Primary Care Foundation Blog

Andrew Anderson and Martin Roland have recently reported that experienced GPs can reduce the number of referrals from NHS 111 (Potential for advice from doctors to reduce the number of patients referred to emergency departments by NHS 111 call handlers: observational study - http://bmjopen.bmj.com/content/5/11/e009444.abstract).

It is no surprise that GPs can, when re-triaging patients after an NHS Pathways assessment by non-clinicians, arrive at very different dispositions.  The typical proportion of ‘Contact GP’ dispositions that an out of hours organisation using GPs to carefully assess each call can convert to ‘telephone advice’ is over 50% for example.

 But what is a surprise is the bullishness of the reports about the article in suggesting that there is a significant cost saving for commissioners.  For example Pulse reporting on the paper said that “there could be a cost saving for commissioners of £11,112 if GPs were employed in the call centres, comparing savings made for A&E with GP salary costs”. 

The saving of £11,112 comes from subtracting the cost of the GPs providing the triage (£41,416) from the saving in A&E attendance (£52,528).  But Pulse has ignored the cost of the outcomes for those patients that are referred elsewhere than A&E (this is perhaps understandable as the original paper fails to mention them either).  There are increases in referral to other services such as Minor Injury Units (using the same set of reference costs at £49/case, this is £3742) and to the OOH service for face to face consultations (at a conservative £34 per case this is £20026, but it is probably rather higher) not to mention an unknown number of cases completed over the phone but advised to contact another service or their own GP next day.  However, even ignoring the latter, once these figures are added in there is no saving – the overall cost did not fall by 15%, rather it rose by at least 17%.

Of course, this is not the full story.  A reduction in cost is far from the only thing that we are interested in and we may be happy to pay more to reduce the pressure on A&E and to improve the patient experience.  And we don’t know other things too such as how many patients follow the advice that they were given (which the authors rightly point out would be useful to know as part of a full cost assessment) or how many of those referred to the MIU or OOH service were subsequently referred on to hospital.

And it may be that there is potential to reduce costs using GP assessment as described.  This was a small scale trial, had it been carried out at scale the cost per case might have been lower.  We have recently completed a model for NHS England, commissioners and providers that looks at the cost of integrated urgent care, focused particularly on NHS 111 and OOH services.  Using a figure of £15 or as an indicative cost per case for phone assessment by GPs the cost for the 1474 cases would have been low enough that a modest saving might possibly have been made.

The concluding sentence in the paper says that there is a need “for further research to establish the cost-effectiveness of different approaches to triaging telephone requests for care”.  I agree with this. 

But the approach adopted and described in the paper is far from the only one. 

Proponents of NHS Pathways would point to the importance of training call-handlers in probing when asking questions – rightly they would say that many existing 111 services have much to do in training call-handlers about how to do this - and that this would make a big difference in referrals to A&E and the Ambulance Service. 

There are also opportunities to increase the proportion of clinicians directly supporting call-handlers whilst they are taking the call.  Not all 111 services do this – and they certainly don’t have the number of ‘floor walkers’ to be able to give timely support to the majority of cases whenever the call-handler approaches an A&E or Ambulance disposition.  To avoid delay to the despatch of an ambulance this would seem to be the right point to intervene in the process at least for Ambulance dispositions (and the cost of these dispositions is much more expensive than those to A&E) so it would seem sensible to look at the benefits of this approach.

The new commissioning standards allow commissioners and providers these sorts of freedoms (provided proper risk assessment, management and audit is included as part of the change).  What is required as people experiment with different ways of working is a careful review of the impact looking at ALL of the costs and ensuring that this is reported widely so that others can learn.

I have sympathy for anyone interviewed on television - I bet I would get it wrong!

