Southend A&E 4 hour performance

There was a chart circulated at yesterday’s Place Scrutiny that I thought could have been simpler and easier to assimilate. So, here is my version:

Apr – Jun 2014 Jul – Sep 2014 Year to date
Basildon & Thurrock 95.95 94.68 95.34
Southend 94.14 96.46 95.28
Essex 93.45 93.10 93.28
Colchester 94.78 91.16 93.01
Mid Essex 91.20 92.23 91.70
Princess Alexandra 90.91 90.27 90.62

All figures are percentages and show how many were dealt with within the four hour target. This is not about being initially seen (triaged) but the length of time from entry into A&E to leaving.

Southend does not come out too badly as compared to other Essex hospitals, although any missed target is not good.

The target is 95%. The numbers in bold red are where the target was missed.


3 Responses to Southend A&E 4 hour performance

  1. But is the target appropriate? And if it is not, is it appropriate to chase the target?

    It depends on the mix of patients presenting and how they arrive.

    You would hope that the triage process meant that people coming in complaining of pains in the chest would go straight through to treatment (which even then may take more than 4hrs), whilst someone complaining of say a sore foot waited. What would you want to do about the drunks? Get them dealt with quickly so they do not cause problems in the waiting area, or make them wait whilst other more medically serious cases got treated first?

    I suspect if you arrive by ambulance you will get treated quicker than someone presenting at the front desk with identical symptoms (if only because you want to free up the ambulance).

    If someone is getting close to the 4hr limit, the law of unintended consequences comes in. You recognise that they will not meet the target so you let them wait (if necessary well over the 4hrs) and deal with other cases first which you know can meet the 4hr target. It “get’s the numbers right”. With any target it is worth pondering “how can I subvert this target?” – and then consider whether is is sensible.

    A single target is a blunt instrument. Process control theory says as a minimum you have a control limit and a tolerance limit. This might mean saying 95% within 4hours and no one over 6hrs. (I bet they do not measure “process capability” however.) Separate targets for different triage levels may actually help the hospital improve (rather than just produce numbers for councillors to debate). If life-threatening cases are being dealt with within target, but “cuts and grazes, dislocations and minor breaks” are not, perhaps there is a need for a minor injuries clinic to take the strain off A&E? The numbers then help make the case.

    Do you measure over a month (which averages out the Friday & Saturday evening hot spots and also averages out the impact of say a major road accident), or do you measure over a day – or even over an hour? Do you want to just produce “the right figures” (even if they hide some regular periods of extreme waits) or do you want to highlight that there is a problem on Friday nights?

    It is quite possible looking in more detail at the above figures that people in “non-conforming” hospitals may feel “all things considered” that their A&E did better than one which had “got the numbers right”.

  2. I accept that targets have their shortcomings, but how else do you measure performance? Bearing in mind that only 19 out of 20 cases have to meet the four hour limit this does allow for those exceptions that you have so eloquently described.

    What targets do very successfully is measure trends. If you have hit the target for, say, six years and then suddenly miss it then that does need explaining.

  3. May be I’m a cynic because whilst I accept in theory what you say, I have seen crude “target” measures used in appropriately.

    If you have hit the target for, say, six years and then suddenly miss it then that does need explaining.

    If your target was 95% and you got 95.1, 97.2, 95.1, 98.6, 95.2, 97.0, 94.9, I don’t think the 94.9 needs explaining but the political headline might read, Pass, Pass, Pass, Pass, Pass, Pass, FAIL!, which is unfair.

    The average of the figures immediately above is approximately 96% but the standard deviation is about 1.3%. What does this mean? Very crudely you can expect about 1/3 of results to be more than one standard deviation away from the mean and 5% to be more than two standard deviations from the mean.

    So let’s look at 96% (the mean) +/- 1.3 (1 standard deviation): so,
    – 2/3rds within the range 94.7% and 97.3% (Note that our FAIL figure of 94.9 is within this range)
    – 1/3rd outside this range, so say 1/6th below 94.7 and 1/6th above 97.3
    The important question is why will they be outside the range?
    Is it incompetence, negligence or recklessness (low figures)?
    Is it excess resources (high figures)?
    Or is it pure chance?
    Statistical theory would point towards chance.

    I have taken some liberties with statistical theory here to keep the numbers reasonably simple, but the point is that if you have a process that is highly variable, a measurement below some politically set target could be due purely to chance and as such it is inappropriate to take action as any corrective active may be so crude as to cause other problems (e.g. wasted resources or resourced diverted from another field).

    In industry control limits are often set at two standard deviations.

    I guess the conclusion is that political targets and management are not easy bedfellows.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s

%d bloggers like this: