Feature / Counting the costs

29 November 2013

Login to access this content

The quality of reference costs is improving, with more data checks and greater use of patient-level costs. Steve Brown gives a quick tour of the latest data for 2012/13


It is not often that an increase in costs represents a good news story – especially in today’s environment of tight public finances and rising demand. But a small increase in the overall costs relating to the treatment of non-NHS patients, submitted as part of the 2012/13 reference costs, is exactly that. It represents a small but significant step in the journey towards improved costs.

The publication of 2012/13 costs at the end of November reveal that the overall costs submitted rose by £1.7bn. But nearly £130m of this was because NHS trusts were required to exclude the actual costs of providing non-NHS services (£758m) rather than net off the income for these services (£888m) as a proxy for costs. If the costs have been accurately identified, this suggests NHS providers made a good profit on non-NHS patients and, given that tariffs are based on reference costs, that previous tariffs for NHS services have been set too low.

Before anyone gets too excited, the figures are small: £129m is about £500,000 per trust and represents about a quarter of 1% of the £55bn of NHS expenditure in the 2012/13 reference costs. The real winner is the costing process. The NHS has known that the netting off of allowable income – to use the reference costs jargon – distorts costs. The new publication represents the first step towards eliminating this distortion as there are plans to apply the same approach to the more complex funding streams of education and training and research and development.

The reference costs provide significant detail on NHS costs, accompanied by increasing levels of analysis and explanation to guide the non-accountant through the phone directory-sized publications. The headline figures, while undeniably too high-level to ever be used as robust benchmarks, make interesting reading for anyone wondering about the link between what the NHS does and how much it costs.

Some £55bn (54%) of the £103bn NHS budget for 2012/13 is covered by the costs, accounting for 5.6 million data items submitted by 244 NHS trusts and foundation trusts. Detailed costs were submitted for 2,100 treatments or procedures covering 15 million finished consultant episodes in admitted patient care alone.

The average cost of a day case was £693 (up from £682 in 2011/12). An average inpatient stay cost £3,366 (up from £3,215), while for non-electives this was £1,489 (£1,436). These figures do not include excess bed days, the average cost of which was £273 (£264). An outpatient attendance cost £108 (£106), while an A&E attendance cost the NHS an average of £114 (£108). These increases (1.6% for an average day case, 4.7% for elective inpatients) should be seen against a 1.8% cut in tariff prices for 2012/13, combining a 2.2% allowance for inflation, and a 4% national efficiency requirement.

The real value for NHS providers comes at the more detailed level. A reference costs schedule shows both the average cost for each healthcare resource group (the currency for admitted patient care) and the range of costs across all providers. So, a tonsillectomy for an under-18-year-old (with no complications or comorbidities) remains one of the service’s modestly high-volume activities with modestly high costs (just over 33,000 operations, with slightly more day cases than elective inpatients). The average cost was £1,316, with elective inpatients costing £330 more on average than a day case. The interquartile range for elective cases reported with a service description of ear, nose and throat (as opposed to general surgery or paediatric ENT, for example) ranged from £1,201 to £1,641.

The publication also compared the overall costs of different providers using a reference cost index, with 100 equating to national average costs for same mix of patients. The full range stretches from 63 to 137 (though this ignores two outliers where data quality was a known issue).

Reference costs have suffered a poor press over the years. Criticisms have been levelled at the historically top-down approach to cost calculation, with, for example, different treatments of key cost components such as overheads making comparisons meaningless. The netting off of allowable income has also attracted flak and the inability to drill into patient specific costs has often made it difficult to investigate any apparent cost differences.

But things are changing. There has been a 30% increase in the number of trusts that have implemented patient-level costing, with almost all of them using the more detailed costs to ‘underpin some or all of their reference costs return’.  Similar numbers are also following the HFMA Clinical costing standards, which although not mandatory, set out best practice in cost allocation and costing practice.

The Department of Health, which continues to collect reference costs on behalf of Monitor (now responsible for price setting), has also been tightening up on data validation with further checks introduced for 2012/13. At the highest level, for the first time boards were required to approve the costing process for their organisations’ submissions. And finance directors, historically required to sign off the costs themselves, this year had to sign off against revised requirements. These included confirmation of board approval, as well as that clinicians had been engaged in the costing process and that a self-assessment quality checklist had been used.

The checklist confirms that all but one of the submitting organisations reconciled their costs to within 1% of the signed annual accounts. It also forces trusts to review and justify very low and very high unit costs.

