Unwarranted variation – are we missing something?

Collectively, the Sustainability and Transformation Plans highlight the prevalence of unwarranted variation in the NHS today.  Many of the plans reference the negative effect of this on patient outcomes, and the overall cost to the NHS – estimated at £4.5bn in England.

But are we missing something?

The Atlas of Variation has revealed the variation in prescribing of key antibiotics across CCGs as being 4.0 fold at its most extreme and 2.5 fold when outliers are removed. We also have good data on cancer diagnosis rates in accident and emergency departments, revealed yesterday as 3.0 fold by the HSJ.

What really underlies these figures is a variation in GP Referrals.  Unless data is captured on referral rates and types, however, variation is extremely hard to analyse and address.

As Professor Jane Metcalf (deputy medical director, North Tees and Hartlepool FT) recently commented, ‘if people are not aware of what [the] variation is how can they compare themselves?’.*

This is the problem that many GPs face.

What can be done?

One approach is the implementation of a referral management service or single-point of access. By establishing a single pathway for referrals, it becomes possible to accumulate the data necessary to identify unwarranted variation. Once data has been collected it can be shared with referring GPs.

This enables people to compare themselves, a critical feature if we are to achieve progress in reducing unwarranted variation.

Referral management services are not simply designed to assess the appropriateness of referrals. If properly established, they should identify GPs who refer significantly less than their colleagues, and the data produced should form the basis of discussion. In many instances, the level of referrals from a GP may accurately reflect the patients they are seeing – but there’s a possibility that a GP does not refer enough for certain specialties.

By understanding GP referral rates for suspected cancer – and identifying any outliers – we can begin to address variation at the route. In doing so we should be left with smaller margin of variation in A&E cancer diagnostics – and potentially a higher rate of early detection.

Without the data produced from a single-point of access or referral service, the cost and process of understanding unwarranted variation in general practice is increased dramatically.


* Recent HSJ roundtable on unwarranted variation