Feature / The inside story

Reference costs tend to be roundly criticised in finance circles but the activity and costing data provides interesting information. In the first of an occasional series applying an HRG4 lens to NHS specialties, Steve Brown looks into maternity

No-one pretends that reference costs are the most robust source of detailed costing information. Used as benchmarking data to compare local costs for specific activities with the national average, the current tariff or with costs in peer group organisations, it can only provide a starting point for detailed analysis and interrogation.

Differences in casemix (not picked up by the more granular HRG4 healthcare resource groups) and costing, coding and recording practice (despite costing manuals and data recording guidance) mean that attempting to identify high- and low-cost organisations is a complicated business.

But come up a level or two from the local cost data and the reference cost data can help provide a detailed picture of activity in different healthcare specialties. It can also shed light on how the NHS spends its money.

Maternity provides a good example of this. Programme budgeting information suggests that the NHS in England spent just over £3bn of its £97bn gross expenditure on maternity and reproductive health. But reference costs data for the same year shows that nearly £2.3bn of this was tied to maternity care.

Some 46% of this expenditure (£1.05bn) related to deliveries – normal, assisted and caesarean (not including home births). And understandably the vast majority of these deliveries was undertaken in NHS trusts.

In fact just 5,806 deliveries were recorded by PCTs out of a total recorded 625,256 deliveries.

Code discrepancy

Hospital maternity statistics suggest a higher level of around 650,000. The discrepancy is because hospital episode statistics (HES) data uses the delivery OPCS codes, while the hierarchy arrangements within HRG mean that some deliveries would be hidden within groupings with a higher hierarchy.

For example, if a patient has a normal delivery but then haemorrhages, leading to a hysterectomy, the episode would be included within chapter M (female reproductive system). 

The fact that there were any deliveries at all in PCT land is largely down to three PCT-run, midwife-led birthing units in South Staffordshire, Gloucestershire and Devon.

It is not clear from the reference costs whether there are other PCTs that run midwife-led birthing units or whether the activity is simply not being captured.

There were a further 23,504 home births recorded in reference costs (within trust and PCT community midwifery services).

The HES maternity statistics (www.ic.nhs.uk) provide good detail on maternity interventions, but reference costs gives a financial angle. For instance, according to reference costs, 63% of all deliveries (not including home deliveries) were ‘normal’, compared with caesarean sections at 24%. But normal deliveries only accounted for 47% of delivery costs, while caesareans consumed nearly 40% of the total. Assisted deliveries – for instance, using forceps – accounted for 12% of activity and 14% of costs.

According to HES data, the caesarean rate did not increase in 2008/09 and remained at a constant 24% from the previous year. However, this rate is still – as with many other developed countries – far higher than the 10% to 15% recommended as a maximum by the World Health Organisation.

Average costs

On average (across all normal delivery categories including all age groups and with or without complications and comorbidities), a normal delivery costs the NHS £1,256, while a caesarean costs £2,715 – so ensuring appropriate use of caesareans has a financial imperative as well as the clear patient benefit.

Childbirth is, relatively speaking, an unplanned event. And this is reflected in the two admission codes specific to maternity admissions – 31: admitted antepartum and

32: admitted postpartum – which are both non-elective admission codes.

There are limited circumstances in which a delivery could be classed as elective. For instance, if a woman had been admitted electively for a separate procedure and had gone into labour during her inpatient stay, this would class as an elective episode.

But even planned caesareans or ‘booked’ admissions to induce labour should map to non-elective activity. All of which suggests

that a significant proportion of the nearly 13,000 deliveries recorded as elective inpatients or day cases have more to do with coding than actual practice.

In terms of the delivery HRGs, the new HRG4 groupings were consistent in the first two years of use. But there is a minor change in the third year. The NHS should already be aware of this as it has collected reference costs on this basis for 2008/09, but it will perhaps sink home fully when it forms the basis for payment in 2011/12.

The change is in the age split. Initially HRG4 only recognised differences in resource consumption below and above 19 years of age. However, the 2008/09 reference costs split costs into two groups – from 16 to 40 and anyone outside this range.

This recognises that real differences in resource use start at 16, not 19, but also reflect the potential for more difficult cases and higher costs involved with mothers over the age of 40. The tariff will seek to reflect these cost differences from 2011.

Because the costs will also split differently between the different age splits, there should be no financial impact on trusts. A quick analysis suggests that 3% of all deliveries in 2007/08 mapped to the higher cost 18 years and below HRGs, and therefore would have attracted a higher tariff.

In 2008/09, just 2% of all deliveries fall into the new below-16 or over-40 categories. However, the activity mix in the different years will have been different, as demonstrated by the change in the proportion of delivery HRGs where age is not specified (rising from 19% to 28% over the two years).

