum of audiences. The diversity of data analysis examples throughout the text together with the R package will make this book a great study and reference text for applied researchers and graduate students in statistics,biostatistics, and epidemiology. The carefully organized sections on the technical formulation of model development, estimation and inference procedures serve as important foundational materials for stimulating further research in SCCS studies forfuture generations of researchers."
—Danh V. Nguyen, University of California, Irvine
The SCCS likelihood
The standard SCCS model
Checking model assumptions
Further SCCS models
Extensions of the SCCS model
Design and presentation of SCCS studies