Technical Guide to the formulae for 2014-15 and 2015-16 revenue allocations to Clinical Commissioning Groups and Area Teams Insert heading depending on line length; please delete other cover options once you have chosen one. 14pt 0
NHS England INFORMATION READER BOX Directorate Medical Operations Patients and Information Nursing Policy Commissioning Development Finance Human Resources Publications Gateway Reference: 01356 Document Purpose Resources Document Name Author Publication Date Target Audience Technical Guide to the formulae for 2014-15 and 2015-16 revenue allocations to Clinical Commissioning Groups and Area Teams NHS England Strategic Finance 25 March 2014 CCG Clinical Leaders, CCG Chief Officers, CSO Managing Directors, NHS England Regional Directors, NHS England Area Directors, Directors of Finance Additional Circulation List Description Cross Reference Superseded Docs (if applicable) Action Required Timing / Deadlines (if applicable) Contact Details for further information #VALUE! This is a guide on how CCG and AT allocations have been calculated for 2014-15 and 2015-16 Equality Analysis for 2014-15 and 2015-16 revenue allocations to Clinical Commissioning Groups and Area Teams N/A N/A N/A Michael Chaplin Analytical Services (Finance) 5E40 Quarry House Leeds LS2 7UE 0113 825 3680 Document Status 0 This is a controlled document. Whilst this document may be printed, the electronic version posted on the intranet is the controlled copy. Any printed copies of this document are not controlled. As a controlled document, this document should not be saved onto local or network drives but should always be accessed from the intranet 1
Technical Guide to the formulae for 2014-15 and 2015-16 revenue allocations to Clinical Commissioning Groups and Area Teams First published: 25 March 2014 Prepared by: NHS England Analytical Services (Finance) 2
Contents Executive summary... 4 1. Introduction... 6 Overview of the weighted capitation formula... 7 2. Clinical Commissioning Group Allocations... 10 2.1 Populations... 10 2.2 Hospital and Community Health Services weighted populations... 11 2.2.1 General and acute... 12 2.2.2 Mental health... 13 2.2.3 Maternity... 15 2.2.4 Unavoidable cost... 15 2.3 Prescribing... 17 2.4 Unmet need adjustment: SMR<75... 18 2.5 Overall CCG weighted capitation formula... 19 2.6 Changes to CCG weighted populations since August 2013... 19 2.7 Pace of change... 20 3. Primary Care Allocations to Area Teams... 22 3.1 Primary medical care services component... 22 3.2 Dental services component... 23 3.3 Pharmaceutical services component... 23 3.4 Overall Primary Care formula and target allocations... 23 3.5 Pace of Change... 23 4. Better Care Fund 2015-16... 25 Annex 1: References... 27 Annex 2: List of documentation published alongside the technical guide... 29 3
Executive summary This guide provides an overview of the calculation of: 2014-15 and indicative 2015-16 recurrent programme allocations to Clinical Commissioning Groups (CCGs); 2014-15 and indicative 2015-16 primary care allocations to Area Teams (ATs); and funding for the 2015-16 Better Care Fund. Recurrent allocations to CCGs and primary care allocations to NHS England Area Teams were agreed by the NHS England Board at its meeting on 17 December 2013. There were four steps in the calculation of actual recurrent allocations to each CCG and primary care allocations to each Area Team once the national budgets had been set: determine target allocations based on relative need; establish baselines (the previous year s allocations); calculate opening distances from target (baseline minus target); determine pace of change policy, that is how far CCGs/ATs are moved closer to their target allocation within the year through differential growth. Pace of change policy balances providing stability in funding for all organisations with moving those furthest under target closer towards their target. The basic building blocks of the formula for CCG target allocations are the size of the population of each CCG and adjustments, or weights, per head for differential need for health care services between CCGs. The weights per head are based on: need due to age (the more elderly the population, the higher the need per head, all else being equal); additional need over and above that due to age (this includes measures of health status and a number of proxies for health status); an adjustment for unmet need and health inequalities; unavoidable higher costs of delivering health care due to location alone, known as the Market Forces Factor (this reflects that staff, land and building costs are higher in e.g. London than the rest of the country). The formula for primary care target allocations follows the same principles and approach whilst using data specific to primary care. Actual allocations are then determined from target allocations through pace of change policy, which for 2014-15 and 2015-16 set a minimum growth in allocations for all CCGs and higher growth in allocations per capita for those furthest under target. Pace of change for primary care allocations also set a minimum rate of growth and higher growth for those furthest under target. 4
The paper also provides a brief guide to the accompanying spreadsheets which show the calculation of target and actual allocations for each CCG and AT. We welcome comments on the formulae which may be sent to england.finance@nhs.net 5
