Feature / Making a mark

Benchmarking built on robust patient-level cost data can be a powerful tool in the search for improvement. Suzanne Robinson reports from cancer specialist the Christie NHS FT

The challenge to keep improving – both in terms of service quality and efficiency – gets harder. It would not be unusual for trusts with years of improvement behind them to be asking how they can improve more. At the Christie NHS Foundation Trust, this drive for efficiency has been helped by the development of patient-level costing data, which has enabled the trust to undertake robust benchmarking to identify real opportunities for improvement.

The Christie is a specialist cancer centre with the largest clinical trials unit in the world and treats rare, complex cancers. Making the best treatment available to patients through technological advancement, research and innovation is at the heart of what the Christie does.

But cancer care isn’t cheap. Understanding the cost base is key to the effectiveness of the organisation. The trust has pursued better costing through a patient-level costing and information system (PLICS) and introduced a service line management structure.

Christie director of finance and business development Ian Moston says: ‘It is how organisations use the information to improve services that is important and getting real clinical ownership of the data is critical to delivering any change.’ Two and a half years after going live with its patient-level system, he says, the trust is seeing sustained benefits.

Part of achieving recurrent benefit is about embedding PLICS information into day-to-day business. The Christie believes it has done this through the following:

l Monthly performance reviews provide an opportunity to challenge divisions on their high-cost healthcare resource groups (HRGs).

l The distribution of income is based on cost allocation from PLICS, providing an incentive for divisions to validate methodologies used.

l A proportion of cost improvement programme (CIP) targets are based on service line position – those with an adverse position have more opportunity to make savings.

l A trust-wide Directory of care covering all patient pathways enables the benchmarking of patient-level activity against national or locally agreed treatment protocols. It allows variations – the number of scans per patient, say – to be challenged within clinical teams.

l All trust business cases that involve investment are required to include an analysis of the service line financial position using PLICS. The trust will not invest in loss-making services without a full understanding and rationalisation of the drivers.

Building on its costing work, the trust joined the Patient Cost Benchmarking (PCB) Group run by Albatross Financial Solutions in late 2010. It was keen to take PLICS to the next level, recognising that an organisation can only go so far with internal assessment of data. To really understand whether the trust was simply inefficient or justifiable costly, another layer of comparison was needed. 

The benchmarking group is used by more than 25 NHS trusts and the numbers are increasing monthly. The data has been used to support national projects undertaken by the Department of Health, Monitor and the Information Centre. The analysis from this work enables comparison by cost category, direct/indirect/overhead and activity measure (length of stay and theatre times).

From a clinical coding perspective it facilitates a better understanding of variation in practice by comparing procedure, diagnosis and HRG comparison across peer groups. And there is consistency in cost definition, which comes from working with the user group and HFMA clinical costing standards group.

Using the data

The trust has developed a simple step-by-step process to identify the right HRGs to focus on (see chart, previous page). This uses many of the reports available from the PCB group.

There are several variables that could affect a loss-making HRG and it’s vital to eliminate all possibilities. By using a hypothesis testing approach and an open mind to the results, the following are considered:

• Costing Is it accurate? This requires clinicians to understand and validate PLICS outputs.
• Coding Case notes are analysed with a multi-disciplinary team including information, finance, coding, service managers and clinicians to validate the accuracy of coding.
• Grouper Technical experts sense check the behaviours and trigger points of the local payment grouper.
• Specialism Once the above have been tested, the analysis into the cost drivers can be assessed. The trust’s Directory of care is used to identify spells with high volumes of tests, scans and treatments. (The possibility that the national tariff underfunds activity or fails to recognise complexity is also considered.)

The haematology and transplant unit at the Christie is one of the largest centres in the country, specialising in the treatment of haematological malignancies. The unit sees more than 1,200 new patients and in excess of 9,000 outpatient cases per year.

Although the department can demonstrate some of the best clinical outcomes in the UK and has an excellent track record of positive patient experience, the financial position appears less favourable. With income of close to £3.7m across all points of delivery, it makes a loss of £2.9m, with its emergency work making the biggest contribution to this.

The service was struggling in terms of capacity, with rising demand and the need for investment to rationalise and improve facilities. The numbers from PLICS suggested the service was running at a loss. And with a corporate requirement for all business cases to include a PLICS financial assessment, there would be little hope of board approval without a full understanding of what was driving the loss.

Complexity and length of stay are the key considerations when assessing why income is insufficient to cover costs, but the hard part is proving that. And even where complexity explains some of the issues, the need for efficiency cannot be overlooked. Any opportunities to be more efficient have to be realised. This is where benchmarking has helped.

Sarah Vassie, income and costing manager at the Christie, has been using the benchmarking reports alongside other analysis. ‘The reports allow us to compare data from a large number of organisations across the country, helping to pinpoint where the trust differs from the average. This can be used in discussions with the trust’s managers and clinicians,’ she says.

