Socio-economic determinants of casualty and NHS Direct use
Journal of Public Health |
Socio-economic determinants of casualty and NHS Direct use
S. M. Shah 0
D. G. Cook 0
0 Division of Community Health Sciences, St George's University of London , London SW17 0RE , UK
A B S T R AC T Background There is limited evidence on the social determinants of A&E use and concerns over the equity of NHS Direct utilization. Methods We analysed data from the 2004 - 05 British General Household Survey, which included 20 421 participants. Logistic regression was used to examine individual casualty use in the last 3 months and household NHS Direct use in the last year. Results Casualty use was higher for individuals living in rented accommodation or without car access, lower income groups, unskilled manual workers, current smokers and for individuals with limiting illness. In contrast, NHS Direct use was lower in households with older residents, low income, no car access and where the head of household was from a manual occupational group, a minority ethnic group or born outside the UK. The odds ratio for use of NHS Direct for households in the lowest equivalized income quintile was 0.67 (0.55 - 0.81). Adjustment for limiting illness increased the effect of socio-economic factors on NHS Direct use. Conclusions Reduced access to A&E services will disproportionately affect poorer individuals, whereas increased investment in telephone services will benefit affluent populations. Current national policy may widen inequities in access to emergency care.
Emergency care is one of the most contentious areas of
NHS provision with increasing pressure for centralization of
A&E services.1 Accident and emergency or casualty is the
most widely used hospital service with 18 million new
attendances in 2006 – 07, but current NHS systems or
research databases do not allow a reliable description of
national utilization patterns.2 Little is known about the
epidemiology and determinants of A&E attendance, but
available evidence suggests higher utilization in more
deprived areas.3 This compares with primary care and
inpatient utilization, which have been widely studied using a
range of routine and research sources.4 – 6 These studies
suggest that overall primary care and emergency inpatient
utilization increase with decreasing socio-economic status,
reflecting higher morbidity, but also indicate more complex
variations for specific services or types of care.3 – 6
NHS Direct has provided emergency telephone advice
nationally since 2000 and received 6.8 m calls in England in
2005 – 06. Early commentary on NHS Direct, including a
National Audit Office report, raised concern that certain
groups may be less likely to use the service despite higher
need.7,8 There is some evidence from local surveys and
ecological analysis on the effect of deprivation on NHS Direct
use, but these do not give a clear pattern.9 – 12 A recent
survey in a single region suggests that NHS Direct use is
positively related to socio-economic position and
circumstances with greatest use among the most affluent.13
National and local policy on emergency care provision
needs to be guided by an understanding of utilization
patterns and the impact of changes in provision on equity of
access. In this paper, we present the first national
description of casualty and NHS Direct use in the same population
using a secondary analysis of data from the General
Household Survey (GHS).
General Household Survey
The GHS is an annual survey of private households in
Great Britain, which is undertaken by the Office for
National Statistics to support planning by government
departments and other agencies.14 The dataset for the
survey is made available by the UK Data Archive to
researchers to allow secondary analysis for specific research
The GHS selects households using probability stratified
clustered sampling to ensure that they represent the general
population. In 2004 – 05, 12 149 eligible households were
sampled and 8700 (71.6%) households participated. In 7659
households, all adult members were directly interviewed
with complete coverage in a further 790 households
achieved through proxy interviews. In 251 households, at
least one adult member was not included in the survey. This
gave a household response rate of 71.6% and a complete
individual response rate of 63.0%. Information on children
was obtained from adults in the household. In total, 16 175
adults and 4246 children were included in the survey.
The GHS asked the participants about their use of health
services in defined periods before the interview. All
participants, including children and proxy interviewees, were asked
whether they had visited a hospital casualty department as a
patient in the last 3 months.
Adults who completed a direct interview were asked
whether they had used NHS Direct (NHS 24 in Scotland)
in the last 12 months. This question did not distinguish use
for the individual interviewed or for another member of the
household, including children. Therefore, NHS Direct use
was analysed at household level based on whether any adult
household member had used NHS Direct. Only households
where information on all adult household members was
available were included.
The GHS asks a range of questions on individual and
household circumstances including income, receipt of
benefits, type of employment, tenure and access to material
goods. In addition, the survey asks about individual health
including whether an individual has long-standing illness,
disability or infirmity, and whether this limits activity.
For household analysis, appropriate individual level
variables, such as self-reported long-standing illness or smoking,
were aggregated to identify households where at least one
individual reported illness or smoking. For ethnicity and
country of birth, the status of the household reference
person (HRP) was used for analysis. The HRP is the
member of the household in whose name the
accommodation is owned or rented or, in the case of joint ownership,
the member with the highest income.
All analyses were undertaken taking account of the sampling
design of the study including weighting, clustering at area
and household level and regional stratification using SVY
commands in Stata version 9.2. Determinants of casualty
and NHS Direct use were examined using logistic
Individual casualty use
Responses on casualty use were available for 20 106 (98.5%)
individuals and 876 (4.4%, 4.1 – 4.7%) reported use as a
patient in the last 3 months. 5.0% (4.5 – 5.4%) of men and
3.8% (3.4 – 4.2%) of women reported casualty use in the last
3 months. Casualty use was highest among children under
five, teenagers and young adults, and lowest between the
ages of 35 and 54.
