Methodological issues relating to case study site selection for a study of emergency department waiting

Abstract This paper examines the purposive selection of case study sites in health services and medical research. It describes different data that can inform case selection, and considers some of the challenges encountered when choosing cases for a study of waiting in English NHS emergency departments (EDs). Case study is a well-established naturalistic research design that can generate a rich, multi-faceted understanding of complex healthcare environments. While the method does not seek to produce statistically predictive results it does aim to produce findings that are transferable to other settings and contexts. The methods literature outlines a number of ways of selecting cases, all broadly purposive, but little practical detail is provided on using data to inform selection. Research papers typically report that purposive sampling was used, and sometimes that ‘maximum variation’, or typical or unique cases were sought. Few reports provide detail about how data or other evidence was used to inform case selection. This paper describes the different kinds of publicly available statistics and information, and local health service data which is available only to researchers, that can inform case selection for a study of a health service or facility. We examine the different understandings these different data provide and discuss the particular challenges these pose when the unit of analysis is an ED. Throughout the paper, we illustrate our arguments by focusing on data available for the area of Luton, England and the ED associated with this area. We hope to address the gap in the methods literature about case study sampling, encourage more discussion of case selection decisions and indicate the particular problems when sampling EDs for case study research.