Chapter 8 - What are the key design choices for capitation?

Unlike block contracts, which are allocated to providers that perform a particular set of services, capitation is done on the basis of population groups. Where a block contract might be awarded to provide a portion of services for a whole population, a capitated model provides all the services to a subset of the population. Unlike block contracts, this allows providers to make trade-offs between services to personalise care according to the needs of an individual.

To allow the best possible care for individuals, it is generally best to focus on groups of people with similar needs. This allows outcomes to be more meaningful and providers to adapt the model of care accordingly. Local areas could choose based on:

• Groups defined by need, age or condition
• Geographical boundaries
• Administrative ownership like a GP list.

The co-design phase proposed ten groups around which Whole Systems Integrated Care can be organised. These groups were chosen, amongst other reasons, because they represent a way of segmenting groups of individuals who can be best served by a similar model of care.

This is an example of the sort of analysis which can be carried out to support the design of a capitation system. Here is shown part of the analysis for the integrated data set from Hammersmith and Fulham. Further details are explored in Supporting Material E: Review of Analyses During Co-design Phase.

Capitation can include some or all of the services (and their associated budgets) used to support the current population group. Commissioners need to decide which budgets they wish to pool into a capitated fund. The services which represent the majority of spend vary by segment, so the selection of services in scope substantially affects both the overall spend and the population groups affected.

The exhibit below is based on individual level data from our integrated data set, gathered across Hammersmith and Fulham. If all CCG, social-care and primary-care budgets were included, this gives an indication of the total size of the capitation for a population group of around 184,000 people and the current spending mix across settings. The full data pack, along with a discussion of assumptions and limitations, is attached in Supporting Material E: Review of Analyses During Co-design Phase.

The working group suggested the inclusion of all services delivered in residential, domiciliary and ambulatory settings for all primary care, home-based care, community mental-health, social-care services (including personal budgets) and third-sector services. They wanted to include acute ambulatory and nonelective care but exclude elective procedures.

Local areas could set a capitated budget for one, some or all of the people within the covered population. In doing so, they make a trade-off between the benefits of increased scale and the benefits of a smaller organisation.

Capitation on the basis of a smaller number of people:

• Improves accountability because it is easier to see who has had an impact on a group of people
• Improves the possibility for meaningful professional connections to be built between service providers, which improves care
• Increases the technical feasibility of building the organisation.

On the other hand, an organisational configuration based on capitation for a larger number of people:

• Increases the ability to operate at scale and gain the economies that improve efficiency
• Improves the ability to model risk and predict the budget accurately, as statistical fluctuations in individual costs average out across large populations (see Exhibit 8.5 overleaf).
• Improves the ability of the providers to spread risk because of their larger size

A statistical analysis using individual level data linking health- and social-care costs in Hammersmith and Fulham was conducted to estimate the likelihood of a provider going over budget. It showed that the risk of overspend decreases by increasing the size of the population.

Commissioners can use contingency insurance to mitigate some of the risks. Insurance works by protecting providers from overspends above a certain level, such as 5 percent of the total budget, or above a certain spend on an individual with exceptional care costs (e.g., more than £30,000). These types of insurance can be funded using a top slice from the original capitation budget. More detail is provided on this in question six below. Local areas have to decide the optimum trade-off between risk, accountability and insurance.

In order to calculate a full budget, it is essential to collect granular activity and cost data. However, it is possible to phase in a capitation programme without having already collected the necessary data.

In the data pack found in Supporting Material E: Review of Analyses During Co-design Phase, initial estimates of capitated budgets for each area based on the costs in Hammersmith and Fulham are provided. In order to calculate these figures, first a sample for each population group was taken and the total cost of care across all settings was found. This number was divided by the total to find the per person average.

Then, for each area, the average was adjusted based on the differences between Hammersmith and Fulham versus each other area. Differences considered include the relative proportion of people in each on population group, the relative prevalence of various conditions, total population sizes and relative spends n different care settings in each local area. More information about this adjustment process is available in Supporting Material E: Review of Analyses During Co-design Phase.

A shadow-capitated budget can be created based on this figure. In the first year, payment happens as normal, but performance can be compared against the shadow budget to allow fine-tuning of the capitation budgeting process before it goes live.

Initially it is not proposed that further adjustments are made by commissioners to the capitated budget for a population group. However, providers will need to prioritise the resources they receive for a population group according to the magnitudes of needs across the individuals for which they are providing care. Over time, as more data is gathered and the impact tested against shadow budgets, this initial assumption will need to be tested, particularly if there are concentrations of needs within some provider organisations but not others.

An example of the results of the capitation calculations based on the costs in Hammersmith and Fulham are below. More details can be found in Supporting Material E: Review of Analyses During Co-design Phase.

Commissioners can set the growth of budget relative to peer group performance, historic trends or an external objective like the growth rate for CCG funding. This can be set to allow predictable changes in budget over time. If performance is judged relative to peers, it is important to account for the differences in peer situations so that the mechanism allocates funding fairly for the populations. Budgeting relative to historic trends and requiring set savings is predictable and easier for providers to plan around. It can also be used to allow shared savings that incentivise cost control.

There also needs to be a mechanism, particularly early on in the transition process, to reset the budget if it was calculated incorrectly, either by changing capitated rates or by altering risk adjustments.

The working group suggested one ought to map the target budget growth to the growth in funding of commissioners. For the first three years, half the growth would go to providers, and half would be paid into the contingency fund used to cap deficits.

Commissioners can agree with providers to share any "upside” savings below the capitated budget. This provides an incentive to providers to improve care and manage costs so that they can reinvest any savings. "Downside” deficits or risks can also be shared between commissioners and providers. If providers overspend their capitated budget, commissioners may not cover the costs of the additional care. This provides an incentive to manage costs.

These agreements can be made independently, so commissioners and providers could agree to share savings, but protect providers from any deficits. Commissioners and providers can also use contingency insurance arrangements to cap the risks and savings. Insurance can be provided so that providers never have a deficit beyond an agreed amount. If providers have substantial under spends, commissioners and providers could agree that some of these savings are returned to commissioners’ support costs of contingency insurance or for other investments.

Commissioners can also choose whether "stop-loss” arrangements are done at the level of individuals or overall budgets or both. For example, commissioners might insure providers against costs above, say, £30,000 for each individual. Alternatively, they might only insure the provider against costs that are more than 5 percent above budget. A mixture of the two approaches could be used as well.

An example of how a mixture of deficit protection and shared surpluses, which was considered by the working group, is given below. Local areas will need to agree their arrangements depending on the capacity of providers to bear risk (e.g., how much cash they have set aside to cover deficits).

In addition, the fact that early adopters will be experimenting with a new system means that arrangements should be put in place to review contracts in the event that situations develop differently from the expectations of parties entering the agreement.

Commissioners and providers can choose where performance is judged, such as at a list level, at a practice, a provider network or across a borough. Smaller performance pools give clearer accountability for the outcomes, though risk random events hiding or pretending to be impact. Larger performance pools reduce accountability and risk successful providers being hidden, and as a result implicitly subsidising poor performance by others.

Commissioners and providers will both need strong performance management systems if capitation is to succeed. More discussion on these arrangements are in Chapter 7: How can we commission integrated care? and Chapter 10: How will provider networks develop and support new models of care?

Commissioners and providers can agree to either set the budget:

• Prospectively: Providers are given a budget at the start of the year, which is the benchmark to calculate any savings or deficits. This works best when providers are taking on risks of surpluses or overspends as the process is transparent.
• Retrospectively: Commissioners continue to pay providers using current mechanisms. At the end of the year they compare actual spending to the benchmark and either share any surplus or adjust future payments to cover any deficits. This benchmark could be calculated as relative performance between providers, rather than as a fixed target. The working group would recommend calculating the capitated budgets prospectively.

Commissioners need to determine which outcomes they want to achieve, the metrics they will use to measure the outcomes and how they will manage the performance of providers. More information is provided on this in Chapter 7: How can we commission integrated care?

Commissioners can choose mechanisms to link outcomes and finances such as:

• "Hurdle” levels that must be met before providers receive any savings
• Penalties for providers who fail to meet quality measures
• Linking proportions of shared savings or deficits to quality measures where providers providing high-quality care receive greater shares of any savings and are responsible for smaller shares of any deficits.

