Feature / Don’t count on it

At one time NHS reference costs were seen as useful, if complex, benchmarking information. But with the advent of service line reporting and patient level costing, have reference costs lost their value? Steve Brown reports  

What can or should we make of the reference costs, published quietly by the Department of Health during May? On the face of it, this telephone directory-sized publication of data exposes the relative cost performance of England’s 400 NHS trusts, foundation trusts and primary care trusts. But for some in the service, the exercise is simply about meeting a central requirement, and the information provides little insight into how to run their organisations locally.

The 2007/08 reference costs cover £44bn of expenditure – an expansion of around £2.5bn on 2006/07. This is also the second year that the reference costs were collected using the new HRG4 healthcare resource group currency – the currency that has been used as the basis for the new tariff in 2009/10. The newly published national average costs for the 1,400 or so HRGs, plus outpatient contacts, will form the basis for the national tariff in 2010/11.

Reference costs have two distinct parts. First there is a reference cost index (RCI), which effectively ranks organisations on their relative costs for delivering the same activity. An organisation with an RCI of 100 has costs equal to the national average. A score of 125, for example, means that costs are 25% above the average, while a score of 92 shows that costs are 8% below the average.

Then there are the reference costs schedules. All organisations submit their average costs for providing different treatments and procedures and the Department of Health publishes figures showing the national average cost and the interquartile range – the range of costs exhibited by the middle 50% of NHS providers.

The RCI range for PCTs stretches from 66 to 280. Even stripping out this top figure, where there were identified data problems, the range runs from 34% under national average costs to 59% above. However, it is widely recognised that the RCI is pretty meaningless for PCT providers given the different approaches and practices in describing and basic recording of activity.

Hospital RCI

It is in the hospital sector that the RCI has traditionally been regarded as having more validity. The full range for NHS trusts stretches from Great Ormond Street NHS Trust at 162 (62% above national average costs) to Southend Hospital NHS Foundation Trust and South Warwickshire General Hospital NHS?Trust, both at 81 (19% below national average costs) – a range of 81 percentage points. (This is using the market forces factor adjusted index, taking account of the unavoidable cost differences facing different organisations in different parts of the country. The MFF has been revised from that used in 2006/07.)

But even here the reference costs have to be read with a major health warning. For a start, there were specific data problems at Great Ormond Street – radiology was not being unbundled and there was unplanned cross subsidisation between inpatient ward costs and critical care and high dependency units.

More generally, as with community providers, mental health trust RCIs are difficult to compare given the data collection inconsistencies from trust to trust. And specialist trusts – orthopaedic and children’s, for instance – have historically fared badly in the national comparison, partly because even within the same HRGs they would have a tendency to undertake a more complex and therefore more expensive casemix.

Stripping out mental health trusts, single specialty providers and ambulance trusts, the range for acute providers (from basic district general providers to those with a high level of tertiary activity) stretches from University Hospital Birmingham NHS Foundation Trust (UHB) at 116 down to Southend and South Warwickshire (81). This range of 35 percentage points shows a tightening of the range from 2006/07, when a 58 percentage point range ran from 80 to 138.

However, there are significant moves within this range. UHB’s RCI, for instance, moved from a near national average 103 to 116. Madeleine Parmar, head of costing contracts and income, says the two years’ index scores are ‘comparing apples and pears’. There is often a tendency with annual assessments to blame the data for any change in the wrong direction, but many make a good case for this argument with reference costs.

First, there have been changes in what is included within the overall RCI between the two years. So 2006/07’s overall RCI calculation excluded adult critical care and a range of unbundled services, while 2007/08’s included critical care but excluded a different range of unbundled services. A particularly influential change for Birmingham was the exclusion of accident and emergency, as the inclusion of its relatively efficient A&E-specific index of 73 would have brought its overall RCI down.

Ms Parmar also points to the use of newly introduced reserve codes as a contributory factor to the change in relative performance. Reserve codes have been used  where trusts cannot report robust unit cost data at HRG level. Ms Parmer suggests that the greater use of reserve codes in some organisations makes some costs difficult to compare.

The use of the new revised market forces factor (which reduced the relevant MFF for UHB from 0.9967 to 0.9645) also played its part. ‘I can explain the changes through a combination of changes in methodology and some things, in retrospective analysis, that we’d have done differently. For instance, there was some additional activity that would have increased the denominator and there was a problem with excess bed days,’ she says.

Despite this Ms Parmar maintains there is some value for services in looking at how they compare to peer groups (provided in large tertiary teaching hospitals) and that cost comparisons do feature in management accountants’ discussions with services – although as patient level costing and service line management develops, reference cost data is playing a less prominent role.

Many finance directors cast similar doubts on the usefulness of drawing national comparisons from the national RCI tables. Tony Whitfield, finance director at Salford Royal NHS Foundation Trust (RCI 96), which has made good progress with patient level costing, believes that the movement of the organisation in the national cost rankings measured by RCI is not the key focus for finance directors.