1. Introduction The NHS England Board meeting on 17 December 2013 agreed 2014-15 and indicative 2015-16 programme allocations to Clinical Commissioning Groups (CCGs) and, for primary care, to NHS England Area Teams (ATs). The agreed allocations reflect the work of the fundamental review of allocations commissioned by the NHS England Board in December 2012. The decisions taken by the NHS England Board were based on the Mandate to NHS England requiring a transparent allocations process based on the objective of equal access for equal need, and the Mandate also requiring NHS England to have regard to reducing inequalities in access to and outcomes from healthcare. This guide provides an overview of the calculation of: 2014-15 and indicative 2015-16 recurrent programme allocations to Clinical Commissioning Groups; 2014-15 and indicative 2015-16 primary care revenue allocations to Area Teams; and funding for the 2015-16 Better Care Fund. There were four steps in the calculation of actual recurrent allocations to CCGs and primary care allocations to ATs once the national budgets had been set: determine target allocations based on relative need; establish baselines (the previous year s allocations); calculate opening distances from target (baseline minus target); determine pace of change policy, that is how far CCGs/ATs are moved closer to their target allocation within the year through differential growth. Pace of change policy balances providing stability in funding for all organisations with moving those furthest under target closer towards their target. Target allocations are based on the weighted capitation formulae recommended by the independent Advisory Committee on Resource Allocation (ACRA). A weighted population is calculated for each CCG and each CCG s monetary target allocation is the CCG s share of the total weighted population for England multiplied by the national budget for CCG allocations. The calculation of monetary target allocations for ATs follows the same approach. An overview of the weighted capitation formulae immediately follows. The subsequent sections provide more detail on the formula and pace of change policy. ACRA s recommendations for the formulae from 2014-15 are published alongside this technical guide. These build on their earlier recommendations set out in their letter of October 2012. The formulae recommended by ACRA are based on independent academic research and references are provided in Annex 1 to this research and other relevant publications. ACRA s remit does not include pace of change policy, which is set by NHS England. 6
This guide also provides a brief introduction to the accompanying spreadsheets which show the calculation of target and actual allocations for each CCG and AT. Due to the large size of the spreadsheets, many values have been hard copied rather than driven by Excel formulae. Where this is the case, the notes explain the relationship between the columns in the spreadsheets. Due to the large file sizes, the calculations have been set out over a number of separate files rather than two or three files. A list of the accompanying files is at Annex 2. Overview of the weighted capitation formula ACRA is an independent committee of NHS managers, GPs, academics, public health experts and local authority representatives. The formulae recommended by ACRA builds on the former formula for PCTs and on the former practice based commissioning toolkit. The weighted population for each CCG is based on: the size of each CCG s population; a weight, or adjustment, for need for health care services related to age (all else being equal, areas with older populations have a higher need per head); a weight, or adjustment for need over and above that due to age (all else being equal, areas with poorer health have a higher need per head); a weight, or adjustment, for unmet need and health inequalities; a weight, or adjustment, for unavoidable differences in costs due to location alone (higher unit staff, land and building costs) and the higher cost of providing emergency ambulance services in sparsely populated areas. The weighted capitation formula estimates the need of each CCG s population relative to other CCGs and is also known as the fair shares formula. It does not seek to measure an absolute level of need for each area, only relative need between areas. The populations used in the formula are the registered lists of all GP practice members of each CCG. People do not have identical needs for health care services. A key difference is that need varies according to age and gender, and in particular the very young and elderly, whose populations are not evenly distributed across the country, have a higher need for health services than the rest of the population. The weighted capitation formula therefore takes into account the relative need per head of different age-gender groups and the different age structures of local populations. Even when differences due to age are accounted for, populations with the same age profiles display different levels of need. An additional adjustment to reflect the relative need for health services over and above that due to age is therefore necessary. 7
Observing need per head directly has not proved possible to date. Instead statistical modelling by academic researchers has examined the relationship between the utilisation of health services on the one hand, and the characteristics of local populations and the area where they live on the other hand. These models have been used to decide which factors to include in the formula to predict future need per head. Typically the models estimate age related need and additional need over and above that due to age as a single set of weights rather than separate weights for age and additional need. This is because additional need varies by age group. The models also include supply variables to take account of the greater availability of health care services generally leading to higher use. While the supply variables are included in the models, they are set to the national average when calculating weighted populations. This means areas are not penalised in the formula for lower utilisation due to relatively lower capacity. As the need for different types of health services varies, there are separate formulae for general and acute, mental health, maternity and prescribing. These are combined to form the overall need weighted capitation formula for CCGs. The costs of providing health care unavoidably vary across the country due to different input costs, in particular staff costs and the costs of land and buildings. The weighted capitation formula includes an adjustment for these unavoidable costs, known as the market forces factor (MFF). These costs are due to location alone, not need. In addition, the emergency ambulance cost adjustment (EACA) adjusts for unavoidable differences in the costs of providing these service across the country, particularly in sparsely populated areas. The models typically assess need as it is currently met by NHS