There were two new mandatory validations and several new non-mandatory validations built into the templates. In most cases, the Department says the validations show reducing numbers of issues over the past three years, suggesting data quality is improving.

Also helping comparability of cost data is the use of a revised currency. HRG4+, which pushes the HRG count up beyond 2,000, gets its first run out in the 2012/13 reference costs. The currency aims to take greater account of complexity and comorbidities, as well as providing more age splits, and should help separate out the costs (and in future the prices) of specialised activity from more routine cases.

In theory, the new HRG4+ currency could provide the basis for the national tariff in 2015/16 – though the decision to base 2014/15 prices on those in 2013/14 rather than on 2011/12 reference costs must throw this into doubt. So, for trusts looking to get ahead of the game on planning, the new reference costs could give a glimpse of future income levels.

Index and schedules

The reference cost index (RCI) provides a measure of relative efficiency of providers, showing the actual cost of a trust’s casemix compared with the same casemix delivered at national average cost. A trust with costs equal to the national average scores 100; a score of 110 suggests costs that are 10% above the average; 90 equates to 10% below the average.

The range this year stretches from 63 (Tavistock and Portman NHS Foundation Trust) to 137 (Royal National Orthopaedic Hospital NHS Trust) – though this excludes two organisations whose index was influenced by known data quality issues. Almost half of all trusts have an RCI within five points of 100.

Stripping out trusts with data quality issues at both ends of the spectrum, the RCI range for general acute trusts stretches from 88 for the Princess Alexandra Hospital to 118 for Mid Staffordshire NHS FT. This again excludes some organisations with data issues.

The index is seen as a crude measure of efficiency, but many trusts still report it to their boards, and trusts are encouraged to use the index to ask questions and explore their performance relative to comparable organisations. The reference cost schedule shows the national average costs and the interquartile range of costs across NHS providers for all healthcare resource groups.

The three highest volume elective HRGs according to reference costs were major knee procedures, major hip procedures and major open upper genital tract procedures (see table). The range of costs for HB21C knee procedures, reported within trauma and orthopaedics (the schedule for the first-time reports back HRGs by the service description provided by trusts) is £5,101 (lower quartile) to £6,319 (upper quartile) – from 10% under national average costs to 11% over. 

High-volume elective inpatients
HRG  Currency description  Activity  Unit cost  Total cost 
HB21C  Major knee procedures for non-trauma, category 2, without CC  41,089  £5,675  £233,179,772 
HB12C  Major hip procedures for non-trauma, category 1 without CC  30,802  £5,880  £181,119,053 
MA07G  Major open upper genital tract procedures with CC score 0-2  25,591  £3,255  £83,310,881 

Looking across all settings, the highest volume activity (not including outpatient attendances) was direct access plain film X-rays (reported alongside HRGs but not strictly an HRG itself). There were 5.3 million of these diagnostic events at an average cost of £28. Two emergency medicine HRGs (VB09Z and VB11Z) also break the 4.5 million activity mark, with average unit costs of £92 and £68 respectively. Of more than 69 million outpatient attendances, physiotherapy tops the activity charts, with 7.1 million attendances costing an average of £42. Some 6.3 million trauma and orthopaedics attendances had average unit costs of £110.

The second year of cluster-based reference costs for mental health shows overall unit costs per day in cluster (excluding initial assessment) ranging from £5 for cluster 18, Cognitive impairment (low need), to £103 for cluster 14, Psychotic crisis. The high unit cost is driven by the high percentage of admitted to non-admitted care days compared with other clusters.



2013/14 draft guidance

The Department of Health published draft Reference costs guidance for 2013/14 alongside the 2012/13 data publications. This confirmed the already trailed mandatory cost collections for education and training costs (in January 2014 and around next year’s reference costs collection). However, neither will have any impact on the 2013/14 reference costs, which will be submitted net of education and training income as usual.

Other changes include a collection of HIV adult outpatient attendances against three groupings as part of the introduction of year of care pathway currencies. There also new currencies, including a wheelchair services currency and changes to promote costing in community services.

With trusts spending an average of 93 days collating and submitting reference cost returns, the Department is committed to taking steps to simplify the reference cost collection – in part a response to efforts to reduce the burden of central returns more generally.

So, for example, consideration is being given to removing the distinction between non-elective long and short stays for next year’s cost return. The separated data has informed lower prices for short stays. However, the adjustment has only been updated periodically and could be performed using other data. Consultation on the draft guidance runs until 13 December.