Ethnicity and health: Deficiencies in ethnic monitoring data
Challenges
The London Health Observatory report Ethnic
Disparities in Health and Health Care, published in
October 2003, highlights some of the deficiencies in ethnicity data
which impede progress in describing and monitoring ethnic inequalities in health
and health care. For example:
 |
Within the hospital sector where ethnic monitoring
has been mandatory since 1995, there is still enormous variation
in meeting minimum coding requirements. |
|
Whilst there is NHS commitment to implement the requirements
of the new Race Relations Amendment Act, the strategy for doing so
is inconsistent. |
|
Because of the poor quality data, there has been a reluctance
to develop analytical tools to help understand what ethnic differentials
in access to health care mean for commissioning services. |
|
Ethnicity data collected by some agencies, such as the
Police and Fire authorities, do not reflect the same ethnic information
as that collected in the census. This makes the analysis of different
sources of information very difficult to interpret. |
|
Ethnicity is not at present recorded on either birth or
death certificates in the United Kingdom. The recording of ethnicity
at birth would facilitate calculation of infant health outcomes as
well as fertility rates for the population by ethnic groups. Ethnicity
recording at death will help to provide answers to important questions
such as whether a particular ethnic group is more affected than others
by particular causes of death. |
Some of the above problems are further elaborated in the
more recent London Health Observatory report on Ethnic
Disparities in Health and Health Care. For example:
|
Hospital episode data for the year 2002-3 shows
that, of the total of 12.9 million episodes in England, 68.1% had
a valid ethnic code. This was an improvement on the previous year.
However, in addition to incompleteness, there may also be problems
with data quality. For example, a significant proportion of records
in England were still using the 1991 Census classification for ethnic
group. Moreover, there is survey evidence that some trusts are using
staff observation to assign ethnicity rather than patient self-identification
as required by Department of Health guidance. |
An analysis of the quality of ethnicity data from East
Midlands NHS providers is attached as Appendix
5.
Addressing the challenges
Following a recent Office for National Statistics (ONS) consultation
on Modernising
Civil Registration,
the inclusion of recording of ethnicity as part of birth and death registration
is now more likely but will not proceed universally without piloting.
Comprehensive births and deaths data with attached ethnicity is thus
still some way off.
The collection of ethnicity data within hospital episode statistics
is seen as an important duty by the Healthcare Commission and is included
within the Commission’s performance
indicator set for Acute and Mental Health Trusts. Active performance
management of Trusts by the Strategic Health Authorities, facilitated
by the Healthcare Commission will hopefully improve the current situation.
The English Strategic Health Authorities' race
equality performance framework for the NHS
provides further useful information on improving data quality including
success criteria.
The Department of Health are facilitating data quality improvement in
a number of ways. For example:
Within the East Midlands, the Local Health Group in Wellingborough
are taking forward a patient profiling project. The project is briefly
described in the Northamptonshire
Heartlands PCT Public Health Annual Report for 2002/03 (page 110).