services and therefore may not capture unmet need or inappropriately met need. There is therefore an adjustment for unmet/inappropriately met need, based on a measure of population health (the standardised mortality ratio for those under 75 years of age (SMR<75)). This is applied to the population of each small area and then aggregated to CCG level. Applying the measure at the small area level takes into account unmet need/health inequalities within as well as between CCGs. ACRA s recommendations are largely based on independent academic research. However, due to the lack of robust quantitative evidence which is comprehensive and consistent between services and across the country, ACRA s recommended measure to be used for the unmet need adjustment was largely based on judgement. ACRA considered a range of measures of population health for the adjustment for unmet need. These were found to be highly correlated with each other. The SMR<75 has the advantage that it can be updated regularly at small area level, while other measures can only typically be updated at small area level using data from the 10 yearly Census. The SMR<75 was recommended as an indicator of the health of the whole population of areas, including morbidity and all age groups. The use of the SMR<75 was an interim recommendation and ACRA wishes to undertake further work in the area of unmet need. 8
ACRA was unable to recommend the share of the overall weighted capitation formula that should be based on the unmet need adjustment. The NHS England Board meeting of 17 December 2013 determined the share should be 10%. The weighted capitation formula for primary care allocations follows the same principles as that for CCG allocations. There has not previously been a primary care formula covering primary medical care, dentistry and pharmaceutical services, and ACRA viewed the primary care formula recommended as the best available presently but requiring further work for future allocation rounds. It is not possible in this overview to give full details of the models provided by the academic research. However, references to this research are provided in Annex 1. 9
2. Clinical Commissioning Group Allocations The following sections set out for CCG allocations: the population base; the weights for need, unavoidable costs and unmet need; how the weighted populations for G&A, mental health, maternity, and prescribing are combined into unified weighted populations for each CCG; and pace of change policy. 2.1 Populations The starting point for the weighted capitation formula is each CCG s population. The populations used are the registered lists of all member GP practices of the CCG. Registered lists are used irrespective of the patients place of residence or where they use NHS services. This follows the guidance Who pays? Determining responsibility for payments for providers (NHS England 2013). The weighted capitation formula is based on October 2013 registrations (by age-sex group) and CCGs member GP practices as published on 15 October 2013 by the Health and Social Care Information Centre. The weighted populations for 2013 are projected forward to give weighted populations for 2014-15 and 2015-16. This is undertaken by estimating the percentage growth in each CCG s registrations from 2013 to 2014 and 2015, and applying the same percentage growth in CCGs registrations to CCGs weighted populations. Registered lists for October 2013 are projected forward to give estimated registered lists for 2014 and 2015. They are projected forward using ONS s projections for resident populations for local authorities, which are mapped to CCGs, to give projected resident populations for CCGs. The percentage growth in CCGs registrations is taken to be same as its estimated percentage growth in its resident population. In essence this means: Where a CCG s geographical area exactly matches a local authority s, the percentage increases in registrations from 2013 to 2014 and 2015 are the same as the percentage growth projected by ONS for the local authority s resident population. Where a CCG s geographical area covers more than one local authority, the average of the percentage growth for each local authority s population is used. The weight of each local authority in the average is its share of the CCG s resident population. The ONS projected populations for local authorities used are the interim 2011 based sub-national populations for unitary and lower tier local authorities. These projections are based on the 2011 Census. The 2012 ONS mid-year population estimates for the LSOAs are used to map local authority resident populations to CCGs. No LSOA crosses a CCG or local authority boundary. LSOAs are an ONS designed small geography, with 32,844 in England. Weighted populations for 2014-15 and 2015-16 are those for 2013 uplifted by the estimated percentage growth in registrations from 2013 to 2014 and 2015 as 10
described above. The growth in local authorities total populations is used, not separately by age-sex groups. 1 It is recognised that using registered lists does not take into account people who are not registered with a GP practice. ACRA considered whether an adjustment should be made to the formula for unregistered populations, but due to the absence of reliable data available on the size of the unregistered population by area and their healthcare needs, concluded it is not presently possible to do so. A - Registrations by GP practice and CCG October 2013 - Excel file This gives the number of registrations by GP practice and CCG, broken down by age-sex group in October 2013 used for the CCG weighted capitation formula. B Calculation of CCG population growth rates - Excel file This shows the mapping of local authority to CCG populations and calculation of the percentage growth rates. 