NHS gooroo blog

Archive for March, 2011

Are 18-week pressures concentrated in particular parts of England? It turns out they are. Here is a map of the underlying pressures in Ophthalmology:

The map shows a coloured marker for each Trust where at least 50 patients were admitted during January 2011. Trusts with low waiting time pressures are shown as little green circles (i.e. Trusts where less than 10 per cent of patients have already waited over 18 weeks, based on incomplete pathways). But if 10 per cent of patients have already waited longer than 18 weeks, then that Trust gets a big pin stuck in it; the colour varies according to how long those 10 per cent have waited: blue for at least 18-24 weeks, purple for 24-30 weeks, and red for over 30 weeks.

As you can see, there are hotspots of multiple pressured Trusts in London and the North West, with scatterings of pressure elsewhere in the country as well.

What about other specialties? Here they all are. And if you want to look at any particular Trust in detail, you can download up-to-date reports here or click the REPORTS link in the menu bar above.

 

Analysis conducted using data from the Department of Health.
Maps created using Google fusion tables
Thanks to Toby Hillman for pressing me to improve the graphics

We’ve just published comprehensive waiting time pressure reports, at Trust and specialty level, based on the January data released by the Department of Health this morning.

The reports are available here or from the Reports menu above (free registration required). Reports are available for every Trust and surgical specialty in England where at least 50 patients were admitted in January.

Here is a sample report; now go and spy on your neighbours!

 

The January waiting time figures were published this morning, so what do they show? In short, things haven’t got worse, but they haven’t got better either. Is this worrying? Yes. Does it mean 18 weeks is doomed? No, we will still have to wait and see what happens.

So what do the figures say? Here are the trends for the time within which 90 per cent of patients are a) admitted and b) still waiting.

The headline target (that 90 per cent of admitted patients should have waited less than 18 weeks) is still being met (just). The solid line remains below 18 weeks.

More importantly, though, look at the dotted-line trend for those patients who are still waiting, which is a more leading indicator of the underlying waiting time pressures in the system. The December uptick was not reversed in January, so this may be more than a blip. Because this indicator remains over 18 weeks, it suggests that pressures have built up in the system that will lead to the headline target being breached, unless either: a) those pressures are reduced; or b) Trusts “achieve” the headline target by favouring shorter-waiting patients for admission (more about that possibility here). So to that extent I would disagree with Andrew Lansley’s recent assertion that “waiting times are stable”; if something is precariously balanced, it may not be moving, but that does not make it stable.

Let’s drill down to Trust level now. The format for the following video was explained here, and it shows the history for each Trust in England.

 

The final chart in the video, which shows the latest statistics for January 2011, is this one:

 

In about 10 per cent of English Trusts, to the right of the chart, things are getting worse for those patients still waiting. They are starting to break away from the pack, and a step is forming in the red line. To the credit of some of those Trusts, the blue dots are starting to “boil up” underneath them, which means that some Trusts are actually treating the long-waiters, instead of doggedly continuing to “achieve” the admissions-based target and hold their blue dots below the 18-week line.

What about the place where the red line crosses 18 weeks, which indicates the proportion of Trusts who are achieving 90 per cent of still-waiting patients within 18 weeks? Here is the trend:

When this indicator dropped two years ago, in December 2008, it recovered in January. When it dropped again to December 2009, it recovered in January. But this year it didn’t recover; it dropped to December, and continued to worsen very slightly in January. Which doesn’t mean that 18 weeks is doomed, but it doesn’t mean that things are getting better either.

 

What should you do if you have a large and growing waiting list backlog, and no way to clear it?

You could devote as much capacity as possible to treating the longest-waiting patients. That would keep maximum waiting times as short as possible, and keep to the principle that patients of similar clinical priority should be treated in turn. Unfortunately, admitting so many long-waiters would also fail the 18-week target: that 90 per cent of admitted patients must have waited less than 18 weeks. For this, you could be fined up to 5 per cent of your elective care revenue by your commissioners.

What’s the alternative? You could choose to achieve the target regardless. That means you could only devote up to 10 per cent of your capacity to the longest-waiting patients. The number of long-waiting patients would grow inexorably, but because you would be achieving the admission-based target you would not be fined.

When faced with this choice, different Trusts go different ways. Here is a scattergram for Ophthalmology (it’s a similar picture in other specialties too). The chart shows the time within which 90 per cent of patients are admitted (vertical axis) plotted against the time within which 90 per cent of patients are still waiting (horizontal axis). (Data is for December 2010 from Department of Health.)

 

The rump of Trusts are nestled in the desirable quarter of the 18-week gridlines. But some aren’t, and two of them are picked out in different colours: Royal Berkshire NHS Foundation Trust (in red), and Western Sussex Hospitals NHS Trust (in green). I’ve chosen them because they illustrate the two options quite well.

