Welcome to HSJ’s Performance Watch expert briefing. Our fortnightly newsletter on the most pressing performance matters troubling system leaders. Contact me in confidence here.

The NHS has rightly focused on cutting patients’ length of stay. But without a thorough understanding of emergency readmissions the job is only half done, writes James Illman.

This month NHS Digital published experimental emergency readmissions data for the first time in five years – a move which has split opinion among health sector leaders.

Supporters say the measure – which looks at readmissions within 30 days of discharge – could be a crucial yardstick to help fully understand pathways and service integration for patients who are often older and frailer.

Healthwatch England also raised concerns in November about avoidable emergency readmissions.

Critics, however, say its reintroduction is a retrograde step which could cause clinicians to be overly risk averse in how quickly they discharge patients and fuel “PJ Paralysis”.

This month’s data covers overall numbers for CCGs between 2013-14 and 2017-18. NHSD is also working on a far more granular set of metrics to try and drill down into the reasons for readmissions, which it hopes to set out in the autumn.

The data suggests national emergency readmissions rose by 1.3 percentage points from 12.5 per cent in 2013-14 to 13.8 per cent in 2017-18.

There were wide local variations in the 2017-18 rates, ranging from 18.2 per cent and 10.3 per cent. The highest were Slough (18.2), Stoke on Trent (17.2) and North Staffordshire (17). The lowest were Harrow (11.5), Isle of Wight (11.5) and North East Lincolnshire (10.3). See the attached spreadsheet for all the figures across all five years.

NHSD said rates were generally higher in more deprived areas, which will come as little surprise, but the Nuffield Trust said the variation is concerning, nonetheless.

But some senior NHS figures have told me that looking at the year-on-year figures was an apples and oranges comparison.

Patient pathways have developed over this period with an increasing number of patients in and out of frailty and ambulatory care units as part of their treatment, and this makes comparisons troublesome.

Moreover, the data does not distinguish between avoidable and unavoidable readmissions, so it’s basically impossible to identify what a “good” level is.

Some go even further. Glen Burley, the well-regarded chief executive of three Midlands trusts, described the move, announced in January, as a “retrograde step which could further fuel a dangerous culture of hospitalisation”.

Responding to the HSJ article, he added: “We need to encourage our clinicians to get patients home and not see it as a failed discharge if they occasionally need to come back in… We should actually be looking at the providers with very low readmission rates and question whether they are putting patients at risk of PJ paralysis!”

Mr Burley’s concerns are valid. And emergency readmissions figures should not be viewed simply as a high rate equals bad, and low equals good. Perhaps areas with low rates need to assess why this is and if they are discharging patients quickly enough.

The NHS still needs to better incentivise swifter discharge, despite setting targets to cut the number of long-stay patients (so-called super stranded patients), a programme on which some progress has been made.

If readmission rates hamper this agenda, that would indeed be an unintended and perverse consequence.

But, despite the legitimate concerns, dismissing readmission data based on these arguments would be wrong.

As leading geriatrician Ian Sturgess, who invented the stranded patient metric and has long warned of the dangers of patients deconditioning in hospital, none of these metrics mean much in isolation.

He told me: “When measuring the stranded and super stranded metric, you need to measure three other key metrics: quality, readmission rates, and post-discharge community and social care costs.

“You have to measure them all together, in what is described as a virtuous cycle of metrics.”

And that’s the nub of the issue: a system which discharges patients in a safe and timely manner must be keeping tabs on all these things.

Healthwatch England deputy director Neil Tester said that, while there was a lot for NHSD to iron out, the new data could “allow systems to identify more easily the reason for admission and then the reason for readmission, whether it’s the same root cause and what the relationship between the two admissions was”.

He said a balanced approach was needed so patients got home as quickly as was safely possible, but “the whole system [must] keep an eye on what is happening after discharge”.

The Nuffield Trust said it had found evidence that preventable readmissions were rising. Research analyst Jessica Morris said rising rates could be “in part…simply down to an ageing population with more complex illness”.

She added: “There could also be appropriate increases which actually reflect patients having better access to treatment, for example when more ambulatory care is offered.

“But pressures on out of hospital services and social care might also play a role – our recent work suggests despite rhetoric on integration, many metrics of support for patients to stay well are flatlining or getting worse…

“Readmissions are generally not good for patients, so the NHS needs to do more to look at what is driving these trends.”

Developing meaningful and fair emergency readmission metrics is no easy task: NHSD started its work on this in 2015.

But used in the right way, such measures could become part of the virtuous cycle of metrics described by Dr Sturgess and help system leaders integrate services and identify underlying problems.