15 - Observational Investigations from England: The CESDI and SWISS Studies

Summary

Introduction

Emerging evidence in the 1980s of a link between infants positioned to sleep on their front (prone) and sudden infant death syndrome (SIDS) eventually led to a national “Back to Sleep” campaign in England and Wales in November 1991 (1-6). The SIDS rate dramatically fell by two-thirds (67%) in just four years, from a peak of 1,597 SIDS deaths in 1988 to 531 deaths by 1992 (7). To monitor the characteristic profile of this reduced number of deaths, and identify further potential risk factors associated with SIDS, a case-control study was commissioned as part of a Confidential Enquiry into Stillbirths and Deaths in Infancy (CESDI) and conducted via the University of Bristol between 1993 and 1996 across a third of England (8, 9). A similar study in a smaller geographical area, the South West Infant Sleep Scene Study (SWISS), collecting additional details surrounding the infant sleeping environment, was conducted 10 years later between 2003 and 2006 by the same team (10). Both studies have been instrumental in providing the evidence base for SIDS risk reduction campaigns worldwide and reducing these deaths nationally by a further 60% to 212 SIDS deaths in 2014. This is a review of the main findings, the changes in risk profile over time, and what can be learnt from combining the data from these two studies.

The Optimal Study Design

Despite SIDS being one of the leading causes of post-neonatal infant death, it is a relatively rare event and thus the optimal study design is the observational case-control study. With this design, rare events can be captured as they happen (cases) and compared to suitable controls, although it is often difficult to establish whether significant associations are causal or whether the observations collected are subject to bias. With SIDS investigations in particular, the importance of the final sleeping environment and lack of an immediate causal explanation means recall and misclassification bias can be problematic.