The next link is a drilldown into Royal Berkshire’s data. They are comfortably achieving the admission-based 18 week target. But since mid-2009 the waiting list has been growing inexorably, with 68 per cent of incomplete pathways (i.e. patients still waiting) over 18 weeks. Nevertheless, the Trust has (almost) consistently met the 18-week target by not admitting long-waiters in significant numbers. This is not a criticism of the Royal Berkshire; they have been under massive pressure to do this. But it is an indictment of the admission-based 18-weeks target.

Royal Berkshire NHS Foundation Trust: Ophthalmology

And here is a drilldown into Western Sussex’s data. Their waiting times have also grown since mid-2009, and an even higher proportion of their incomplete pathways (75 per cent) are already over 18 weeks. But this Trust has devoted a lot of capacity to treating those long-waiting patients, so that few are waiting longer than 30 weeks. Unfortunately their efforts to treat long-waiters mean they look very bad against the admission-based target, achieving only 53 per cent of admissions within 18 weeks. This puts them at risk of maximum fines from their commissioners, even though they are doing the right thing for their patients.

Western Sussex Hospitals NHS Trust: Ophthalmology

The admission-based 18-week target is clearly a problem when waiting lists grow. It deters hospitals from treating their long-waiting patients; it violates the principle that patients with similar clinical priority should be treated in turn; and it misleads the public with statistics that make everything look rosy.

The solution, I suggest, is to change the target from “90 per cent of admissions within 18 weeks” to “90 per cent of incomplete pathways within 18 weeks” (i.e. 90 per cent of those still waiting, at any point in time, should not have waited longer than 18 weeks). Now that central enforcement of the target has ceased, this could be negotiated locally between commissioners and providers.

It would mean tackling the underlying issue, of course, and that would be expensive at Trusts like the Royal Berkshire and Western Sussex. But isn’t tackling the underlying issues and improving outcomes for patients meant to be what the reformed, clinician-led NHS is all about?

The 18-week target was achieved in 2008, and has been maintained ever since. But for how much longer? The financial squeeze is putting waiting times under increasing pressure, and the Government and NHS are looking anxiously for signs that the 18 week target might be about to blow.

There are no definitive signs of this so far. But if there were, what would they be? In this post we will investigate whether the manner of the triumph might herald the manner of the fall: whether the way in which 18 weeks was achieved in the first place might hold some clues about how it might eventually unravel. As we shall see, some strange things happened as 18 weeks was achieved, and there is every chance they might happen again.

Let’s start with the official version. Here is the historical record, as published by National Statistics and the Department of Health:

The headline target (that 90 per cent of admitted patients should be treated within 18 weeks) was achieved in June 2008 and has been achieved ever since. The most recent figures (up to December 2010) show no sign of approaching failure. So far so good.

But what about those patients who haven’t been treated yet, and are still on the waiting list? The official record shows:

Again, everything looks rosy. Although if you look very closely, the number of over-18 week waiters does seem to be a little higher than in was a few months ago. It might be worth taking a closer look at this.

Instead of taking the 18 weeks bit of the target as our starting point, what happens if we take the 90 per cent bit? The next chart shows the time within which 90 per cent of English patients were admitted; it also shows, for those patients still waiting, up to how long 90 per cent have waited so far. We might expect the two lines to follow each other fairly closely; after all, if you are trying to bring down waiting times then you devote as much capacity as possible to the longest waiters, right? But in fact…

How curious. It looks as though there were still lots of patients waiting much longer than 18 weeks, when the target was achieved in mid-2008.

How can this be? It could be that some Trusts achieved the target by carefully admitting only 1 over-18-week-waiter for every 9 under-18-week-waiters, even if there were still plenty of patients waiting for ages. That would be a classic example of achieving the target but confounding the objective.

To find out, we need to drill down to Trust level. We’ll pick a Trust that achieved the headline target early, but still had long-waiters on the list. So the next chart shows data for June 2008, for the Countess of Chester Hospital NHS Foundation Trust, all specialties combined:

The over-52 weeks column for the number waiting doesn’t fit on the chart, and it’s 17,943 patients high. Three-quarters of the patients on the waiting list had been waiting more than 18 weeks. Yet this Trust was “achieving” the headline target, with 90.8 per cent of admitted patients being treated within 18 weeks.

If the objective were genuinely to achieve short waiting times for patients, you would expect as much capacity as possible to be directed towards the longest-waiters (while protecting clinical priorities). Instead the headline target throttles it, limiting to 10 per cent the capacity devoted to over-18-week-waiters.

We can only learn so much by looking at a single Trust. So let’s look at the picture across the whole of England.

In the following time-series, each English Trust admitting over 50 patients in the month is represented by two points: a blue one showing how many weeks 90 per cent of patients were admitted within, and a red one showing how many weeks 90 per cent of patients still waiting had waited within. Each monthly chart is sorted, so that the red (still-waiting) points rise continuously.

While the video is running (it’s only 2 minutes long), notice two things: firstly, the blue points falling like rain in early 2008 to “achieve the target”; secondly, watch where the red line crosses the 18-week gridline, and how it moves to the right much later as waiting times genuinely improve.

