Feature / On the level

Patient level costing can play a key part in delivering the quality and cost improvements critical in meeting the productivity challenge in the NHS. Steve Brown hears about one trust’s experience implementing the programme

David Nicholson’s £20bn challenge for the NHS will demand that significant levels of savings are released from frontline care. The government’s own estimates for running costs (commissioning) only amount to £5.1bn. And hitting the savings target – set out in the latest operating framework – to reduce these costs to £3.7bn over the next four years will only contribute £1.4bn to the overall total. The simple fact is that the bulk of the savings will need to come from clinical services – eliminating unnecessary referrals, investigations and treatment, optimising pathways and tackling avoidable variations in approach.

There is no argument that clinicians need to be heavily involved with this process (although there are differences of opinion about exactly how this involvement should be taken forward). But clinicians cannot take the kind of decisions they will need to take – eliminating waste and designing better pathways – without support. They will need detailed information telling them the most likely places to look for improvement opportunities and a detailed understanding of the current make-up of costs incurred by existing activities.

That is where patient level costing comes in. It can enable information to be organised at a level clinicians easily relate to – the patient. In cases where costs are above average or appear significant outliers, it can provide the ability to understand why. This can help clinicians working alongside managers to understand whether these costs are appropriate or perhaps offer an opportunity to improve services for patients or reduce costs.

Southampton’s goal

This was the goal for Southampton University Hospitals NHS Trust when it took the patient level costing plunge back in 2006. It had already organised itself into a more business-like organisation – reporting on a service line basis – and felt it already had the clinical and management engagement in place. It simply didn’t have the right tools and information to help it meet its joint objectives of financial control and care quality optimisation.

Fiona Boyle, the trust’s finance manager and leader of the patient level costing implementation project, says the problem with service line reporting was that the existing reports were static.

‘We couldn’t drill down to interrogate the data and get the data we needed to improve services and performance,’ she says. ‘It didn’t enable us to understand the profitability position, where and why we were profitable and where and why we were making a loss.’

Finance and investment director Alastair Matthews arrived at the trust in 2007 with a background in the private sector. He admits he was struck by the contrast between the NHS and the commercial sector. ‘I was surprised the NHS didn’t know the profitability of its various activities,’ he told the HFMA annual conference in December. This was something he was keen to put right at Southampton.

Southampton’s patient level costing solution has been built on the Ardentia system, commissioned in 2007. Full roll-out was achieved in 2010. Ms Boyle admits the implementation took longer than anticipated. She puts this down to several factors.

There was work to do on the trust’s own patient administration and various feeder systems – which Ms Boyle says ‘hadn’t been used in anger’ – before linking in with the new costing system.

The trust had also chosen the Ardentia system in the knowledge that it still needed developing. This provided good value for the trust, but implementation was always going to be slower as a result. Subsequent implementations at other trusts have been faster.

Even with full board and senior management buy-in to the project and dedicated resources – a pre-requisite for success according to Ms Boyle – there were distractions. Reference cost submissions, for instance, still had to be made and relied on many of the same staff.

‘In the first year we double ran,’ she says. ‘We did reference costs on our own system while implementing patient level costing separately.’

Staff time is unquestionably the biggest investment the organisation has made in the project. It estimates total running costs of about £500,000 over four years. And some of the tasks – mapping finance data to core activities for instance – were far more time consuming than had been anticipated. Some of the costs have been non-recurrent; others have now been built into general running costs.  For instance, in addition to Ms Boyle working half time on the project, there were four band 5s assigned to the project – one in finance and three in IT. Funding for two of the IT roles and Ms Boyle’s half-time contribution stopped with the roll-out of the project.

Rules of engagement

The time taken can also be attributed to a determination to get it right and not launch prematurely. At the very outset the trust drew up a ‘rules of engagement’ document with clinicians and one of the key commitments was that the systems and data would be properly tested. ‘We made it clear we would only use the data when it meant something to the clinicians,’ says Ms Boyle.

The trust acknowledges there were times when it was tempted to push some of the patient level data out into the trust, to get people working on some of the apparent variations in costs, practices and pathways. However, it resisted and continued to work with clinical champions to ensure the data was as robust as possible and resistant to the ‘I don’t believe the data’ response that has sometimes characterised attempts to use reference costs data in the past.

‘We knew that first impressions were very important,’ says Mr Matthews.

Ms Boyle suggests the simple process of implementing patient level costing has benefits in itself, even before the costing outputs are put to any use.

‘We uncovered data problems and corrected them,’ she says. ‘Although patient costing systems need good data, they also help you to create good data.’ For instance, as part of the project, the trust significantly improved its neurology theatre data. This gave it better utilisation figures, which were useful to management even before the data was fed into the costing system.

In fact, there have been several success stories throughout the project – some with a financial impact. For instance, the high costs incurred on some orthopaedic patients highlighted that coding drift had led to some activity not qualifying for a specialist service top-up that the trust was entitled to.

It has also helped improve the trust’s reference costs submission, with the added benefit of being able to drill down into the data for the first time to understand variations compared with peers or the national average.

But what about informing decisions to eliminate waste or drive productivity, one of the key goals for patient level costing?  Here too there have been some early successes – despite the fact that clinicians only received their first data in September last year.

The data has highlighted variation in costs at consultant level. Some of this is explicable – for instance, a less experienced surgeon taking longer in theatre. But there were other cases of surgeons using more expensive prosthetic implants than others. When presented with the information, the relevant operational manager acknowledged it was an issue that was already known about. However, the patient level costing data provided both the opportunity and the evidence to actually push through a change in practice.

Length of stay

In another example it was spotted that an oncologist had high costs driven by a very long average length of stay. Further investigation revealed that the oncologist was only on site one day a week and was working off site the rest of the time. There was no robust discharge process in place in his absence and so patients were being kept in hospital until he was back on site. Now a process has been put in place to enable colleague or nurse-led discharge. This provides a three-way win – the patient is discharged home earlier, the trust frees up beds and reduces costs, while the PCT minimises the chance of paying excess bed day payments.

Ms Boyle accepts that some of these decisions could have been informed by comparison of activity data – comparing length of stay information by consultant rather than attaching costs to all parts of the patient episode. But she suggests that having a financial currency really helps.

‘People become very aware when the money is attached – there is something about looking at the financial implications that gets people’s attention,’ she says.

The trust has made good progress. Patient level costing data is now used in the board strategy and finance discussions. There have been discussions about profit- and loss-making areas and broad causes. The costs have been used to set some local prices (outside areas covered by the national tariff) in the current year. Clinical champions have used the data to examine ‘average’ patients on normal pathways and look at the outliers. But, even so, the trust recognises that it is still in the early days of using the data.

Some of the conversations are only just getting under way. And while clinical reaction to the data has been generally positive, the trust knows there will be some challenging groups and some feisty conversations ahead.

It also knows that there are still data quality issues that need to be resolved – it will after all be an iterative process.

But the trust is already happy with its decision to pursue a patient level system. ‘Service line reporting without patient level costing just identifies the problems but doesn’t give you the solutions,’ says Mr Matthews. ‘This provides a common language for clinicians, planning and finance. ‘

He concludes: ‘You shouldn’t underestimate the time and resources needed. But the prize is worth the investment. I can already see the payback today.’