For the working group, there would be an agreed set of outcomes that would be tracked. Providers were required to reach a minimum standard. If they failed, these standards commissioners would use contractual levels to intervene, such as penalties or by working with the providers to reallocate resources to support an amended model of care. In addition, the working group believed that performance against outcomes could influence the proportions of loss protection and savings kept by the providers.

Local areas planning to adopt capitated budgets will first need to develop full business cases that evaluate the anticipated financial impact of the new payment model and the design choices. These should be built around the expectation of a long time horizon. Payment systems for integrated care are typically phased in over a moderate period, and make use of extended contracts of five or more years. For example, in Valencia, capitation contracts were agreed for 20-year periods.

During the transition, commissioners and providers will need to work together to:

• Develop the new payment model fully with complete design choices and collect the data required for an initial business case
• Discover, through shadow budgeting, the impact of trialling the new payment model on provider activity, deficits or surplus
• Adjust business systems to support the new payment model.

Commissioners and providers need to agree:

• Speed of adoption: How fast the system is set up depends on many factors, including will, technical ability to manage new arrangements and availability of infrastructure such as information systems
• Speed of risk sharing: Providers and commissioners need to agree how quickly providers are at risk for overspends. This is influenced by the capacity of providers to bear risks and change their behaviour to mitigate them.

International examples of moving towards new payment models phase in payments in a number of steps, managing the speed of adoption and risk sharing on a range of three to five years:

• ChenMed: Created a capitated system for elderly patients with complex chronic conditions with its individual clinicians. The capitated budgets for individuals are determined by Medicare based on a stepwise regression of diagnosed conditions to vary budgets according to anticipated needs and likely costs. ChenMed use these aggregated individual budgets to invest in an intensive primary-care service that includes central hubs with specialist staff on site. Care interventions provided to its population by other organisations, such as local hospitals, are reimbursed according to Humana (its payor) existing rates and deducted from its capitated budget. This avoids ChenMed needing to renegotiate contracts.

Humana provides stop-loss insurance so ChenMed is responsible for the first $40,000 of care in a year for individuals and is protected from costs over this amount.

ChenMed also developed individual clinician performance incentives and risk share. In the first phase, ChenMed physicians were paid a fee for service but with performance management to track their outcomes. In the second phase, physicians were moved onto a salary plus a share of upside risk so they could share in any savings from improved care quality and lower resource use. In the third phase, physicians began to share full risk but could receive greater share of savings.

• CMS Pioneer Accountable Care Organisation (ACO): This programme moved to a capitated payment model over a number of years. It began with three hospitals which had moved to a new payment model, where 50 percent of upside-only savings of at least 2.5 percent below the budget target could be kept by providers. This provided incentives for improvements whilst allowing the system to be fine-tuned. In the second year, it expanded to nine hospitals and the model changed to sharing 70 percent of upside savings and downside deficits above or below the budget. In the third year, the ACO migrated to a capitation model and all its hospitals received 50 percent revenues from risk-share contracts.

More information about the next steps for Whole Systems Integrated Care is available in Chapter 12: Next Steps.

Payment of these funds will be performance related and can only occur if an agreement between an LA and a CCG has been signed off by a Health and Well-being Board.

• How will you calculate your capitated budget in the first year and then for following years?
• How will you share savings and deficits relative to budget?
• How will you ensure quality?
• How will you transition to new payment systems?

To develop a capitation payment model you will need to plan for and complete the following:

• Meet as commissioners, providers and lay partners to discuss your vision for Whole Systems Integrated Care locally
• Decide what functions a payment model will need to enable this vision (more information on the payment models is available in the reference material)
• Work through the 10 considerations for capitation payment design based on the needs of your local area to develop a business case for capitation (further information on these issues is available in the reference material)
• Collect the local statistics to replicate the analyses, described in the supporting material, to build a shadow budget to test the new payment model with
• Get sign-off on early stage plan and invite early adopters to work on new models of care, potentially with additional investment available
• Monitor outcomes and use to inform the business case for continued investment
• Share learning with other localities to improve their planning in implementing capitation

See Chapter 4: What population groups do we want to include? for more details.

See Chapter 6: How do we innovate a new model of care working with users and carers? for more information

See Chapter 11: What informatics functionality will we need? for more details