2.2 Hospital and Community Health Services weighted populations There are three steps in calculating HCHS weighted populations. The first is to weight or adjust registered populations for need, the second is to weight or adjust for unavoidable costs due to location, and the third is to adjust for unmet need/health inequalities. The basic approach in calculating need weighted populations is to multiply the population for each age-sex group for each GP practice by the relative need per head estimated by academic researchers. The products for each age-sex group are summed to give the relative need weighted population for each GP practice. The weighted populations for GP practices are summed to give the relative need weighted populations for each CCG. The approach for weighting for unmet need is based on the standardised mortality ratio for those under 75 years of age (SMR<75) applied at small area level to take account of inequality in health outcomes within as well as between CCGs. The two adjustments for unavoidable costs due to location are the market forces factor and the emergency ambulance cost adjustment (EACA) are then added at CCG level. The SMR<75 weighted population combined with the need and unavoidable cost weighted population gives the relative overall weighted population for each CCG. Within HCHS, need per head differs for general and acute, mental health and maternity, and the HCHS weighted populations are therefore built up from these separate components. Each is discussed in turn below. 1 2013 weighted populations uplifted by the percentage growth in registrations gives a different total for England than unweighted registrations. Each CCG s weighted population for 2014 and 2015 is scaled by the same percentage so that the England total for weighted and unweighted registrations are equal. This leaves each CCG s share of England weighted registrations unchanged. 11
2.2.1 General and acute Need per head for general and acute was estimated by the Nuffield Trust using a person based approach, building on the research for the former practice based commissioning toolkit. The person based approach uses data at the individual level (anonymised) to provide more accurate estimates of need for small populations, such as GP practice lists. The Nuffield Trust research estimated jointly need related to age and additional need over and above that due to age. The Nuffield Trust s research covered inpatient and outpatients for general and acute and also covered accident and emergency. Mental health, maternity and specialised services were excluded. CCGs are not responsible for commissioning specialised services, which are directly commissioned by NHS England. The Nuffield Trust estimated relative need based on past patterns of utilisation of health services compared with the characteristics of patients. The quantified relationships found are taken to be predictors of relative future need for health care services, with the exception of the supply variables. The Nuffield model used anonymised data on the diagnoses for each patient admitted to hospital in 2007-08 and 2008-09, their age, and their GP practice. The numbers of anonymised registrations by age-sex group were also obtained for each GP practice to provide information on the proportions of a GP practice s list using health care in 2007-08 and 2008-09. Other data included in the model were data from the population census and attributed to individuals based on their place of residence - these are data only available for small geographical areas (LSOAs) rather than for each individual. They include data such as the proportion of people from black and minority ethnic groups, and the proportion of people aged 16-74 in semi-routine occupations 2. The model estimated the relationship for individuals between these explanatory variables for 2007-08 and 2008-09 and actual costs for the individuals in 2009-10. The modelling tested from a wide range of potential variables which were the best in statistical terms, and were also plausible indicators of need, to be included in the final model. It was found that diagnoses and age were the most important variables in the model. The utilisation of health care may also be affected by the relative availability of health care services. Variables were included in the model to adjust for this, known as supply variables. These variables included for example distance travelled to outpatient appointments and the number of operating theatres. While these variables were included in the models as they affected utilisation, they were not included in the formula to calculate weighted populations, instead their value for each area was set 2 At the time the research was taken the results from the 2011 Census were not available. One of the variables included by Nuffield was the ratio of ONS populations to GP registrations. This variable was updated by NHS England to take into account the results of the 2011 Census. 12
to the national average. This means if an area has lower use of health care services because of lower capacity or longer distance, this is corrected for in the formula. It is assumed that the predictors of need in 2009-10 are also accurate predictors of need in subsequent years. The Nuffield model includes nearly 400 variables which were applied to each individual to provide need per head for each age-sex group for each GP practice. The Nuffield Trust provided the need per head values for each age-sex group for each GP practice directly to NHS England. NHS England did not perform (nor holds) the calculation due to the large size of the data set and to protect the anonymity of individuals data. A small number of GP practices have been newly formed since the Nuffield research was undertaken. NHS England used the average CCG need per head by age-sex group for these GP practices. C Need per head (General and Acute) - Excel file This shows: - the need per head for each age-sex group for each GP practice provided by the Nuffield Trust; - where the need per head was not available from the Nuffield research for new practices, and the average need per head by age-sex group for the CCG was used; - each GP practice and CCG s registrations weighted for need (general and acute); - the variables included in the Nuffield model and their coefficients. 2.2.2 Mental health The need per head for mental health services was estimated by a team led by Manchester University and followed a similar person level, utilisation based approach to the Nuffield Trust, and is known as the person-based resource allocation for mental health (PRAMH) formula. Similarly, it jointly estimates need related to age and additional need over and above that due to age. The PRAMH model is based on analysis of the Mental Health Minimum Dataset (MHMDS) over the period 2008-09 to 2010-11. The MHMDS covers specialist mental health services within hospitals, outpatient clinics and the community. Very specialised mental health services which are not be commissioned by CCGs but by NHS England were excluded from the research. As a relatively small percentage of the population use mental health services in a year, the researchers recommended a two-stage model. The first stage models the proportion of individuals who use mental health services, and the second stage models the need based costs for the service-using population. Additional data on patient characteristics were available for the second stage. The modelling for both stages was undertaken for each age-sex group for each GP practice. 13
