Background
Equity has been a guiding principle in the British NHS, which is
unique among health systems in distributing the majority of its
revenue resources between geographical areas in relation to population
needs. However, financial equity does not guarantee equity in access
to health care and marked variations have been recorded in both
GP referral rates and hospital admissions. Due to the difficulties
of defining and measuring 'need', the reasons for and implications
of such variations for health care equity are poorly understood.
This lack of information has important implications for health care
equity. Where relatively high rates of use are underpinned by legitimate
health demands, policies directed at securing spatial equity in
health care expenditure can result in clinical inequity.
This dilemma
has been drawn into sharper focus with the introduction of primary
care groups (PCGs). Not only has the scale of resource allocation
been reduced to the sub-Health Authority level, but monitoring of
resource utilization has intensified within PCGs (i.e. at the general
practice level). Unless better information becomes available about
how and why general practice populations vary with regard to health
service utilization, such efforts may undermine the government's
commitment to promote equity in health care.
Aims and
Objectives
The aims of this study are to:
- examine variations
in health service utilization according to 'need' by GP populations;
- explore the
extent to which inequities that are found arise from systematic
supply effects, are associated with the population characteristics
of general practices, or reflect variations in medical practice
at the local level;
- explore the
equity implications of replacing resource allocation based on
historical activity with allocations based on weighted capitation.
Study Design
The study area comprises twelve health authorities in contrasting
locations in the United Kingdom. The general practice is the unit
of analysis. In the first phase of the project, a series of needs
estimates will be generated, including census-derived rates of self-reported
illness and socio-economic status, expected rates of major cancers
and indicative prevalences of specific conditions. These will be
linked with health authority on-line prescription and contract minimum
data sets in order to calculate use-need ratios for total referrals,
admissions and prescriptions and for specific treatments or procedures
for specific conditions. Using the use-need ratios as the dependent
variable, the second phase of the project will determine the extent
to which inequities in utilization arise from systematic supply
or demand characteristics or are independently distributed. Targeting
residual practices whose patterns of health resource utilization
diverge from the expected in ways that cannot be explained and which
would stand to gain or lose significantly under a localized system
of financial equity, Phase III of the project involves a series
of semi-structured interviews with GPs.
The study design has a number of key methodological features, including
its use of the general practice as a functionally meaningful unit
of analysis, the examination of the impact of variations in the
relative mix of primary and secondary care, the emphasis on cause-specific
morbidity and the incorporation of cost measures.
Policy Implications
An important objective of health policy is to ensure that disadvantages
conferred to health status by socio-economic position are not compounded
by inequities in health care. However, the fact that equity objectives
relating to different levels of the health care system can sometimes
conflict is rarely acknowledged. By providing information on how
and why variations in service use occur, this research has been
designed to help policy-makers identify ways of promoting a fair
allocation of resources without exacerbating inequalities in access
to health care.
Project Summary
The government has made clear its commitment to introducing changes
into the way the NHS distributes its resources. Changing the resource
allocation process in ways which both move resources down to local
level (to primary care groups) and which target health inequalities
is a key component of the government's public health strategy.
The project
addresses questions central to the development of new resource allocation
mechanisms. It examines the factors underlying variations in the
use of primary and secondary health services and treatments between
general practice populations, and explores the equity implications
of introducing resource allocation mechanisms based on weighted
capitation at more local scales. Using a range of sources, it has
generated a dataset which links estimates of health care need for
specific conditions to data on health service utilisation at the
level of the general practice and primary care organisation.
Key findings
- A method
of using existing epidemiological data to apply morbidity prevalence
and incidence rates to targeted populations has been developed,
based on age, sex and socio-economic status. This 'indicative
prevalence methodology' provides estimates of need which are not
based on utilisation, the major limitation of current weighted
capitation formula.
- The method
has been used to measure the degree to which different standard
proxies of health service need offer good or poor approximation
of disease prevalence. For example, analyses indicate that age/sex
estimates are more reliable than standard deprivation indices
for conditions such as angina; in contrast, standard deprivation
indices are more acceptable proxies for mental health services.
- Because the
use of epidemiological estimates allows one to distinguish between
that part of practice-level variation that is due to variations
in need and that part which is due to variations in use relative
to need, this method lends itself well to the monitoring of equity
and fair access. However, some of our findings (based on analysis
of the activity of over 550 practices serving a population of
3.5 million patients) are at odds with the received wisdom that
the use of NHS services is characterised by an inverse care law.
For example, examining use in relation to need for cardiology
services, urban practices that are close to acute hospitals and
whose practice populations are highly deprived use inpatient services
to a higher level than expected. The pro-poor bias in service
uptake holds when factors such as demography, service accessibility
and rurality are controlled for.
- Age and sex
emerge as significant determinants of service access. Age-specific
use/need ratios for emergency and elective admissions for coronary
heart disease and for key procedures, like percutaneous transluminal
coronary angioplasties (PCTAs) and coronary artery bypass grafts
(CABGs), fall off progressively with age. Men are more likely
to receive treatment than women relative to need in all age categories.
- These results
suggest that consideration should be given to how populations
living in peripheral rural areas can obtain fair access to specialist
care. However, the assumption that higher rates of hospital care
are necessarily a 'good thing' should also be revisited. Within
the urban setting, the pro-poor bias in levels of hospitalisation
may reflect poorer primary and community management, suggesting
a need for a stronger public health focus. Differences in treatment
according to age and sex should also be investigated further to
ensure that such differences are clinically justified. To this
end, more robust approaches to monitoring fair access will be
required.
- In addition
to applying the indicative prevalence methodology to investigations
of health service use in relation to need, the project demonstrated
the feasibility and impact of using this method to set health
care capitations for specific clinical programmes. If resource
allocation is to be used to target health inequalities, it is
important to distinguish between funding streams for the treatment
of existing disease and funding streams that are directed towards
prevention (including the strengthening of primary and community
management).
- Symptom-based
capitation estimates for the treatment of coronary heart disease
within individual primary care organisations (PCOs) have been
compared against actual resource use and against that which would
be allocated using the Hospital and Community Health Services
(HCHS) formula. Compared to current expenditure, the HCHS formula
results in a slight shift of resources towards PCOs serving less
deprived populations, whilst a needs-based capitation methodology
results in a substantial shift of resources away from disadvantaged
areas.
- The indicative
prevalence approach provides a methodology with which to target
areas where populations are at risk of developing premature illness.
However, such an allocation process would have to be underpinned
by decisions concerning the appropriate balance between the treatment
of existing disease and programmes to reduce inequalities in the
determinants and prevalence of disease, as well as about the content
of such programmes.
|