POPULATION-BASED FUNDING FORMULA

FIVE YEARLY REVIEW SUMMARY

2007-08

MINISTRY OF HEALTH

Pre Publication Version

Contents

Acknowledgments

Preface

Executive summary

Overview of PBFF

Review – scope and process

Proposed changes

Cost per person calculations

Adjusters

Final Results by DHB

Appendix One: Data sets used in analysis

Appendix Two:The New Zealand index of deprivation and ethnicity

Acknowledgments

The Ministry would like to thank the following members of the PBFF advisory group.

Gary Jackson / Northern region, Counties Manukau DHB
Ratana Walker / Northern region, Waitemata DHB
Peter Lowry / Northern region(reserve), Auckland DHB
Julie Wilson / Midland region, Waikato DHB
Helene Carbonatto / Midland region, Tairawhiti DHB
Brian Gibb / Midland region(reserve), Taranaki DHB
Patricia Harvey / Central region, Wanganui DHB
Joce Carr / Central region , Hawkes Bay DHB
Gary Allen / Southern region, Canterbury DHB
Wayne Champion / Southern region, West Coast DHB
Mary Hart / Southern region(reserve),Nelson Marlborough DHB

Thanks also to the Wellington School of Medicine, Statistics New Zealand and the New Zealand Health Information Service who supplied much of the data used in this review.

Preface

This is a pre-publication version of the report on the population based funding formula review, 2007/08. This version is for a DHB readership.

The final published version may change as a result of editorial input and to make the document more accessible to a wider readership. The final published version will be made available to DHBs when completed.

Executive summary

This report presents the results from the first five-yearly review of the population based funding formula (PBFF).

The PBFF has been used to allocate funding to District Health Boards (DHBs) since 2003/04. PBFF is designed to fairly distribute the available funding among DHBs according to the relative needs of their populations and the cost of providing health and disability support services to meet those needs. Cabinet directed that the formula be reviewed every five years.

The review process involved extensive consultation with DHBs. The feedback received was overwhelmingly supportive of most aspects of the PBFF model, and consequently only minor changes to the formula were proposed to Cabinet. The results of the review were approved by Cabinet in November 2008 and the results of the subsequent technical update were approved by the Minister of Health and Minister of Finance in December 2008.

Cabinet approved the continuation of the current factors and adjusters used in the PBFF model. Cabinet also approved that the unmet need adjuster would increase by a fixed annual amount and that this funding would be targeted, initially at primary care.

Future work was approved on the following areas:

  • Investigation of the introduction of an adjuster for the capital charge costs on land
  • Investigation of possible ways tochangeMental Health funding to partially separate the implementation of new funding for mental health services from new funding for other services
  • re-estimation of the rural adjuster component for facilities.

As a result of the technical update to the PBFF model, the proportion of the total funding that each DHB receives will change. Implementation of the new model will mean that DHB shares will more closely align with current evidence regarding the health needs of the DHB’s population and costs of meeting this health need. Of particular note is the continuing evidence of higher cost weights for Maori and Pacific, compared with Other (ie non-Maori, non-Pacific) New Zealanders; and the higher cost weights for those living in areas of higher deprivation, as measured by the New Zealand Index of Deprivation 2006[1] (NZDep2006).

The revised and updated formula will be used to fund DHBs, commencing in 2009/10.

Overview of PBFF

The PBFF determines the share of Vote Health funding to be allocated to each District Health Board (DHB). Cabinet has previously agreed the aim of PBFF as being to fairly distribute the available funding among DHBs according to the relative needs of their populations and the cost of providing health and disability support services to meet those needs.

As a result of this review Cabinet has re-confirmed the existing structure of the PBFF and the population factors to be used.

Under the PBFF, each DHB’s share of funding is determined by:

  • costs per person multiplied by the DHB’s population (this component is also known as cost weights).

Plus the following three adjusters:

  • a policy-based weighting for unmet need, which increases the population weightings for Maori, Pacific people and those who live in areas of high deprivation
  • an additional adjustment for the unavoidable differences in costs that DHBs face in providing or funding services for rural communities
  • an additional adjustment for the unavoidable differences in costs that DHBs face in providing or funding services to eligible overseas visitors.

The groups used in the PBFF cost per person calculation are age, sex, NZDep2006, and ethnicity. There are three ethnic groups used, Maori, Pacific and Other (where other is defined as non-Maori, non-Pacific).

