Written evidence from Central, North and South Manchester CCGs and Manchester City Council (PEX 0008)

This evidence relates to resource allocation in the NHS and the inquiry consideration of: ‘The effectiveness of the mechanisms by which resources are distributed geographically in the NHS’.

SUMMARY

A. CCG Allocations

1. We believe that the NHS England’s duty on health inequalities should be increased from ‘..reducing inequalities in access to and outcomes from healthcare..’ to the former duty on the resource allocation policy of the Department. of Health to ‘..reduce avoidable health inequalities..’. NHS England are open to such an expansion.

2. From 2009/10 to 2011/12 a health inequalities formula was added to NHS PCT resource allocation formulae and we believe a similar formula should be re-instated for 2014/15 allocations to Clinical Commissioning Groups (CCGs). If this is not done and the current utilisation formulae are used alone then resource share will move in the ‘wrong’ direction in terms of reducing health inequalities between CCGs i.e. from poor health to good health CCGs.

3. It seems generally accepted that utilisation models give only a very approximate measure of health need and additional formulae for unmet or inappropriately met need are therefore necessary in addition to adjustments for health inequalities.

B. Primary Care Allocations

4. Attempts to reduce inequity in the geographical distribution of GPs (and hence of primary care resources) in England since 1948 have only been partially successful. A standard method of measuring inequity indicates that it is still significant and favours areas of better health.

5. The current inequity in GP geography is likely to lead to variation in the quality of primary services and to variation in the effectiveness of CCGs.

6. The centralisation of the budget for GPs under NHS England provides an opportunity to manage initiatives to reduce the inequity in GP geography.

7. Possible initiatives to reduce the inequity include additional central funding for new GPs in under-doctored areas, greater incentives for GPs to work in these areas, and introducing a limit to the weighted list size per full time equivalent GP.                            

A. CCG ALLOCATIONS

1. CCG allocations strategy should include the reduction of health inequalities between CCGs in England

1.1 We believe that the quality and quantity of all aspects of healthcare in an area commissioned by CCGs influence the average health of the population of the area. This is particularly true as financial pressures affect staffing levels and influence rationing decisions. Although arguments can be made that some specialties e.g. community services and maternity may have a greater affect on population health than others, it might be difficult to argue for any speciality that increasing resources could not improve quality of outcomes, all other things being equal.

1.2 There is a very wide range of average health status between different areas (e.g. CCGs, LAs) in England which is an important dimension of health inequalities, and which is widely regarded as highly problematic. If, as proposed above, all aspects of healthcare in an area influence the average health of the population of the area, then the philosophy and strategy underlying CCG allocations should include the reduction of geographical health inequalities as a major aspect. This point seems obvious when stated baldly like this. But because many other issues affect health inequalities e.g. the local economies, lifestyles and prevention, it can get lost or relegated. The previous allocation strategy statement is clearer in supporting the reduction of health inequalities than the 2012 Act which depends on the interpretation of ‘reducing inequalities in …. outcomes from healthcare’, though as NHS England point out in Annex D of the Fundamental Review papers,[1] ‘While the inequalities duties on the board are limited to reducing inequalities in access to and outcomes from healthcare, there seems to be nothing to prevent NHS England from seeing its duties on inequalities as wider…’.

1.3. In view of the above points the ‘equal access for equal need’ allocation philosophy may not allow CCG allocations to help reduce health inequalities. A contribution from CCGs to such a reduction is likely to require above average healthcare in poorer health areas, and hence an additional allocation above ‘equal access for equal need’ for these areas.

1.4 We believe that although the public health budgets which moved to local authorities in April 2013 will have an important long term role in reducing health inequalities, this should not detract from the similarly important and more immediate role of CCG budgets.

(Also, ironically, the new Public health allocation formulae[2] move allocation share in the ‘wrong’ direction from poor health areas to good health areas so unless this is corrected these public health budgets will be less fit to reduce inequalities between areas in England than when they were subsumed within PCT allocations).

2. A formula for health inequalities should also include an element for ‘unmet’ or ‘inappropriately met’ need

2.1. It is generally accepted that poorer health areas are more likely than better health areas to have some needs which are not met by health services. Allocation formulae based on the use of services will not measure all need for services where there is ‘unmet’ or ‘inappropriately met’ need. It is therefore important to include additional formulae to allow for this in poorer health areas in order to actually get the starting philosophy of ‘equal access for equal need’, in addition to consideration of reducing health inequalities.

3. The DFLE formulae or a similar formula should be re-instated to avoid resource share moving in the ‘wrong’ direction

3.1. There is a history of adjustments to allocations for health inequalities and/or unmet need. 2001/3 allocations contained additional funds for health inequalities and 2003/8 allocations used a main formulae where it was considered that unmet need was accounted for within the formulae.

