Email your comments

The Patients at Risk of Re-hospitalisation (PARR) tool was commissioned by the Department of Health, developed with New York University and Health Dialog, and is now heavily promoted by the King’s Fund.

It is essentially number-crunching software, with data from the previous 4 years’ hospital admissions used to predict the likelihood of a further admission in the next 12 months. The latest version, PARR++, was released in November, and a launching conference was held at the King’s Fund. PARR is not the only system available: Dr Foster Information has developed a similar system, known as High-impact User Manager Tool, developed by Imperial College, and again with New York influence, and uses similar principles.

Haven’t we heard this before?
This sounds reminiscent of Evercare, charged with case management for those who had experienced two or more emergency admissions in the previous year. Evercare delivered improved quality of care for the targeted patients, but an evaluation failed to show that the model led to reduced admissions across the health community. Part of the reason may have been that the targeting was too crude, and that “frequent flyers” may regress towards the mean the following year – perhaps finally “sorted” by one of the admissions lucky enough to hit a specialist old age medicine ward – or indeed may die. The King’s Fund review in 2004 concluded that there was insufficient evidence to support case management in the UK. So have they changed their mind?

PARR++
The latest version of the tool is great fun. It has wonderful graphics to display the data, and can be easily manipulated before your eyes. It certainly provides a visual tool for inspecting patterns of admissions within a PCT, across age groups, diagnoses, different practices etc. It is free software and can be ordered from the King’s Fund website. It will import data from Excel or Access, but may require some IT expertise to get started. Although it acquires a large array of datafields, the strongest predictors for future admission remain age, diagnoses, previous admissions, and the re-admission rate at the local DGH.

As well as reviewing past admission patterns, it risk stratifies the population who have experienced an admission in the past. One can produce a list of the highest risk 3%, 5%, 10% or whatever, displaying their previous admissions and attached diagnoses. Rather worryingly, the system also quantifies and displays the recent hospital costs for individuals, and predicts their future costs – but is this going too far for an NHS? Perhaps it carries the potential for rather dubious practices ethically.

How does it work?
Many community matrons and case managers are using PARR to define their caseloads. Good examples are from Newham, where the community matron reviews the high risk printout each month with GP and nurse colleagues, to agree on who should be targeted. The community matrons use the principles of comprehensive assessment, similar to Evercare. They work collaboratively with an attached social worker, and they link with a broader long-term conditions management team – a network of specialist nurses. To this, they are also adding in telehealth systems.

Virtual ward
Croydon has pioneered the Virtual Ward concept. This model employs a systematic, data driven approach to the care of people with complex needs. It was adapted from the hospital model of multi-disciplinary team working. The PARR tool combined with GPdata (the Combined Tool) has been used to provide risk scores for each patient and those with the highest score are managed within the Virtual Ward, which is led by a community matron. Team meetings are held each day to discuss the plan for each patient. The team is composed of existing members of the primary health care team with additional new resource from the community matron and a ward clerk. Each ward has 100 ‘virtual beds’ and covers up to 10 practices. The Croydon team are now considering the addition of a matron with mental health expertise as this has been identified as a gap in the service.

Is it effective?
No new evidence base was presented at this meeting. There are 3 whole system pilots funded by the DH which are evaluating PARR with community matrons plus telehealth. Again, this sounds reminiscent of Evercare, and no doubt has received pump-priming. Evaluation will be over the next 2 years or so. The Combined Tool (PARR plus GP data) is a new development. It is likely that this tool is more accurate in identifying the highest risk patients within practices than the present demographic data. For example, the model was able to identify that within a population of patients with heart failure, only 48% were being prescribed a beta blocker. The model will enable resources to be targeted more effectively but also will identify those patients who have not yet experienced hospital admission allowing for more proactive, preventative models to develop. But as yet, there is no conclusive working example confirming the benefits.

But what can we say to the sceptics? Community matrons can deliver comprehensive assessment and improved quality of care to vulnerable and frail older people, and this must be welcomed. As in the example of drug therapy and cardiac failure, there will be many examples where these models can ensure the delivery of proven evidence-based care. Of course, the frailest, most vulnerable subjects require more complex care and clinical judgements than just delivering the evidence-based drug. Clearly there are good examples of where integrated care is being delivered, where the community matron is working with social care, and within a network of specialist nurses, and with access to secondary care. This is probably the nearest we can get at present to evidence-based care for frail older people living in the community and not yet in crisis.

The Primary and Continuing Care SIG within the BGS is always interested to hear of best practice examples, and certainly of local evaluations of these models of service.

Ian P Donald

BGS Newsletter, December 2007
Issue 14 ISSN 1748-6343 14