There are separate models for males aged 20-64, females aged 20-64 and those aged 65 and over. This is because relative need differs between these groups, the latter being heavily influenced by dementia and related illnesses. The explanatory variables in the models include for example age, psychiatric diagnosis, severe mental illness prevalence from the quality and outcomes framework (QoF), categories of condition of mental health severity, the proportion who are single, and ethnicity. As for general and acute, supply variables were included in the model but set to the national average in the calculation of weighted populations. The supply variables included for example the existence of a nearby mental health provider and distance to mental health team base. The available data for the research did not cover those aged under 20 and so an alternative method was used by NHS England for calculating mental health need per head for the four age bands under 20. The method used bed days and outpatient data with mental health diagnostic codes from HES to estimate the national hospital cost per head by age-group. The costs for those aged under 20 were expressed as a percentage of those aged 20-24. These national percentages were applied to the need per head from the PRAMH project for those aged 20-24 for each GP practice as estimates of the need per head for the age-groups aged under 20. As the use of mental health services by those aged under 20 is relatively low, it is unlikely this approach significantly affects the overall mental health weighted populations. The research team provided need per head values for each age-sex group for each GP practice directly to NHS England. NHS England did not perform (nor holds) the underlying calculation of applying the need variables to each individual due to the large size of the data set and to protect the anonymity of individuals data. A small number of GP practices have been newly formed since the PRAMH research was undertaken. NHS England used the average CCG need per head by age-sex group for these GP practices. D Need per head (Mental Health) - Excel file This shows: - the need per head for each age-sex group for each GP practice provided by the PRAMH research team, and the estimates for those aged under 20 as described above; - where the need per head was not available from the PRAMH research for new practices, and the average need per head by age-sex group in the CCG was used; - each GP practice and CCG s registrations weighted for need (mental health); - the variables included in the PRAMH model and their coefficients. 14
2.2.3 Maternity The maternity model is based on the number of births and the need weighted cost per birth. The model is from the Combining Age Related and Additional Need (CARAN) report. The need weighted cost per birth is based on a model which found the best explanatory variables to be the proportion of births which are low rate births and mean house price in the local area. The CARAN report is based on data for 2004-06. The number of births used are new registrations due to births for the 12 months to April 2013. E Need (Maternity) Excel file This shows: - the number of new registrations for births, the value of the variables in the maternity model and their coefficients; - each GP practice and CCG s registrations weighted for maternity need; - the variables in the maternity model and their coefficients. 2.2.4 Unavoidable cost There are two adjustments for unavoidable costs, the market forces factor (MFF) and the emergency ambulance cost adjustment (EACA). Market Forces Factor The MFF adjusts for the unavoidable cost differences between areas due to their geographical location alone. For example it typically costs more to run a hospital in a city centre than in other areas due to higher staff, buildings and land costs. This adjustment is for higher, unavoidable input costs alone, not due to higher costs due to higher need. The MFF currently used is based on that calculated for former PCT allocations and the methodology is set out in Resource Allocation: Weighted Capitation Formula, Seventh Edition. There are four components to the MFF, unavoidable differences in cost across the country due to each of: medical and dental staff; other staff; land; and buildings. The staff component (non-medical and dental) was based on the HERU research report The Staff Market Forces Factor component of the weighted capitation formula: new estimates. In the NHS, pay rates are determined by national pay structures and therefore differences across the country are relatively small. However, indirect pay costs faced by providers differ significantly across the country, such as vacancy rates, staff turnover rates and use of agency staff. The HERU research report used differences in pay rates across the country in the private sector, which were found to be highly correlated with these indirect staff costs faced by NHS providers. The 15
