Feature / Perceived wisdom

Birmingham Children’s Hospital NHS FT had to address the actual and perceived accuracy of its patient cost data before it was accepted as a useful, wide-ranging tool to support the development of patient care. Steve Brown reports

A few months into 2013, it may be painful for some to be reminded that gym membership is not the same as going to the gym. In much the same way, having a patient-level costing system is not the same as undertaking patient-level costing and using that data to underpin decision-making.

This was the case for Birmingham Children’s Hospital NHS Foundation Trust until 2010. They were members of the patient-level costing club – ticking the ‘implemented’ box on the annual Department of Health patient-level cost survey – but it was not an active membership. There was no real commitment to the cause.

‘We had a system,’ says trust head of costing Julia Gray. ‘But it was in the background. We’d just hovered around patient-level costing. Then finance director David Melbourne [now interim chief executive] decided the trust was going to fully engage and ensure it had full board support.’ The epiphany for the trust was a case involving two girls with lesions known as arteriovenous malformations – cases where treatment options included significant surgery or a new, expensive drug. The finance team was asked to work up the financial case to demonstrate to commissioners that drug treatment was not only clinically beneficial, but the better financial option too.

Ms Gray remembers a very labour-intensive, time-consuming process that needed the involvement of the costing team, service managers, clinicians and financial management. But the case was built to demonstrate that the drug treatment approach did indeed deliver both the best outcomes and the best long-term financial costs.

‘This was a really worthwhile effort, but from a business perspective, you want to offer the same gold standard to as many patients as possible.’ she says. ‘And if you are going to do that, you need to be able to turn the information around much more quickly – ideally making it available at the desktop.’

This is where patient-level costing comes in. ‘With patient costing in place, if we want to undertake this kind of pathway comparison, we can look at previous patient “bills” and compare with costs associated with a different pathway,’ says Ms Gray. ‘Crucially, we can do this quickly and link back to clinical outcomes.’

The decision to raise the profile of patient costing and roll it out across the trust was followed by a series of practical steps. Accurate data is key with patient costing if it is to be used for local decision-making. But this breaks down into two elements, Ms Gray says – actual and perceived accuracy. Both must be dealt with. ‘In terms of actual accuracy we took information from all our main databases and visited all specialties to find out what patient-level data they collected – whether in dedicated systems, spreadsheets or in notebooks,’ she says.

So, in addition to the main systems such as outpatients, inpatients and emergency department attendances, the trust had data on tissue typing, stem cell transplants, intensive care transfers and orthotics. The next task was getting the data into the costing system and validating it. ‘We’ve developed a complex system of matching rules. So, for example, at a high level certain cardiac information can only be matched to patients with specific procedure codes or intensive care transfers only match to an inpatient.’

Process of refinement

In reality, matching has been a constant process of refinement. Coders give the costing department masterclasses in coding – showing how codes relate to the data seen by the finance team. But involving clinicians has been key.

Ms Gray says: ‘We have clinical champions – in particular an assistant medical director for quality with responsibility for patient costing. We have monthly meetings where we go through patient bills and check the matching, comparing what has been charged against a patient’s activity with the care described in the patient’s notes. He’s an oncology consultant, so we may look at whether chemotherapy drugs – known to be a difficult area in costing – are being allocated to the right patients. If not, can we refine the system further to ensure better allocation? We walk away with a list of fixes and review the impact at the next meeting.’

These meetings – and others with clinicians as part of patient costing user groups – tend to focus on the five most complex patients in the previous period. The complexity of the cases means they will cover a lot of activities, giving an opportunity to test whether allocations and matching are working in wide-ranging areas.

One-stop shop

In addition, there are bimonthly meetings of a patient costing clinical leads development group. Ms Gray says patient-level costing brings lots of different issues together – variations in treatment approach, variations in performance, coding and data issues as well as costing. So this working lunch with IT, coding and finance staff provides a one-stop shop for clinicians to raise issues on related issues.

Clinical leads also receive quarterly patient cost reports with cost highlights for their specialty. This brings together costing data for performance, financial and patient analysis, setting out, among other things, theatre costs, costs of ordered tests and expensive patients.

Like other trusts implementing patient costing, Birmingham has had to work hard with clinicians to get the data as accurate as possible and to get them to accept it. Ms Gray says the costing team’s motto is ‘persistence will prevail’ – if clinicians dismiss the data, you need to fix the problems causing errors or demonstrate the data is right. But in general, she says, doctors like the evidence base behind patient costing and each success with the data can help to further increase confidence.

One clinician argued that the trust was being underpaid for combined liver and kidney transplants. Although the activity was rare, the doctor said costs had never been right. The team worked with the clinician to refine allocations and improve costing and produced costs to back up the clinician’s argument. The information was passed to the contracting team, which used it in discussions with national and regional commissioning groups to revise funding in this complex area.

The positive result not only improved funding for the trust, but demonstrated the value of the patient-level cost data. User group meetings have also produced wider benefits – for example, picking up on documentation issues. In one patient bill review exercise, it came to light that comorbidities or complications were not being recorded for clinical haematology cases. While this affects income, it can also make the trust look expensive in benchmarking as its more complex patients will be hidden inside a healthcare resource group for a less complex casemix.

There have also been good gains in gastroenterology, where the cost data has been used alongside outcomes to inform development of standard trust-wide protocols for children with inflammatory bowel disease. And in cleft lip and palate, where patient-level cost data has provided the service with its first consolidated view of income and costs, the data has helped establish the amount of activity undertaken outside the payment by results regime, again helping to inform discussions with commissioners about non-PBR contracts.

Savings implications

However, Ms Gray is clear that patient cost data must also help the trust find ways to reduce costs as well as ensuring accurate payment. This can be a more delicate area. The chief medical officer is pushing a demand management initiative across the trust, looking at demand for tests. It is not the costing team’s place to challenge the validity of specific tests, but the patient cost data can highlight which clinicians are ordering most tests or using the most expensive tests. This provides a starting point for understanding usage.

The dermatology department has used the data to focus on timing of tests. With patient cost data highlighting some duplication of ordered tests, it was realised that a lack of co-ordination around the timing of a test and the follow-up appointment was often leading to clinicians ordering more up-to-date tests.

A change in scheduling practice will reduce tests and costs. Similarly, patient cost data alerted one oncologist to the high costs of one test typically ordered in a batch of five. Yet this one test – the most expensive of them – was in fact only needed if the other four indicated it was required. It is now only ordered if necessary. No reduction in patient care, but good savings on cost.

Overall, the response from clinicians is positive – they see patient costing as a supportive tool. Martin English, associate medical director for quality, says clinicians want to engage in processes that improve patient experience, outcomes and efficiency.

‘Patient-level costing has meant departments can identify where there are current costs in the system and quantify the major contributors to those costs,’ he says. ‘This can rapidly lead to audits and service improvements to reduce costs. More subtly, the same systems can be used to define the financial benefits realisation of any project that begins with the aim of improving experience and outcomes.’

Ms Gray adds: ‘They have confidence that if they voice a concern it will be put right and if they invest time, they will see improvements. We wanted it to be a broader tool than just costing and we think that is the way it is seen – giving transparency around coding and documentation and supporting work around clinical standardisation, validating income and cutting costs.’

For Birmingham at least, being a member of the patient-level costing club looks set to deliver value.