‘We do our reference costs in a professional and proper way,’ he says. ‘That doesn’t necessarily mean we feel they are an overly useful tool. The best they can make is a modest contribution to the overall efficiency debate of the organisation internally. Organisations need to understand their own costs against their own drivers – that is where the real effort and energy is needed.’

Another foundation trust finance director admits that his organisation would spend no time reconciling any movement between years. ‘We do not use the reference cost index as an indicator of efficiency. And although at one time it was, it is no longer reported to the board because it is generally a discredited indicator.  The reference costs are currently simply a mandatory exercise to supply data to the Department for them to construct a national tariff.’

While everyone spoken to by Healthcare Finance recognises the need for caution in making comparisons and drawing conclusions from reference costs, the RCI continues to be monitored in some organisations. ‘We do report it [RCI] to the board,’ says Paul Assinder, finance director of the Dudley Group of Hospitals NHS Foundation Trust, ‘and we have a corporate objective to reduce RCI.’

For 2007/08 the trust matched its 2006/07 score of 94, despite the various methodology changes and a revised MFF.

And at University Hospitals Coventry and Warwickshire NHS Trust, finance director Andy Hardy believes the RCI can at least indicate direction of travel and reflect realised cost improvements.

‘Our reference costs went up in 2006/07 [from 97 to 118] partly as a result of our move into a new private finance initiative hospital. Our cost base went up and that was seen in our reference costs. As a result we had to find some £30m of savings to break even in 2007/08. We achieved that and that is clearly reflected in the improvement this year [RCI 109].’

Similarly Tim Bennett, finance director at University Hospitals of Morecambe Bay NHS Trust, says that the trust’s move from an RCI of 110 in 2005/06 to 93 in 2007/08 – ignoring a score of 138 which was attributed to data problems in 2006/07 – reflects real cost improvement.

‘We’ve had one of the highest levels of cost improvement programme delivery that I’m aware of and we had to do that because we were coming out of turnaround,’ he says.

But he agrees that internally the focus is not on the national RCI. While the RCI is reported to the board, comparisons at individual HRG level are seen as more useful. However, even here the trust is moving more towards use of its service line reporting data and patient level costing rather than HRG costings.

As Mr Bennett explains, clinicians relate much more to individual patient episodes than to average costs for undertaking an HRG. ‘[Reference costs] can point out areas that offer opportunities for improvement but it doesn’t tell you exactly how to get that improvement. For that you need to drill down into much lower levels of detail to understand if it is your length of stay, the cost of prosthesis, the number of patients you treat per consultant or the casemix that is driving the cost.’

Some organisations do use their own HRG costs to inform service line costs to marry up with service line income. But many organisations also use the local reference costs in the local contracting process.

At Dudley, Mr Assinder says: ‘We use reference costs mostly to determine local prices, with a lot of our activity sitting outside payment by results.’

In particular this year, with the tariff moving from HRG version 3.5 to HRG4, the local reference costs have helped inform discussions about what is a reasonable transition price. And local HRG4 costs for new obstetrics admission HRGs – for admissions not related to delivery – have been used by local agreement in place of the national tariff, an area that has been causing problems across the country.

Ad hoc uses

Leeds Teaching Hospitals NHS Trust also says reference costs are put to a range of ad hoc uses. Its RCI moved from 123 to 114, although it suggests that the 2006/07 figure was artificially high. It lists local uses for reference costs as contracting, contract monitoring, benchmarking and for business cases.

Chris Raspin, development manager at the Audit Commission, believes that reference costs do still have a valuable role in providing useful management information.

‘Reference costs provide a wealth of information about cost efficiency in health, which simply does not exist in other sectors,’ he says.  As such, it is extremely valuable, not only for its most recent purpose of providing information on which to base tariffs, but also as a tool to help local providers identify efficiency improvements.’

However, he suggests that growth in the coverage of reference costs and the introduction of HRG4 means that NHS providers now have to put a great deal of effort into compiling their submissions. 

‘While each new development of the system has aimed to improve the accuracy and comparability of the outputs, in practice it

may also have increased the likelihood of error and made analysis more difficult rather than easier,’ Mr Raspin adds.

Plans to review the quality of data – a step short of formally auditing reference cost submissions – could help eliminate some of the errors and inconsistencies.

However, generally around the service there is a feeling that the value of reference costs is on the wane, with trusts looking to other data sources for their assurances on cost efficiency.

For many finance directors the real point of reference costs is to enable the creation of a national tariff. But the Department is already moving towards a few best practice tariffs for 2010/11. And there are longer term plans to inform the tariff using patient level costing data from a sample of organisations (see Healthcare Finance April 2009, page 5).

It would appear that reference costs – at least in their current form – face a somewhat uncertain future.

To download a pdf of this article as it appeared in Healthcare Finance, click here