private sector pay rates used were adjusted for differences across the country in age and sex of employees, occupation, industry and level of responsibility of the job. Indirect staff costs for medical and dental staff were found not to differ across the country as they do for other staff. Instead the medical and dental component was based on the direct, higher costs of employing medical and dental staff in London, i.e. on the London pay weighting. The building component was based on relative location factors calculated by the Building Cost Information Service (BCIS) from an analysis of tender prices for public and private contracts at local authority level. The land component was based on the land value per hectare calculated for each Trust. This used the net book value of land in providers audited summarisation schedules and land areas reported in ERIC returns. The data for each of the four components were for the period 2007-10, the exact years varying by component. The MFFs last calculated for each provider using the approach outlined above is the starting point for the calculation of MFFs for CCGs. The MFF for each CCG is calculated from the MFFs of providers where each member GP practice s patients received inpatient and outpatient treatment. The CCG MFF is the weighted average of providers MFFs, where the weights are the activity (as recorded in HES and costed using payment by results tariffs) undertaken by each provider for patients registered with each CCG s member GP practices. The weights are often known as the purchaser-provider matrix. The weights use the 2013-14 tariff (excluding best practice tariffs) and 2011-12 HES activity. The CCG MFFs are expressed as an index, with the England average set to the value 1.0. The same MFF index value is applied to each of general and acute, mental health and maternity. Emergency Ambulance Cost Adjustment The Emergency Ambulance Cost Adjustment (EACA) adjusts for unavoidable variations in the cost of providing emergency ambulance services in different geographical areas, and in particular sparsely populated and metropolitan areas. The EACA currently used is that for PCTs mapped to CCGs based on geographical boundaries The methodology for the PCT EACA is set out in Resource Allocation: Weighted Capitation Formula, Seventh Edition. The CCG EACAs are expressed as an index, with the England average set to the value 1.0. The same EACA index value is applied to each of general and acute, mental health and maternity. 16
F Market Forces Factor - Excel file This shows: - the % of each CCG s costed inpatient and outpatient activity with each provider; - the 2013-14 provider payment target MFFs; - the scaling to rebase CCGs MFFs so that the England average equals 1.00. 2.3 Prescribing The prescribing component covers medicines prescribed in primary care. There is first an adjustment for need related to age and sex and then an adjustment for additional need over and above that due to age. The adjustment for age and sex uses the weights developed by the Prescription Support Unit (PSU) known as ASTRO(09)-PUs. The second adjustment made, for prescribing need over and above that due to age and sex, is from the Report of the Resource Allocation for Mental Health and Prescribing (RAMP) project. This is the same as used for the former PCT formula, and which was originally calculated for each GP practice. The model for additional need includes both need and supply variables as for the other components. The variables are set out in the RAMP research report. The need variables include for example the Low Income Scheme Index (LISI), the proportion of those aged 70 years and over claiming disability living allowance (DLA), and the standardised mortality ratio. The RAMP report used data for circa 2008-09. The RAMP model and ASTR0(09)-PUs were applied for each GP practice and the GP practice weighted populations summed to give the CCG weighted populations. The prescribing component is not adjusted by the MFF or EACA as the costs of prescribed medicines are the same throughout the country. G - Need (Prescribing) - Excel file This shows: - the calculation of registrations weighted for age, sex and additional need for each GP practice and CCG; - where the variables in the model were not available from the RAMP research for new practices, and the average per head by age-sex group in the CCG was used; - the coefficients and variables in the RAMP model. 17
2.4 Unmet need adjustment: SMR<75 In the absence of robust quantitative evidence which is comprehensive and consistent between services and across the country, ACRA s recommendation of the measure to be used for the unmet need adjustment was largely based on judgement. ACRA was unable to recommend the share of the overall weighted capitation formula that should be based on the unmet need adjustment. The NHS England Board meeting of 17 December 2013 determined the share should be 10%. ACRA considered a range of measures of population health for the adjustment for unmet need. These were found to be highly correlated with each other. The SMR<75 has the advantage that it can be updated regularly at small area level, while other measures can only typically be updated at small area level using data from the 10 yearly Census. The SMR<75 was recommended as an indicator of the health of the whole population of areas, including morbidity and for all age groups. The SMR<75 is a measure of how many more or fewer deaths there are in a local area compared with the national average, having adjusted for the difference between the age profile of local areas compared with the national average. It is applied at small area level (middle layer super output area (MSOA)) and then aggregated to CCGs. This allows for inequality within, as well as between, CCGs to be taken into account. Each MSOA was assigned to one of 10 groups based on its SMR<75 value, those with the lowest SMR<75 values were in group one, and those with the highest SMR<75 values were assigned to group ten. The groups had an equal span of SMR<75 (subject to at least 5% of MSOAs being in the group). The alternative of having equal numbers of MSOAs in each group would have meant very small differences in the SMR<75 values between the middle groups. Each of the ten groups is assigned a weight per head, with the MSOAs in group 10 having a weight five times higher than the MSOAs in group 1. The weight for the intermediate groups increases exponentially, so that group one has a weight of 1.00, group two a weight of 1.20, group three a weight of 1.43, up to group ten with a weight of 5.00. The exponential increase in the weights means the impact of the SMR<75 based adjustment between CCGs depends on how many of its MSOAs are in each of the 10 groups. Each MSOA s population is given a weight of between 1 to 5, and the MSOA weighted populations are then summed to CCG level. This approach follows that used in the formula for public health grants to local authorities in 2013-14 and 2014-15. H SMR75 weighted populations - Excel file This shows: - the weights per head for each of the 10 groups; - the calculation of MSOA and CCG SMR<75 based weighted populations; - for each CCG, the number of MSOAs in each of the 10 groups. 18
