As patients demand more, it is vital to control variation while preserving truly personalised delivery, argue Geraint Lewis and Phyllis Shelton.
A heady cocktail of factors has the potential to drive up demand for treatment in today's NHS: rising patient expectations; excess capacity and workforce; falling waiting times; and a payment system that offers incentives to hospitals to do more work.
And the removal of barriers to entry into the NHS marketplace, with growing roles for the independent sector, new social enterprises and other third-sector providers, all have the potential to increase demand further unless balances are set in place.
Yet there is compelling evidence that some patients – particularly those in deprived areas – still experience deficiencies in certain aspects of healthcare. So one of the key challenges facing the NHS is ensuring patients receive the care they need, and no less, and the care they want, and no more.
Addressing this fundamental issue will enable us to increase equity and efficiency, leading to a more sustainable system.
Where differences are seen between the amount of healthcare resources one patient receives compared with another, the variation is either warranted explained by differences in clinical need or individual preference – or it is not.
Promoting warranted variation leads to truly personalised care. Reducing unwarranted variation could be the key to more equitable care and re-establishing financial control over a system where demand, costs and competition are all set to rise.
An increasingly competitive market with excess supply will keep stimulating demand further unless carefully controlled.
Unwarranted variation is seen where care that is known to be effective, such as the use of beta blockers after a myocardial infarction, is implemented to varying degrees in different parts of the country and by different clinicians. It also occurs where treatment choices that should be sensitive to patients' individual preferences bear little relationship to what individual patients say matters to them.
In both scenarios, it could be argued that doing nothing about the unwarranted variation not only wastes valuable resources but reduces the quality and personalisation of care. But what, in practical terms, can the health service do to tackle unwarranted variation?
Why is it that, among the top 100 US hospitals, the average number of intensive care days spent by patients in their last few months of life varies six-fold from one hospital to another? And in the UK, why do some GP practices request seven times more pathology tests per patient than others?
One cause of unwarranted variation is the so-called 'diagnostic-therapeutic cascade'. This is the whirlpool-like scenario that certain patients can find themselves being drawn into. Simply by virtue of being in frequent contact with the healthcare system, patients with long-term conditions are more likely to mention new symptoms and to have these investigated and treated.
For example, the new combined predictive model shows that patients with a long-term condition are twice as likely to have a diagnosis made relating to a back problem and twice as likely to undergo imaging of the back; and there is a trend towards having more back surgery. The value of such investigations and treatments is questionable. They drive up costs and expose patients to the hazards of the anaesthetic room and operating theatre for little apparent benefit.
A second source of unwarranted variation is seen in preference-sensitive care. When assessing the cost-effectiveness of a particular treatment it is usual to quote the benefits, called the 'utility value' in terms of quality-adjusted life years (QALYs). For many procedures, however, the personal utility of a particular outcome or side-effect varies dramatically from one patient to another.
For example, the three standard treatments for angina – medical therapy, angioplasty and coronary artery bypass grafting – have broadly similar utility values. But stark differences in the profile of potential complications and side-effects for each treatment make the decision highly preference-sensitive.
Professor Al Mulley, chief of the general medicine division at Massachusetts General Hospital, cites the hypothetical example of two patients with different lifestyles.
'Let's imagine two patients who have angina,' he says. 'One is a keen golfer; the other an avid chess player. For the golfer, the dramatic resolution of angina symptoms that can be expected after bypass grafting would be highly attractive. But the chess player is likely to be put off surgery because of the risk of cognitive impairment experienced by many patients following this operation – and so might choose medical therapy instead.'
The risk-benefit trade-offs of having bypass grafting versus medical therapy are quite different for these two patients – even though the aggregated utility of the two options are similar. In other words it would be misleading simply to use the off-the-shelf, average QALY values when advising these two patients about treatment. A much more personalised, preference-guided approach is needed.
A problem shared
Through their work for not-for-profit organisation the Foundation for Informed Medical Decision Making, Professor Mulley and colleagues have become leading authorities in shared decision-making. This is the process of transferring understanding and skills to patients so they are able to make preference-sensitive decisions based on their knowledge about their condition, the available options, personal values and preferences.
The foundation produces a series of decision aids, including booklets, films and web applications that support patients facing a wide range of preference-sensitive decisions.
In the US these resources are offered in conjunction with telephone-based 'health coaching' to millions of people. In the NHS this kind of service has been operating in Norfolk for two years under the management of Health Dialog UK.
Evidence suggests shared decision-making helps patients and healthcare systems in a number of important ways. It reduces unwarranted variation by improving patients' comfort with decisions and reducing decision conflicts; it strengthens relationships between patients and their clinicians; and it enables truly informed consent.
Crucially, however, there is overwhelming evidence that where patients are better informed and supported in preference-sensitive decision-making, on average they choose less invasive options.
For example, in a systematic review of seven high-quality trials of decision aids relating to major surgery, six showed a 21-44 per cent drop in the uptake of invasive surgery – with no adverse effect on health outcomes where patients were offered a patient decision aid.
The authors found that 'one trial that showed a non-significant trend toward increasing the rates of surgery also had the lowest rate of surgery in the control group (2 per cent). This was a UK study that had low referral rates by GPs to surgeons, because of a shortage of urologists.'
These findings illustrate how patient decision aids can be used to reduce unwarranted variation – including boosting activity where background rates were too low.
At Norfolk primary care trust, both patients with long-term conditions and others facing a preference-sensitive decision – such as whether to have surgery for chronic back pain – are offered shared decision-making and telephone-based health coaching to help them make more informed choices, improve self-care and enhance the patient-doctor relationship.
Norfolk GP Dr John Sampson feels strongly that working with patients in this way helps reduce unwarranted variation and can allow healthcare expenditure to be managed.
'This is integrated, patient-centred care,' he says. 'The patient is in control of their condition and results from the Norfolk practices are showing better use of NHS resources with reduced inappropriate use of hospital care.'
If we are to achieve a sustainable, balanced and equitable NHS we must look towards eliminating unwarranted variation. This can be achieved by supporting patients to be co-producers in their own health and care.
There are two practical steps that PCTs could take. First, they should identify occurrences of the diagnostic-therapeutic cascade in their population and take steps to minimise it. The whirlpool of increased activity facing patients already in the system can be predicted in advance by using the NHS combined predictive model.
The second practical step that PCTs could take is to stipulate that they will only fund preference-sensitive procedures – such as angioplasty, coronary artery bypass graft, prostatectomy for benign prostatic hyperplasia – where it can be demonstrated that the patient concerned has been offered a patient decision aid and support as part of the process of informed consent.
The NHS spends millions on these procedures, and the payment by results system encourages increased intervention through supplier-induced demand. There is a clear case to be made for investing a few extra pounds per patient in decision support that will lead to truly informed consent, more personalised care, and huge financial benefits for local health economies.
© 2007 Emap PLC