3.2 PCT allocations for 2009/12 included a health inequalities formula which was also considered to allow, to some extent, for unmet need. The formula was a simple one of 70-DFLE (Disability Free Life Expectancy) where 70 was the best DFLE in England. Because of the difficulty of precisely quantifying the issues ministers were given three options of the weighting to be given to this formula – 10, 15 or 20 %. 15% was chosen for years 2009/11. The new administration controversially reduced this arbitrarily to 10% for 2011/12, (though partly offset by expanding its range). Manchester expressed concern about this reduction to the Health Select Committee[3] which backed Manchester’s concern,[4] and the issue gained national media coverage.

3.3. The DFLE formula in 3.1 seems to have been quietly dropped for the Review working paper allocation figures presumably because ACRA did not consider the evidence underpinning it to be ‘robust’ (though the ‘robustness’ of the utilisation formulae used is also questionable –see section 4 below). This omission seems strange when it was supported by ACRA for three previous years and their current research has presumably not yet come up with an alternative formula. It is primarily this omission which has distorted the working paper figures1 away from deprived urban areas and towards wealthier areas, and also from North to South, and caused such consternation in many urban CCGs, including the three City of Manchester CCGs, which already have some of the worst health in England. See Table 1.

Table 1. Northern Core Cities (LA areas) in order of % loss in target allocation if utilisation models alone are used (2013/14 figures)

3.4. The Department of Health commissioned in 2010 a major academic study[5] on ‘the health inequalities elements of the NHS weighted capitation formulae’ which included unmet need. The study team consisted of twelve experts including Sir Michael Marmot. The recommendations from the study included formulae of similar strength to those used in the 2009/11 allocations. However the recommendations were not taken up which is puzzling. The final paragraph of the conclusion of the report indicated a number of options:

‘We have suggested potential improvements to the current health inequalities formula that might be adopted in the short-term based on the distance from best DFLE method. These are described above and include calculating the adjustment at the small area level and then aggregating it to PCTs, using a predicted DFLE measure (where DFLE is predicted as a function of social determinants of health), and using alternative measures for combining quality of life with the life expectancy measure. We have also suggested a number of adjustments to the current utilisation based approach. For the longer term we have recommended that ACRA consider a method for equalizing health between areas based on the relationship between funding and health outcomes.’

4. There is a need for judgment and adjustments when there are questions about the robustness of the utilisation models which underpin the CCG allocation formulae

4.1. Utilisation models, which underpin most current and past formulae, based on service use, are not exact like laws of physics because the system they are trying to analyse is so complex. They measure use which approximates to demand which in turn is assumed to approximate to need. But these two levels of approximation will themselves vary with geography via culture and supply of services.[6] Also the models vary with the detail of the modelling method chosen, the choice of variables, whether they use national standard costs or local costs taking length of stay into account etc. They have been subject to some fundamental academic criticism[7],[8] and as the detailed methods have changed over time the actual results at health authority level have been very variable, as indicated by Table 1 for the Manchester Health Authority/PCT/CCGs. In this table indices are allocations per head relative to England = 1 so that 1.25 means 25% more than the England average per head and 0.94 means 6% less than the England average per head. The Health Inequalities (HI) Indices would be multiplied by the formula indices to yield an increased index.

Table 2 General & Acute model indices for the City of Manchester

Note York and Area indices adjusted for changes in Manchester demography

These results call into question the stability of the utilisation models. Manchester’s relative health position over this period has been stable at the bottom of the England health league. The population has gradually increased its proportion of young adults relative to England but the older York and AREA figures in the table are adjusted for this. The York model was considered for many years within the academic community and the Department of Health to be the ‘gold standard.’

4.2. AREA was considered to be an improvement which allowed for unmet need. So the astonishing 22% drop in Manchester’s allocation index from AREA to CARAN doesn’t make sense. To retain credibility some explanation is required of this change and in particular why the latter two models are so much more weighted to age rather than deprivation compared with the two earlier models.

4.3. For comparison with the General & Acute indices, the mental health model (covering 13% of spend), also a recent PBRA type model though by a different group of academics mainly from the University of Manchester, gives very different results from the former and often in the opposite direction. For Manchester the General & Acute index of 0.94 compares with one of 1.54 for Mental Health. The biggest disparity is Lambeth CCG with indices of 0.67 and 1.92 respectively, a threefold difference. Taking all CCGs in England the Mental Health indices correlate weakly negatively (r = -0.30) with the General & Acute indices. Figure 1 gives the relationship for all CCGs and shows visually that there is only a weak correlation which is negative.

While the age profile of costs will reverse to some extent between the two specialities it is surprising, for example, that Manchester, with its characteristically highest mortality in England and very high deprivation, should have an index for Mental Health that makes sense and one for General & Acute that is so low by comparison.

Figure 1 Should we expect a better correlation between G&A and MH model results?

