In my work as a clinician (surgeon) I have used time-series visualisation a lot in the process of improving the flow of my service (e.g. reduced waiting times) and I assumed that the venerable XmR chart would work well enough for whatever data I was plotting.  I subsequently discovered that was not the case and that prompted me to dig deeper into the history and design of the XmR chart.  What I discovered is that it was designed for presenting quality metrics, which is necessary, but is less well suited to presenting flow metrics and to diagnosing the causes of problems like long waiting times (A&E, GP access, urgent suspected cancer, elective care etc).  The problems that arise are like the ones we see in diagnostic tests – false positives and false negatives.  For example, the XmR chart is insensitive to small shifts and drifts in flows (i.e. demand and activity) and fail to alert us (a false negative) early enough to avert a catastrophe.  Small differences in flows over a long period lead to cumulative effects that “suddenly” catch us unawares – especially when we are only monitoring failures (e.g. maximum waiting time targets).

So, some years ago I grasped this nettle and set about designing an improved way to use time-series charts to help us diagnose the root causes of the much more complex behaviour of a health care system.  This project has led to the Vitals Chart which has now been thoroughly verified and validated and I’ve attached one of the published case studies which is written by Lesley Wright who has been a champion of flow improvement in the NHS since the start of the Modernisation Agency.

Would anyone on the group be interested in exploring how these linked system behaviour charts (SBC) might be more widely used in health care improvement?  And I mean on all dimensions: safety, flow, quality and affordability because we all need wins on all four dimensions at the same time!