To calculate each DHB’s PBFF percentage share the following steps occur:

  • the costs per person for each population group are multiplied by the number of people in that DHB in that population group to get the expected cost for the group
  • theexpected cost for each population group in the DHB are added together with the adjuster funding to get the total expected cost for the DHB
  • the total expected cost for the DHB is compared to the total for all DHBs to calculate the DHB’s PBFF percentage share.

Review – scope and process

Scope

At the commencement of the five-yearly review process, the Ministry of Health consulted extensively with DHBs. The feedback received strongly supported most aspects of the current model. The report on the consultation phase is available from the Ministry on request.

Based on the consultation feedback the Minister of Health agreed a scope of review that was focused on making improvements to the current PBFF, rather than radical change.

Process

The PBFF review went through several steps between the initial start and the preparation of the final report. These steps were:

  • initial consultation with DHBs and central agencies
  • the DHB advisory group provided expert advice to the Ministry
  • Ministry provided advice to the Minister of Health
  • Minister of Health presented a policy paper to Cabinet who approved the policy
  • Ministry finalised the PBFF model
  • external audit of model completed
  • Ministry provided results of final technical work to Minister of Health and Minister of Finance.

There were relatively few changes made to the model after the Advisory group phase; the population projections were updated, the refugee funding in the overseas adjuster was calculated and there were some minor adjustments to calculations based on feedback from the auditor.

The report

The report covers the final PBFF model as approved through the above process. More detailed reports were made available to DHBs during the project and are available on request from the Ministry.

The remainder of the report covers in turn:

  • investigation of factors for inclusion in the model
  • investigation of additional adjusters
  • review of relationship between Mental Health funding and PBFF
  • calculation of costs per person
  • calculation of adjusters
  • results.

Proposed changes

Investigation of other factors

There was strong support from DHBs for the retention of the existing factors (age, sex, ethnicity and NZDEP2006) in the PBFF model. The subsequent modelling work confirmed the validity of the existing four factors. However the advisory group did investigate some changes and these are discussed below.

Ethnicity

The current PBFF uses ethnicity factors of Maori, Pacific and Other. The review investigated using Asian as a separate ethnic group. The cost analysis showed that people of Asian ethnicity have a lower average cost per person. Health outcome data suggests that the reason for the lower cost is because of the relatively better average health status of Asian people.

The review did not support introducing an ethnicity grouping reflecting lower cost weights for the Asian populationbecause of limitations in the data available. For example, there are some additional hidden costs associated with language and cultural barriers that are not picked up in the straight cost analysis. Further, the health status data available at present represents a ‘snap-shot’. We do not have quality time-series data to enable us to assess trends. Nor do we have quality data on the implications of NZDep in relation to Asian ethnicity.

There are also difficulties in that the Asian population group is not homogenous: it includes people from a variety of ethnicities and background, with different settlement histories. These factors all have an impact on health status and health need, and it is believed that the aggregate comparatively lower cost of the Asian population may obscure individual adverse profiles that may be different across different DHBs.

Deprivation

Changes to how deprivation was used in the model were examined to see if there is a better way to measure deprivation in comparison to the current approach. We investigated using deciles rather than quintiles, and using different combinations of deciles eg, deciles 1-6, 7, 8, 9 and 10. It was decided that continuing to use the quintile method was the most appropriate approach.

Clustering

Clustering refers to the situation when a DHB has a higher than expected level of a particular service. Clusters could be due to the characteristics of the population of the area, migration into the area to access the service or result from a higher level of supply stimulating demand.

The advisory group looked in detail at some example cases of services with DHB variation and also discussed the principles and practicalities involved. The group decided not to include clustering in the PBFF because of the difficulties in attributing clustering to a particular cause and also due to the practicality of either identifying possible clusters or testing for possible clusters.

Direct measures of health need

At present we use ethnicity and NZDep as proxies for health need. It has not been possible to use more direct measures of health need because there is insufficient data for this purpose. This could be considered as part of anyreview.

Aged Residential care:-income and asset testing regime and regional pricing

The amount paid for aged residential care (ARC) clients by the government depends upon an income and asset test. This leads to different DHBs having different distributions of client subsidy types based on the asset levels of their clients.

There is regional variation in price for this service and this also leads to variations between DHBs in total cost.

