The Uncorrelated Primary Care Law – do General Practice Budgets reflect Need in Bristol 2011-12?
Is there any good evidence that practices with higher core incomes reflect patient need?

I will be looking here at data for the 48 PMS and 7 GMS practices in Bristol, using the 2011-12 finance data as this precedes an initiative by Bristol PCT to reduce inequity in GP income. Need and performance data was obtained from Bristol PCT or from the NHS IC information portal NHSIC. We already know that there are wide variations in GP income per patient in Bristol, ranging from £59.7 to £120.1 per Carr-Hill patient per year (Carr-Hill provided a method to adjust list size for demand).

GP practices , both GMS and PMS, already have incentives with proven ability to improve performance , through QOF and enhanced services. In this work I look for evidence that the wide range in core budgets shows any relation to need.

I will use total core practice income from GMS or PMS sources in each of the figures presented. I have excluded income from QOF, premises & IT payments, income from various extended services schemes and the prescribing incentive scheme.


It is useful to note that in all of the following graphs, 6 of the 7 GMS practices are clustered around the £60 mark on the x axis. This report will concentrate mainly on need, using easily obtained practice data on age( elderly and young children), deprivation and chronic diseases. A further analysis will look at income and performance. It is not always easy to say whether any variable is purely need, as the first example shows.

QOF income
QOF income per patient has need (through prevalence) and performance elements and rises as GMS/PMS income rises (Figure 1), so if practices are low earners, they appear to find it harder to hit high QOF scores. There is quite a wide variation in QOF performance. ( note : in 2011-12, PMS practices had 100 points automatically deducted from their maximum QOF target (1000 points) ). The data imply that practices with a low basic income lose more ground to higher earning practices through QOF. Perhaps a higher basic income enables practices to provide better systems and services and thus hit higher QOF targets. Perhaps better practices have managed to obtain a better core contract and also provide better care. There are some significant outliers. There is no evidence that low core earners can make up that loss by doing better at QOF. 


Figure 1 : QOF income vs practice income, R=0.2

Need and income
Deprivation
There is a small negative correlation between practice income and the Index of Multiple Deprivation (IMD, fig 2). Intuitively, it would be better if practices in areas of high deprivation were better resourced. Carr-Hill’s intention was to reflect workload, and there is evidence that patients in deprived areas do not access health care(create demand) in proportion to their health need as much as those in less deprived areas do.
Figure 2, R=-0.04

Elderly
There are less elderly patients in the practices with high incomes. There is a group of practices with high proportions (7-8%) of elderly patients who also receive low core incomes, £60-80 per Carr-Hill patient.

Figure 3, R= -0.30


Young children
Again, a disappointing negative correlation..., despite inclusion in the Carr-Hill adjustment.
Figure 4, R=-0.21

Severe mental illness
Not much correlation between income and prevalence of severe mental health problems. This graph will rely on the coding of severe mental illness, a QOF category which will generate some practice effort and income. It is disappointing to find that practices with high prevalences do not tend to get more core resources.

Figure 5, R= 0.01 

Depression prevalence aged > 18y
Again relies on QOF coding. Depressed patients (if they were feeling positive) might have expected their practices to be better funded.

Figure 6, R= -0.19

Dementia prevalence %
Do high earners not try so hard with coding cases for QOF? The same group of 4 low earning practices at the lower end of the core income scale.

Figure 7 , R=-0.30
COPD
COPD has quite high prevalence in some parts of Bristol, due to the (largely historical) tobacco industry. No evidence that the core money follows that need, although some increased other income would be received for high prevalence practices via QOF.
Figure 8 , R= -0.04

Overall
When we look at some of the available needs data, there is not much evidence that high core budgets reflect high need. Some of these elements will attract extra practice funding through QOF. Some of these elements have been chosen precisely because QOF enables data availability. We have been unable to show any significant evidence that core GP income in Bristol has much association with need. In some instances there are worrying negative correlations. These findings would appear to justify the PCT's attempts to redistribute core income, and they also support National efforts to move GP practice incomes onto a more equitable basis.

click here for PDF