Intended for healthcare professionals

News NHS Postcode Lottery

Patients with type 1 diabetes are missing out on flash glucose devices, finds BMJ investigation

BMJ 2018; 363 doi: (Published 07 November 2018) Cite this as: BMJ 2018;363:k4675

The postcode lottery isn't the biggest problem: the incoherent CCG criteria are much more worrying

Postcode lotteries for care always cause great angst in NHS watchers. But that isn't the biggest problem here. The incoherent criteria written by CCGs and their medicine optimisation committees are far more worrying as they will severely damage our ability to make better decisions in the future.

Part of the postcode lottery happens because some CCGs have not adopted any standards at all for prescribing the Freestyle Libre devices. Other have adopted a variety of strict criteria (like the North of England RMOC example used in the article, but often much worse and less coherent).

But few of those clinical criteria make any sense. They look like a set of clinical excuses cooked up after a financial decision to prescribe Libre as little as possible. They certainly don't look like they are targeted at the groups of diabetics most likely to get short and long term benefits from using Libre (and therefore save the CCG the significant future costs of dealing with higher levels of complications).

For example, most type 1 diabetics don't test their blood glucose often enough (we don't issue enough conventional test strips for all type 1s to come close to the 8-10 tests a day that are required for decent control and the typical diabetic is probably testing only 3-4 times a day with many testing even less frequently). But most criteria start by saying Libre should only be prescribed to those already testing at least 8-10 times a day. And the criteria then impose further limits based on the frequency of serious hypos or DKA hospitalisations. Surely the right target group is the people who would avoid ever being in such serious trouble if they tested their glucose more frequently now? In other words more or less the opposite of the current criteria.

Some CCGs argue that they are waiting for better evidence of the benefits from continuous testing. But the prescribing criteria make it harder to get that evidence as they don't target the groups most likely to benefit. Even if we learned about the benefits to the narrowly defined eligible groups that would tell us almost nothing about the benefits to the majority of type 1 diabetics.

NHS Clinical Commissioners argue, correctly, that "Clinical commissioners have a responsibility to consider the needs of their whole populations, reduce inequalities, and improve quality of care while living within the funding they are given..." which would be fine if that is what CCGs were doing. If their criteria said "we don't know the long term costs and benefits of Libre so we will hold off large scale prescribing until that evidence is clear" then we would have a basis for improvement and a clear set of criteria about what improvements we would have to demonstrate to show that Libre is a major benefit for a limited cost. But the current guidance limits any ability to learn about the benefits for most diabetics by setting ridiculous, incoherent criteria for current prescribing. Criteria which are dishonest in pretending that the rules are clinical when they are actually financial.

If the issue is whether the benefits are large compared to the cost when compared to other treatments then we can have a sensible debate comparing Libre to, for example, warfarin replacements where prescribing costs have risen in the last 5 years from ~£20m/year to over £300m/year because new drugs are 20 times the cost of warfarin (and for modest gains in outcomes). The worst case cost of Libre versus current costs (taking not account savings from lower use of conventional blood tests) are probably £100m-£150m for the whole of England.

Libre is an odd case in cost terms as well which none of the CCG guidance has recognised. IF diabetics used current blood-glucose tests as much as experts think they should (say 8-10 tests/day) then Libre would save money. The only reason why Libre would be more expensive is because most diabetics don't use conventional technology as much as they should because it is painful and inconvenient. It seems an odd goal for the NHS to rely on patient non-compliance to keep costs down, but that is the implication of current guidance.

Yes, a postcode lottery is bad. But incoherent prescribing criteria by CCGs based on spurious clinical criteria is much worse and will have longer lasting and damaging impact on the health of all their populations as it entrenches bad practice in how to make good decisions when new drugs and technologies appear.

Competing interests: self funding Libre user.

08 November 2018
stephen black
data scientist
black box data science
biggleswade, bedfordshire