A model was developed to compare actual to expected costs per DHB allowing for both the income and asset testing and differential prices. This model ignored other reasons that could explain DHB differences: levels of family support, asset testing policy, and DHB choices over models of care.

The review did not recommend any adjustment to the PBFF for income and asset testing. This was for two reasons: firstly the financial impact was only borderline of material, and secondlythere are significant causative factors that we could not include in the model limiting its credibility.

Additional Adjuster Considered – the Land Adjuster

In response to feedback, a number of additional adjusters were proposed for inclusion in the PBFF. Criteria were developed to determine the appropriateness of new adjusters. Following application of these criteria, only one adjuster remained for further consideration: the Land Adjuster. The adjusters considered and the criteria used were made available to DHBs as part of the review.

A land adjuster was investigated because DHBs pay an annual capital charge on their land. Differences in value of land across New Zealand cause costs to vary significantly between DHBs. Analysis was undertaken to determine the effect of averaging the cost, and applying the average cost on an area basis.

There was some support for this approach from DHBs, but technically this work lies outside the scope of the PBFF review. Further work on the land adjuster will continue.

Other Policy Matters: Mental Health

The operation of a mental health ring fence alongside the population-based funding system causes operational problems. The main problem identified by DHB funders is that when DHBs take more than their PBFF share of new mental health funding in one year, they receive less demographic funding for other services in the following year.

As a follow up to the PBFF review an advisory group considered the feasibility of running a separate mental health PBFF to improve the alignment between PBFF funding and Mental Health ring-fence funding. The consensus of the advisory group was that this had considerable risk attached and was too complex a change to be implemented in the time frame. For 2009/10 the Ministry will provide DHBs with a comparison between their current Mental Health ring-fence funding and their PBFF mental health positions. This will be for information only. During 2009 there will be further consideration of a way forward to improve the alignment between the two methods.

cost per person calculations

Overview

The demographic groupings used for cost per person calculations are based on age, sex, ethnicity and NZDep2006quintiles.

Event level data was collected for the 2006/07 period for each expenditure type (for a complete list of data sets used see appendix one).

The costs per person were derived by dividing the total cost for each group defined by age, sex, NZDep06 by the national population for the group. These costs per person were then used to test if there is a residual differential increase or decrease in costs for ethnicity. If there is a residual difference for ethnicity the costs are adjusted. The adjustment is usually made as a percentage adjustment across all age, NZDep06 groups. However, sometimes a better fit to the data is achieved by making the adjustment specific to each NZDep06 quintile.

The adjustment for ethnicity is done in this way because the population groups are too small for many of the Maori and Pacific grouping to allow robust costs per person to be calculated for each age, sex, NZDep06, ethnicity group separately. For maternity expenditure the costs per person were calculated for age, sex, and ethnicity and then adjusted for the residual costs for NZDep06. Maternity was treated differently because for this service ethnicity was a stronger predictor than NZDep06.

The costs per personwere scaled to the 2006/07 expenditure using information supplied from DHB funder view accounts. This was done first by service area and then the service areas were summed to give the total costs per person. Missing data was assumed to follow the pattern for the service area. The adjusted scaled costsper personper service area are aggregated to form the final cost per person to be used in the PBFF formula. These costs per person are referred to as the cost weights.

The service areas used as sub-components in the PBFF model are:

  • personal health primary care
  • personal health other
  • mental health
  • psycho-geriatric
  • health of older people.

The results of the costs per person modelling are reported in the next section for each service area and then for the aggregate results.

The population statistics that have been used are supplied from Statistics New Zealand, and the NZDep06 has been created by the Wellington School of Medicine post 2006 census. A short description of the NZDep06 is included as appendix two.

Personal health primary care

Personal health primary care costs per person were modelled for pharmaceutical, laboratory and the immunisation data. The data sets showed lower cost weights for Maori and Pacific people when compared to Others.

Funding for primary health care organisations is based on a capitation funding model. As part of a technical update of the PBFF in 2006 the primary care capitation model was mapped to the PBFF factors and this was used as the costs per person for this service. Since there had been no change to the primary care capitation model since then these costs per person were used again.

The final primary care model combines the referred services costs and the capitation costs.

Figure one shows the variation by NZDep06 in the primary care data for females in the ‘Other’ ethnicity group. The variation is less than that found for Personal Health Other.