Here is the final (December 2010) chart as a still:

There are still some Trusts where 10 per cent of still-waiting patients have already waited more than 30 weeks. None of these Trusts is devoting a high proportion of capacity to long-waiters. Those six Trusts are (starting at the top right):

• University College London Hospitals NHS Foundation Trust
• South Warwickshire NHS Foundation Trust
• Yeovil District Hospital NHS Foundation Trust
• Southampton University Hospitals NHS Trust
• Royal Berkshire NHS Foundation Trust
• The Whittington Hospital NHS Trust

The place where the red line crosses 18 weeks is an interesting indicator, and it shows how many Trusts are achieving 90 per cent of patients waiting less than 18 weeks. Let’s look at the trend for this indicator:

At the start, back in 2007, hardly any Trusts were achieving 90 per cent within 18 weeks (for those patients still waiting). This rose rapidly as waiting times improved, fell back over the winter of 2009-10, and then rose again; at the peak, over three-quarters of Trusts were achieving 90 per cent within 18 weeks for those patients still waiting. Since then it has fallen again, and in December 2010 – the coldest for a century – it was just below the trough of a year earlier.

Is this a sign that 18 weeks is doomed? No; we need to wait and see what the future data holds. Is this a good indicator to watch? Quite possibly.

We have seen how Trusts were able to achieve the headline admitted-patients target in 2008, even though their waiting list backlogs were still very long. So it would be unwise to look at the admitted patient (or for that matter the non-admitted patient) records as reliable early indicators of mounting pressures on 18 weeks.

Instead, the patients who are still waiting are a better place to look: the so-called incomplete pathways. Even though 18 weeks remains “achieved”, the English waiting list still contains 14,637 patients who have waited over 52 weeks since referral. And because there is such variation between Trusts, we need to be looking at the whole spectrum of Trusts to pick up the early indicators; 18 week pressures will not build up evenly.

The King’s Fund have an excellent waiting times tracker running, showing whether the English NHS is keeping a lid on the 18 week target, or not. So far there is just the merest suggestion of an uptick; nothing that “coldest December for 100 years” wouldn’t explain.

So I thought it would be interesting to see whether this picture is replicated on other measures too. Not wanting to tread on the King’s Fund’s toes, I’ve picked two different measures from the ones they follow:

• the average (mean) referral to treatment (RTT) waiting time, because this indicates the underlying pressure on waiting times caused by the size of the waiting list (according to Little’s Law); and
• the 90th centile RTT waiting time, which corresponds to the operating standard required for admitted patients.

These measures also differ from those monitored by the Department of Health, who have chosen instead to monitor the median waiting time (which I think is potentially misleading) and the 95th centile RTT waiting time (which is a fine measure, but not the one they have specified for admitted patients).

Also, I thought it would be interesting to look at the trends by specialty. (All figures are calculated from the provider adjusted admitted pathways (or the nearest equivalent for older data) on the Department of Health’s RTT performance page.)

Before we look at the numbers, what would we expect the time trends to show? We know from simulation studies that (in a well-managed waiting list) the mean waiting time leads the 90th centile when the waiting list is growing, but lags when the list is shrinking. So on that basis, we might expect the mean waiting time to be a leading indicator of 18-week breaches. However, the national aggregate waiting list could not by any stretch be classified as “well managed” to the strict consistency of our simulation studies, and so that prediction is unlikely to hold in practice.

So let’s take a look at the data: General Surgery first.

 

We can see the dramatic plunge in waiting times as the 18 week target was achieved in 2008, with mean and 90th centile falling in lockstep. Thereafter, the target is consistently met (just). Mean waiting times remain level, indicating that the underlying pressure is not building; although theoretically this indicator could be undermined if long-waiters were being neglected (and therefore not showing up in the admission statistics). No sign of any uptick so far.

The more you drill down, of course, the more variation you find. Here is General Surgery again, with data overlaid for a single provider (Whipps Cross University Hospital NHS Trust):

Whipps Cross clearly had difficulty achieving the 18 week target in this specialty, and only did so when their mean waiting times fell in line with the English mean. However the achievement was only temporary. Mean waiting times have risen, and after battling away through 2009 the Trust was eventually unable to maintain the 90 per cent standard. Their RTT waiting times in December 2010 looked like this:

Going back to the national picture, let’s look at Orthopaedics. This specialty often has the worst waiting time pressures, because the cost per case is very high (raising the temptation to delay patients) and because the proportion of urgent patients is low (so people learn that they can neglect Orthopaedics for quite a while before problems arise). Unfortunately this also means that once a problem has arisen with Orthopaedic waiting times, it is very expensive to fix.

Sure enough, the operating standard that 90 per cent should be treated within 18 weeks has never been achieved England-wide in Orthopaedics. The good news, though, is that the mean RTT wait is not rising, and so the underlying pressures nationally are not getting worse.

Finally, here are the other main surgical specialties. The bottom line? No sign of real pressures building so far. But we’ll keep an eye on these indicators in the coming months and see what happens as the financial squeeze really starts to bite.