2.5 Overall CCG weighted capitation formula The overall weighted populations for CCGs combine together i) the HCHS need and unavoidable cost weighted populations, ii) the prescribing need weighted population, and iii) the unmet need adjustment. The HCHS components are combined based on information from month 6 forecast 2013-14 spend. The same source is used to combine the HCHS and prescribing components. No formula is available for community health services and it is assumed that the general and acute and mental health formulae are also representative of community health services. The NHS England Board decided that the unmet need adjustment should have a weight of 10% and the other components together a weight of 90%. The weighted populations for CCGs have been calculated based on October 2013 registrations. Weighted populations for 2014-15 and 2015-16 are those for 2013 uplifted by the estimated percentage growth in registrations from 2013 to 2014 and 2015 as described above. The overall weighted capitation formula is used to calculate CCGs target shares of the available resources and are a key component in pace of change policy. 2.6 Changes to CCG weighted populations since August 2013 As a contribution to the fundamental review of allocations announced in December 2012, NHS England published in August 2013 the interim CCG formula and weighted populations. The interim formula was considered by the Board of NHS England in December 2012 but was not adopted, instead all CCGs were given the same percentage growth in their 2013-14 allocations. The changes from the interim formula published in August 2013 are: a) updating registrations from April 2012 to October 2013; b) updating CCGs member GP Practices from March 2013 to October 2013 as recorded in the Organisation Data Service; c) a small adjustment in the application of the mental health formula; d) an updated purchaser-provider matrix for the MFF, including 2013-14 tariffs; e) the introduction of the unmet need adjustment based on the SMR<75; f) weighted populations for October 2013 projected forward to 2014-15 and 2015-16 based on ONS projected populations for local authorities. The Excel file I CCG weighted populations gives a breakdown of the change in distance from target from those published in August 2013, which include a) to e) above, and also the update in baselines from 2013-14 allocations announced in December 2012 to 2013-14 allocations based on Month 6 plans. 19
I - CCG weighted populations - Excel file This shows: - the overall weighted population for each CCG based on October 2013 registrations; - how the weighted populations for HCHS, mental health, maternity, prescribing and the SMR<75; - the overall weighted population for each GP practice based on October 2013 registrations (excluding the SMR<75 based adjustment as this not available by GP practice); - CCG weighted populations projected forward from October 2013 to 2014 and 2015; - a breakdown of the change in the distances from target between those published in August 2013 and those based on the updates a) to e) above and updated baselines. 2.7 Pace of change Traditionally, local commissioning organisations such as the former Primary Care Trust (PCTs) were not immediately given the target allocations as determined by the weighted capitation formula, but moved towards their target allocations over time under pace of change policy. Pace of change balances providing stability in funding for all organisations with moving those furthest under target closer towards their target. Pace of change policy for CCGs is set by NHS England. Pace of change policy involves: establishing the baselines (2013-14 allocations based on Month 6); establishing the available national budget (growth of 2.54% in 2014-15 and growth of 2.09% in 2015-16); determining target allocations from CCG weighted populations; calculating distance from target (the difference between the baseline and the target allocation. If the target allocation is greater than the baseline the CCG is said to be under target. If the baseline is greater than the target allocation then the CCG is said to be over target); setting actual allocations through differential growth in allocations between CCGs, using criteria based on distance from target. At its December 2013 meeting, the NHS England Board considered seven options for pace of change. Option four was adopted. The options considered were as follows. Option 1 was uniform growth, under which all CCGs would receive the same percentage growth in their total allocations, and there is no pace of change. Options 2 to 4 were all based on per capita allocations and took as their first step an adjustment of resources for growth in expected population in 2014-15 and 2015-16. 20
