Can doing the right thing give the wrong impression?

If you are using readmission rates to measure hospital performance the answer is 'maybe'...

A recent news headline caught my eye: ‘Scandal of NHS 'production line' as readmissions soar’. The article quotes figures showing that emergency readmissions to hospital have risen by more than three quarters in the last decade. It suggests thousands of patients are being hurried through the system, sent home before they are well enough in order to free up beds, only to be readmitted later on.

But a research project I've been involved with asks whether there’s another way of looking at this data. Could it be that hospitals that work harder to keep frail patients alive are being penalised for their higher readmission rates, whereas others providing lower standards of care look better on paper?

Well, it’s a fair assumption that upon admission to hospital the primary objective of staff is to minimise mortality amongst their patients. A higher quality hospital that is better at keeping patients alive will demonstrate a low (adjusted) mortality rate, therefore, being deemed to have performed well.

However, in keeping patients alive the high quality hospital will simultaneously have created a frailer local population of surviving patients who are more likely to be readmitted to hospital. The high quality hospital may then demonstrate a high readmission rate, thus indicating poor performance. Performance success in one quality measure – such as low mortality – can, therefore, make it much harder to show success in another quality measure, such as readmission rate.

There is, therefore, a potentially significant failing when using readmission data in isolation to assess hospital quality. Ignoring the connection between readmission rates and mortality can lead to misleading judgements about hospital performance.

We’ve been supporting academics from Imperial College, London to research this phenomenon in the NHS. Using routine administrative data over a five year period on fractured hip admissions for 290,000 NHS patients aged 65 and over, the researchers found evidence that, on average, hospitals that perform well on mortality rates performed less well on readmissions.

To me, these findings seem particularly important because, as the article shows, the NHS is increasingly using hospital readmission rates to assess performance. With financial penalties and other sanctions attached to poor performance, this could put some hospitals under unwarranted financial pressure at a time when budgets are already stretched to the max.

It’s difficult to argue against making hospitals accountable for their readmissions. But it needs to be done fairly so that high performing hospitals are not unjustifiably penalised for doing what many of us would say they are primarily there to do: that is demonstrate low mortality rates, or in everyday speak, save lives.

What’s really interesting is Imperial’s proposition that much of the recent rise in readmissions may actually have been due to improved quality, in the form of an increasing likelihood of survival, rather than a general rise in inadequate care on first admission.

It will be interesting to see how having a better understanding of the interdependency between mortality and readmission will affect how we think about high performance. If the relationship between readmission rates and hospital costs can be more accurately investigated, it will help assess whether or not conventional cost measures, such as length of stay or reference costs, offer the NHS meaningful indicators of value for money.  

Darshan is a Research and Development Manager at the Health Foundation.