A really useful and very accessible guide to issues of overdispersion in funnel plots with recommendations for action.  Illustrates how flexible a tool funnel plots can be!

David Speigelhalter, Winton Professor of the Public Understanding of Risk in the Statistical Laboratory at the University of Cambridge writes on how funnel plots can be flexible, attractively simple, and avoid spurious ranking of institutions into ‘league tables’.  Contains technical approaches to dealing with overdispersion.

A review of how understanding variation can help to identify and potentially prevent serious issues, using six case studies including the Shipman case and childrens cardiac surgery at Bristol.

A frequent problem in quality improvement is that people waste time collecting too much data.  In this paper, our goal is to offer some general guidelines to measuring fidelity of implementation on small samples in the face of constrained measurement resources.

The paper presents two experiments that compare lay people’s use of control and non-control charts to make decisions between hospitals (Funnel charts vs League tables) and to monitor changes across time (Run charts with control lines vs Run charts without control lines). As expected, participants more accurately identified the outlying data using a control chart than a non-control chart, but their ability to then apply that information to more complicated questions (e.g., where should I go for  treatment?, and should I investigate?) was limited. The discussion highlights some common concerns about using control charts in hospital settings.

Adapted from the IHI, the 'stages of data acceptance' provides a view on how teams may (or may not!) engage with data and give some understanding of the stages people pass through.  

Healthcare managers, clinical researchers and individual patients (and their physicians) manage variation differently to achieve different ends.  This paper outlines different approaches to understanding and using variation in different contexts and considers what a 'gold standard' approach might look like.