Option 2 gave all CCGs the same growth of 1.66% in their per capita allocation in 2014-15 and 1.23% growth in their per capita allocation in 2015-16. Option 3 sought to maximise pace of change for the most under target CCGs. Growth per capita would be 3.3% for those 5% or further below per capita target, reducing to 1.22% for those between 3% and 5% under target. Growth in total allocations would be limited for significantly (>5%) over target CCGs to 2.14% (= real terms protection per GDP deflator). For 2015-16 the approach is similar, with maximum growth in total allocations of 1.7% for those more than 5% over target (above the GDP deflator of 1.5%). Option 4 introduced a floor so that all CCGs would see their total allocation grow by at least 2.14% in 2014-15 and 1.7% in 2015-16. This reduced the resources available for the most under target CCGs. The maximum per capita growth falls to 2.64% in 2014-15. This sought to balance the challenge of directing additional funding to those CCGs under target on a per capita basis with providing stability in funding for all CCGs. Option 5 to 7 modelled a more rapid pace of change. These were: 0-5%: where CCGs most above target receive flat cash and those furthest below target receive 5% funding growth, with the position of the transition between the two determined by affordability until target is reached (option 5); Cap and collar: where CCGs more than 6.33% above target receive no more than 0.89% (matching average population growth), CCGs within 5% of target receive 2.14% and CCGs more than 5% below target receive an increase in excess of real terms growth rising to 10% for the most underfunded CCGs (option 6); 5 years: where growth is set at a pace that, all else being equal, would bring each CCG to within 5% of target within 5 years (option 7). J - CCG pace of change options - Excel file: This shows the 7 options described above. 21
3. Primary Care Allocations to Area Teams NHS England asked ACRA to advise on a formula for primary care to be used to allocate primary care budgets for the 25 Area Teams. The recommended formula has separate components for primary medical care, dentistry and pharmaceutical services. Each is discussed in turn below. There is not a separate component for ophthalmic services due to the lack of available data, and expenditure on this service is relatively low. As already noted above, there has not previously been a primary care formula covering primary medical care, dentistry and pharmaceutical services, and ACRA viewed the primary care formula recommended as the best available presently but requiring further work for future allocation rounds. ACRA also recognised that the formula represented primary care as currently delivered and that they could not predict future changes in primary care structures. The primary care formula covers the funding in the line Primary Care (to be allocated) in Table 2 of the NHS England Board paper on allocations for the 17th December 2013 meeting. It does not cover the line in the table Primary Care (other budgets). 3.1 Primary medical care services component The primary medical services component is based on the Carr-Hill formula, which is the contractual basis for the distribution of global sum payments in the GMS contract to GP practices. The Carr-Hill formula is based on academic modelling of consultant workload and is largely based on need reflecting the number of people registered and age-sex mix of the practice, as well as factors related to the health status of the population. The formula uses GP practice registered populations which are adjusted for the following factors: Age-Sex Age-sex workload curve (for 14 male-female age bands) used to reflect frequency of home and surgery consultations Nursing and Residential Homes practice count of patients in residential care Additional Needs practice average standard mortality index practice average Limiting Long-Term Illness (LLTI) Index Unavoidable Costs Staff Market Forces Factor for the practice electoral ward. The MFF is applied only to practice staff costs, not to other costs. Weighted populations were calculated for each GP practice and were summed to Area Team level. 22
ACRA recommended excluding the new registrations and rurality variables in the Car-Hill formula as ACRA felt these factors did not represent need. More information on the Carr-Hill formula can be found in Appendix A of the Review of the General Medical Services global sum formula report. 3.2 Dental services component The dental services formula uses national average costs by age, sex and IMD of patients residence for those accessing NHS dental care. This is multiplied together with the number of NHS patients seen (mapped to Area Team based on which Area Team now holds the dental contracts) in the 2 years to March 2013 by age-sex-imd group. Revenues from patient charges were deducted using the 2012-13 ratio of patient charges to the total value of dental contracts, mapped to the Area Team based on which Area Team holds the dental contract. 3.3 Pharmaceutical services component The prescribing formula, which is part of the CCG allocations formula, and based on academic research is used for pharmaceutical services. This allocates resources for the cost of providing community pharmacy services in line with the cost of the drugs dispensed. The prescribing weighted populations from the CCG formula are scaled to the primary medical care population size and aggregated to Area Teams. 3.4 Overall Primary Care formula and target allocations The individual weighted populations for primary medical care, dentistry and pharmaceutical services were combined together in line with 2013-14 forecast spend on these services. The NHS England Board decided that an unmet need adjustment should be applied using SMR<75, and accounting for 15% of the overall primary care weighted populations. The SMR<75 weighted populations for ATs is the sum of those for the CCGs in their area. For 2014-15 and 2015-16 target allocations, each Area Team s overall weighted population was uplifted by the projected change in registered lists of the CCGs in their area. 3.5 Pace of Change Pace of change (PoC) policy determines final allocations. Area Teams furthest under target received the highest growth to move them closer towards their target allocations. Primary Care PoC was based on: overall budget uplift of 2.14% in 2014-15 and 1.70% in 2015-16; those more than 2% over target, growth of 1.6% in 2014-15 and 1.2% in 2015-16; those between 0% and 2% over target, proportional growth between the minimum and average; those under target, growth of at least 2.14% in 2014-15 and 1.70% in 2015-16, with proportionally higher growth for those Area Teams furthest under target. 23
K Primary Care - Excel file This shows: - the calculation of weighted populations for primary medical services, pharmaceutical services and dentistry; - how the weighted populations for primary medical services, pharmaceutical services, dentistry and SMR<75 are combined to give an overall weighted population for ATs; - pace of change. 24