Analytics in healthcare: bringing A&E performance targets into sharper focus
05 February 2018
Over the past few weeks, it’s been impossible to read, watch or hear any UK news bulletin without at least a mention of the NHS’s struggles to hit the 4-hour standard for A&E performance.
On the face of it, the target seems narrow and simple. Achieving it, however, is anything but straightforward. That’s because its apparent simplicity belies a highly complex array of activities that all have to be working well across the system. From A&E triage and streaming to admissions, flow through inpatient wards, discharging processes and even to community-based care, if one or more of those are not working well, there will almost certainly be a knock-on impact on the rest. The result is that the four-hour target will be missed.
And in recent years we’re seeing performance continue to decline. In London over the most recent Christmas period, for example, no trusts were able to achieve the 4-hour target. And the situation is similar across England. So, what is to blame? The answers to that question depend on the views of the respondent. One side points to a lack of funding, resources and people as the chief culprits. Others argue that inexorably rising demand from an ageing population and the increase in chronic diseases are placing almost impossible demands on the system. Some would suggest the NHS needs to get a grip and make better use of the resources it has.
Whatever the ultimate reasons for trusts failing to achieve the target, it’s clear that many different external factors play a role. But what is interesting is that the absolute number of patients presenting at an emergency department is not a very important driver of performance. In fact, most emergency departments cope quite well with higher patient volumes. However, one factor that does have a major impact on performance is the acuity of the patients. More ambulances, sicker, older patients requiring admission will all present greater challenges to A&E departments’ ability to cope.
But it is not easy to tell which factors in particular are having the greatest impact. If an A&E department manages relatively well under considerable pressure one day and rather less well under more benign circumstances, it’s traditionally been hard to unpick which drivers made things better or worse on a particular day. Instead, anecdotal rather than factual evidence tends to be used to excuse or explain performance. That makes it hard to identify whether hospitals are doing well or not.
To help hospitals understand and therefore manage their performance better, we’ve been developing innovative analytical tools. These enable hospital teams to predict with a high degree of accuracy whether individual patients presenting at A&E will require admission and if they will breach the four-hour standard.
The model takes into account a large data set covering a wide variety of drivers and delivers much more definitive judgements of how well the department performed against the demand conditions they faced. This cuts through the anecdotes, increases transparency and accountability and can help create a rapid learning cycle that leads to improvement.
Today, we are using these predictive models alongside behavioural and cultural change programmes with a small number of trusts and achieving measurable improvements. For example, we’ve recorded a five percent better performance than would have been expected (taking in to account the type of demand faced) at one of the largest and most complex sites in the country.
What this will allow us to do is help uncover the drivers of performance and celebrate success where it is due but also highlight underperformance where necessary. And with the challenges that the NHS will continue to face, creating hard evidence to explain and improve performance has never been more important.