4.4. In the light of the above indications that the PBRA modelling gives only an approximate result, it is important to add judgment so that the overall allocation result makes sense. This reason for flexibility in the way models are used or adapted is in addition to the reasons given in the earlier points for also taking account of the need to reduce health inequalities and also to adjust for unmet or inappropriately met need.

B. PRIMARY CARE ALLOCATIONS

5. The distribution of GPs in England is inequitable at both Regional and PCT/CCG levels and in general favours better health areas and the South.

5.1 The geographical distribution of GPs in England has been problematic since the birth of the NHS.[9],[10] Many GPs prefer to live and work in desirable areas. Consequently it has been difficult to reduce the inequity in the distribution of GPs which favours areas of better health. This is despite the past efforts of the Medical Practices Committee, the publicised concern of many past Secretaries of State for Health, the transfer of the primary care financial allocations direct to PCTs and the Equitable Access[11] initiative.

5.2. Figure 1 illustrates recent inequity by Region. For example the North West has a shortfall of 13% or 620 full time equivalent (FTE) GPs, while London has an excess of 13% or 570 GPs. The figure also illustrates a strong north/south divide in GP provision – the five southernmost Regions have 22% more GPs per weighted population on average than the five northernmost Regions.

5.3. The calculation of relative GP provision uses the standard Department of Health method of counting Full Time Equivalent (FTE) GPs per needs weighted relevant population, with analysis carried out by Public Health Manchester. GPs were counted at September 2010 as sourced from the NHS Information centre for Health and Social Care, and the population is the 2010 relevant one from the Department of Health exposition book for 2010/11. The needs weighting used the formulae used for primary care allocations to former PCTs (2010/11 allocations) 12.[12] The benchmark is taken as the England average and the calculation of the Index of provision (England = 100) and the shortfall and excess of GPs are all relative to this benchmark.

Figure 2 % excess/shortfall GPs FTE by Region 2010 data

Table 3 GP provision by Region

5.4. Figure 4 illustrates the situation for former PCTs. Very substantial inequity exists with a greater range than at regional level. The range is from 27% under-provided to 53% overprovided compared with the England average and again calculated using weighted relevant populations. The variation for CCGs is likely to be greater because some PCTs have split into two or more CCGs.

Figure 3 Variation in GP provision for former PCTs (England average = 100)

5.6. Figure 5 shows the correlation between relative GP provision and premature mortality (as measured by the under 75 standardised mortality ratio for persons) for all 151 PCTs. The correlation is significant and the slope negative showing that in general premature mortality increases as GP provision decreases. This demonstrates that the inequity of GP distribution is in general strongly to the disadvantage of poor health areas.

Figure 4 Relationship between GP provision and premature mortality (u75 SMR) for former PCTs

6. Under-doctored CCGs are at a disadvantage but there are possible ways for NHS England to reduce the inequity

6.1. The creation of CCGs brings the problem of GP distribution into sharper focus. Under-doctored CCGs and their patients will be starting at a disadvantage and they may find it more difficult to absorb the new responsibilities for commissioning and budgeting; they will have little control over the situation because the GP budget will be under the control of NHS England.

6.2. These likely future problems make it important for NHS England to both review the problem of GP distribution and to develop a plan of action which would ultimately produce a fair distribution related to the needs of the population of each CCG. This action will be an important factor in addition to CCG allocations in order to set a level playing field for NHS England’s responsibility to hold CCGs to account for health outcomes.

6.3 Possible ways to reduce inequity are suggested as follows:

(i) The most obvious is for NHS England to commission and provide start-up funding for new practices in under-doctored areas. But fundamental problems remain which could hinder this such as a lack of sufficient financial incentives for GPs to work in deprived areas.

(ii) Gradually introduce a limit in weighted list size per FTE GP in each practice. There is currently no limit to the list size or weighted list size per GP so that there is a financial incentive for practices to maximize list size in under-doctored areas, sometimes rising to 3000 or even 4000 patients per GP compared with the England average of about 1700. This proposal should make it easier for such areas to absorb new practices.

(iii) Educate patients on the service standards they should expect, and on the weighted list size per GP in their practice. This is designed to encourage patients to move to a better practice if their current one has too many patients and consequently offers a poor service.

(iv) Give greater financial incentives to attract GPs to poor health areas.

The financial incentive in the Global Sum formula of the GP contract (it excludes the health inequalities factor used for PCT allocations), or salaries offered to salaried GPs, may not be high enough to attract GPs into deprived areas or to offset levels of private income in wealthier areas. Some commentators have also blamed the MPIG ’safety net’ payment for helping maintain inequity,[13] (though this effect should reduce with time), and some suggest that QOF targets are more difficult to achieve in poor areas.[14],[15]

(v) Investigate the possible relationship between where GPs train in practices and where they gain employment as GPs and if this is positive take steps to increase training places in under-doctored areas.