The big change in HRGs in recent years – which has had an impact on both reference cost collection and tariff remuneration – has been away from the delivery suite. Under

HRG v3.5, the largest maternity HRG by volume was N12 – antenatal admissions not related to delivery event – with approximately 700,000 spells in 2007/08 (when tariff was still based on HRG v3.5 – see box overleaf). N12 was a catch-all for all non-delivery obstetric activity, including both antenatal and postnatal events relating to the mother.

N12 guidance

In 2008/09, following concerns about the recording and coding of patients to N12, the Department of Health issued guidance clarifying that to be classed as an N12 a patient had to be admitted. It said a stay of less than four hours should not be classed as an admission, but recorded as a first obstetric outpatient attendance.

Trusts and commissioners were asked to agree appropriate levels to provide a control on levels of usage and cost. HRG4 (introduced for tariff purposes in 2009/10) was trailed as the solution. It provided more granularity to recognise the different activities – from short stays for observation or investigation (for instance requiring an ultrasound) to overnight stays – and the different associated costs.

But in terms of simple cost, it didn’t provide a solution in 2009/10. The cheapest new HRG contact (NZ04 clinical contact for observation) was priced at £499 – just £42 lower than the previous N12 price. The other contacts were priced from £700 upwards in terms of cost.

The PBR guidance for the year noted that there was still a large differential between the tariff for NZ04 (£499) and a first obstetrics outpatient (£145).

More accurate coding of activity will over time correct this issue, but trusts and commissioners remain unclear about whether antenatal day unit activity should be recorded as admissions (NZ04 – NZ10) or outpatient attendances.

The first step in determining the correct treatment of a non-delivery event is to look at the length of stay. With a length of stay of zero (no overnight stay), the event could be an HRG grouping – recorded as an outpatient procedure, a day case or short stay emergency – or an outpatient attendance.

An OPCS procedure code (which includes, for instance, an ultrasound scan) would mean HRG4, while no OPCS code means outpatient attendance.

With the OPCS code, the recording of the event depends on the data set being used – either admitted patient care or outpatient.

And this can be down to historic practice or the way the ward desk is set up – all ward attenders might automatically be recorded as outpatients, for instance.

Different practices on booking in patients will also make a difference. A worried mum-to-be phoning in to the ward could be told to come in. Some trusts would record this as a booking, hence an elective day case, others would treat it as non-elective.

Recording tactics

In payment terms the elective/non-elective issue is immaterial – the tariff in 2010/11 is the same. But recording as an outpatient procedure or a day case/non-elective moves the price paid from £141 to £460. Add the fact that many of the outpatient attendances could have involved similar interventions such as scans that simply weren’t recorded – so attracting an outpatient attendance of £138 – and the significance of the issue becomes clear.

If the episode includes an overnight stay, the issue is clearer. You can’t have an outpatient (or a day case for that matter) involving an overnight stay so the only option is an elective or non-elective HRG.

It is understood that different hospitals within the same trust even operate different recording conventions as a result of different historic practice. Some trusts are still using the four-hour rule of thumb – if an event duration is less than four hours this would be recorded as an outpatient, otherwise an admission – even though this applied to the old N12 HRG under HRG v3.5 and there is no mention of any such approach in the new PBR guidance.

The labels for the non-delivery HRGs have been changed in the local payment grouper for 2010/11 to reinforce the key distinction – zero length of stay or an overnight stay (see table previous page). But while this should reinforce the logic, the HRGs are in fact assigned on exactly the same basis by the grouper, pulling data from the various completed data sets. The service is keen for more guidance.

There has been no indication from the Department about a way forward. However while the general non-mandatory planned same day tariff operating in 2009/10 has been abandoned for 2010/11, the Department’s director of NHS finance Bob Alexander has made it clear that the concept is bound to be re-examined in future (Healthcare Finance March 2010, page 13).

Some believe that the planned same day approach could be applied on a specialty by specialty basis and a combined outpatient procedure/day case/emergency short stay tariff for maternity non-delivery events could provide a way forward.

However, there is a concern about the financial impact of any centrally imposed solution. According to 2008/09 reference costs, the NHS undertook nearly 890,000 non-delivery HRGs at a combined cost approaching £500m. Some form of phased solution may be needed to avoid overnight winners and losers and general financial instability.

More change is also on the horizon within the maternity HRG4 chapter. For a start in 2008/09 reference costs, a new HRG was introduced – NZ10Z (diagnostic and therapeutic procedures on fetus). Previously such an intervention would have mapped to the full investigation admission HRG NZ09. This will lead to a tariff for these procedures from 2011/12. And beyond that the 2009/10 reference costs collection signals more significant revisions.

The delivery HRGs are changing to reflect the different resource implications of deliveries requiring epidurals, inductions and post-partum surgical interventions.