For example, in the SA sub-chapter (haematological disorders), analysis of comparative data across the PCB peer group sample showed the Christie was more expensive than the group average in 17 HRGs, only beating the average on nine HRGs. The result was that the Christie was more expensive than the group average.

Looking at just one HRG for admitted patient care (SA25F – acute myeloid leukaemia without CC) suggests significantly higher unit costs than the benchmark group average (see table above). This savings potential (£642,000) provides a starting point for further analysis. The data can be further cleansed. Significant outliers can be removed (those with very long lengths of stay for example).

In this case, there were also differences between the way the Christie and the other group members recorded regular day attenders (recorded as outpatients at the Christie). This raised the average unit cost across the PCB group. The further table (left) shows a more accurate assessment of the savings potential for this single HRG – still a substantial £376,000.

To make this sort of assumption it’s important to understand the reasons behind the variance. Again this is where clinical engagement is key. This report has generated much debate with clinicians keen to understand why activity varies from trust to trust. This could suggest there is inconsistency in how the activity is counted, which could in turn dilute the costs of this very specific HRG.

The clinical haematology teams use the most advanced methods to treat patients with haematologic or malignant diseases. Patients in treatment are often compromised following extensive cycles of intensive chemotherapy and so more vulnerable to infections and complications. The CC (with co-morbidities or complications) HRG isn’t always sophisticated enough to recognise this level of complexity and in some cases the detail isn’t recognised by the grouper, pushing some activity back to the without CC HRG. There is a definite case for more granularities within tariff, which the benchmarking group is looking into.

The numbers of diagnoses (complexity/co-morbidities) or number of treatments (complications) at the Christie differ from other trusts. Could this be the reason why costs appear high? Analysis (see graph top right) shows that, compared with the benchmarking group average, the Christie has a higher percentage of diagnoses for the SA01F HRG. This suggests a correlation between diagnosis and higher cost and that the Christie has a higher level of specialism within this HRG than the rest of the group. And if there is a more complex diagnosis why does the grouper not recognise this?

The bar chart (above) shows how cost behaves in relation to income and the number of diagnoses. As the number of diagnoses increases so do the costs – but income remains constant. This could be problematic for an organisation where the majority of patients are complex in nature.

The example above demonstrates how benchmarking can be used to facilitate a trust’s understanding of its costs, income and casemix. By drilling down to a point of delivery analysis, comparison can be made on clinical practice. This has helped the  trust identify activity that can be undertaken in an outpatient setting rather than admitted patient care. By removing these costs as part of the business case, the service line loss would reduce by 15% and there would be a lower charge for commissioners.

The clinical teams at the Christie are now far more aware of the consequence of cost in making business decisions and how changing their clinical practice at a granular level can translate to big financial savings down the line.

In addition, the work has provided a much greater appreciation of how complexity and cancer is reflected in national tariff. It provides valuable evidence to support how tariffs of the future might be structured. This is not something that is going to be addressed overnight, but it’s important that the NHS as a whole understands the relationship and ensures a national tariff of the future reflects the complexity of the care.

Adrian Bloor, consultant haematologist at the trust, argues that ensuring income reflects the true costs of the delivery of care is a challenge for both clinical and financial teams.

‘In my practice [high-cost complex interventions] this has been particularly problematic when using a tariff derived from cost averaging across institutions with widely varying practices,’ he says. ‘Utilising PLICS has resulted in a far more detailed understanding of the costs of treatment. This information provides a detailed and transparent foundation to develop reference costs that can be directly compared between hospitals and can hopefully be used to shape the tariff.’

But he adds that close interaction between finance and clinicians is an essential part of this process. ‘[It is needed] to understand the practice, sense check assumptions and outputs and maximise quality,’ he says.

There have been many benefits for the organisation from its involvement in this work. The first relates to costing. The Christie has a much greater understanding of its services from a patient cost perspective. PLICS has helped to create meaningful, easy to understand, reliable comparisons to better understand variances. As a result the trust has improved its clinical coding and recording of activity and changed some of the methodologies in the PLICS system.

In addition it has created a strong network of colleagues through the benchmarking group – it is refreshing to find organisations so willing to share information.

It has also enabled teams to engage with some of the important developments taking place across the NHS. It has created opportunities to work on initiatives such as Project Diamond – a project involving specialist trusts that’s focused on funding risks in London and examines the relationship between cancer care costs and tariff income. The third benefit has been much greater clinical engagement. This has been encouraged by creating a meaningful output in a language that clinicians understand with comparators they can challenge.

Next steps

There’s much more to be done, and potential for even more in the future, but it’s important to keep on track. Over the coming months the benchmarking group will be focusing on developing an acuity/dependency tool with the aim of recording differences in nursing and staffing resources at patient level. There are also plans to introduce diagnostic test benchmarking, which could provide outlier organisations with a best practice marker for clinical teams to work to.

For the Christie there are also real benefits in linking this work to the trust’s Directory of care, where national benchmarking data can be used to compare to local treatment protocols.

Suzanne Robinson is head of income, costing and contracting at the Christie NHS Foundation Trust