But Keith Willett's claim needs correcting. He says: “We know that 15% of people that tried last year to get a GP appointment were unable to achieve one in the time they wanted and half of these turned up at an Urgent Care Centre or A&E”

 A quick sense check reveals this can't be right:
• There are about 300 Million GP consultations a year, roughly 5.5 times a year that we go to our GP
• 15% of 300 Million is 45 Million
• Half of that figure is 22.5 Million
• This is larger than the total number of attendances at A&E, MIU, WIC and UCCs in England which is reported to be 21.7 Million

I feel sure that he was referring to the GP Patient Survey results. If he had said: “The GP patient survey showed that 15% of people that tried last year to get a GP appointment were not able to get one or were offered an inconvenient one and 10% of these said that they turned up at an Urgent Care Centre or A&E” he would have been pretty much spot on.

Reputable reports assembled by the great and good are frequently citing the dramatic increase in emergency admissions.  Recently the evidence assembled for the Bruce Keogh review into Urgent and Emergency Care said that "Emergency admissions to hospital in England are also increasing, with a rise of 40 per cent between 2003/04 and 2010/11". They supported this by referring to the excellent analysis by Ian Blunt, Martin Bardsley and Jennifer DIxon ('Trends in emergency admissions in England 2004-2009; is greater efficiency breeding inefficiency" from the Nuffield Trust). 

But they have got it wrong. Actually the report says in the first bullet point on the first page, under the heading 'Key Points' that:

"The number of emergency admissions in England rose by 11.8 per cent over the five-year period 2004/05 to 2008/09 – a total of approximately 1.35 million extra admissions."

What they seem to be referring to is the rise in emergency admissions through A&E, and they produce the graph shown below which, as you can see, is clearly labelled and defined as 'admissions through A&E'.

 So what is really going on?  I have carried out a simple analysis of the reported data for all England. The graph (below )shows that emergency admissions are increasing, but at a more modest rate than is often claimed with the rise being less than 13% comparing the year to January 2013 with the financial year 2007/8, with the rise in emergency admissions through A&E of nearly 50% being offset by a reduction in non-A&E emergency admissions of approaching 30%.

To provide the numbers;  In 2007/8 there were 4,753,318 emergency admissions, 2,546,425 through A&E and 2,206,893 not through A&E.  In the 12 months to Jan 2013 there were 5,357,115 emergency admissions, 3,781,495 through A&E but only 1,575,620 not through A&E

Another way of looking at this is to say that compared with 2007/8 (when the number of admissions through A&E were 2,546,425 out of a total of attendances of 4,753,318) the switch of patients (631,273 of them) from alternative routes for admission to going through A&E (if that is what we assume is happening) will have accounted on its own for a growth of 24.8% in emergency admissions through A&E and 13.2% in attendances at A&E.

As always in this blog, I am keen to start a discussion.  Do you agree with the picture above?  What am I missing?  What other similar analysis have you seen of this issue?

So what might be going on?  And here I stress that the remainder of this posting is speculation or hypothesis. 

The first thing is to dismiss the suggestion made to me that primary and community care is performing so well and their intensive focus on reducing emergency admissions is meaning that they are referring many fewer patients to hospital for admission - and that the reduction in non-A&E emergency admissions is because of that.  Some of my best friends are GPs and I know that they and their colleagues do a great job.  I wish it were possible, but I cannot believe that any practice could have managed to reduce emergency admissions of their patients by more than 25% (and if they did I would wonder what they had been doing wrong in the past).  I certainly think it is beyond belief that all practices and community services in England could have improved in less than four years to this extent.

So what else?  Could it be that this is a symptom of the increasing occupancy of hospitals across the country (a result as much of the difficulty in getting people out of hospital as the increase in numbers coming in) and that the shortage of beds is making all sorts of admission more difficult - in fact so difficult that the hard-pressed GP or other health professional has, in too many cases, no alternative but to call an ambulance or send the patient directly to A&E?  A&E is, after all, one of the few parts of the hospital that cannot close its doors to patients (labour wards for obstetrics and gynaecology is another example, but here demand for services is not mainly shaped by the way the NHS works but by other factors with a nine month lead-time).  Perhaps you have other possible rationales - it would be good to hear them….