Household NHS Direct use
In 7634 (87.7%) households all adults answered the
question on NHS Direct use with 1624 households (20.7%,
19.7 – 21.8%) reporting use in the last year. Households,
which had used NHS Direct, were more likely to have a
member attend casualty (OR 2.17, 1.8 – 42.87).
Determinants of casualty use
At the individual level, after adjustment for age and sex,
unskilled manual social class, living in rented
accommodation, lower household income, household receipt of
income support, lack of access to a car and current smoking
significantly increased the likelihood of casualty use
(Table 1). Long-term limiting illness was the strongest
predictor of casualty use (OR 1.93, 1.60 – 2.32). After further
adjustment for long-term illness and region, living in
privately rented accommodation, lower income and smoking
remained significant predictors of casualty use. The effect of
other determinants was attenuated by control for
longstanding and limiting illness but the trend remained for
most socio-economic indicators.
Determinants of household NHS Direct use
NHS Direct use was higher in larger households and those
with children but lower in households with older people
(Table 2). After adjustment for these factors, measures of
material deprivation and social status including lack of
access to a car, low-equivalized household income, living in
social housing, manual occupational group and receipt of
Table 1 Odds ratios for individual casualty use in the last 3 months
Table 1 Continued
1.04 (0.75 – 1.44)
1.02 (0.74 – 1.41)
1.08 (0.78 – 1.52)
1.28 (0.89 – 1.82)
1.07 (0.76 – 1.49)
1.21 (0.84 – 1.73)
1.56 (1.01 – 2.42)
1.45 (0.94 – 2.24)
aAnalysis weighted with account for clustering at postcode area and
income support all significantly reduced the likelihood of
NHS Direct use. In addition, NHS Direct use was markedly
lower in households where the head of household was not
white or born outside the UK. Long-standing and limiting
illness in the household both predicted NHS Direct use.
After control for household long-standing and limiting
illness, the relationship between material deprivation and
social status and NHS Direct use strengthened with an OR
of 0.58 (0.48 – 0.71) for the poorest households compared
with the most affluent.
Analysis of individual NHS Direct use was consistent
with findings from the household level. After control for
age, sex, long-standing illness and household structure,
Ethnicity of HRP
Country of birth of HRP
Skilled manual 1805
Semi-skilled manual 945
Unskilled manual 388
Own or use of car
Household income quintile
Household income support
1.27 (1.19 – 1.36) 1.24 (1.16 – 1.33)
2.58 (2.08 – 3.20) 2.83 (2.27 – 3.53)
1.11 (0.91 – 1.36) 1.13 (0.92 – 1.39)
0.50 (0.41 – 0.62) 0.47 (0.38 – 0.58)
0.30 (0.23 – 0.39) 0.27 (0.21 – 0.35)
0.56 (0.38 – 0.81) 0.57 (0.38 – 0.85)
0.36 (0.2 – 10.62) 0.38 (0.21 – 0.70)
0.65 (0.40 – 1.05) 0.68 (0.41 – 1.11)
0.50 (0.40 – 0.62) 0.52 (0.41 – 0.66)
0.96 (0.73 – 1.23) 0.92 (0.71 – 1.20)
1.11 (0.87 – 1.42) 1.09 (0.85 – 1.39)
0.89 (0.68 – 1.15) 0.82 (0.63 – 1.07)
0.71 (0.53 – 0.96) 0.65 (0.48 – 0.88)
0.71 (0.49 – 1.03) 0.62 (0.42 – 0.91)
0.77 (0.69 – 0.87) 0.72 (0.64 – 0.82)
0.66 (0.56 – 0.76) 0.64 (0.54 – 0.75)
0.69 (0.59 – 0.81) 0.64 (0.55 – 0.76)
1.10 (0.93 – 1.33) 1.15 (0.96 – 1.38)
1.03 (0.86 – 1.23) 0.98 (0.82 – 1.17)
0.88 (0.73 – 1.05) 0.80 (0.66 – 0.96)
0.88 (0.73 – 1.08) 0.77 (0.63 – 0.94)
0.67 (0.55 – 0.81) 0.58 (0.48 – 0.71)
0.75 (0.60 – 0.96) 0.65 (0.51 – 0.82)
Table 2 Odds ratios for household NHS Direct use in the last year
Table 2 Continued
1.25 (0.85 – 1.80) 1.22 (0.85 – 1.77)
1.47 (1.15 – 1.89) 1.45 (1.13 – 1.86)
1.35 (1.05 – 1.78) 1.33 (1.02 – 1.74)
1.33 (1.01 – 1.75) 1.29 (0.98 – 1.70)
1.04 (0.79 – 1.36) 1.04 (0.79 – 1.35)
1.10 (0.86 – 1.40) 1.09 (0.86 – 1.39)
0.88 (0.67 – 1.14) 0.92 (0.70 – 1.20)
1.32 (1.00 – 1.75) 1.31 (0.99 – 1.72)
0.73 (0.50 – 1.05) 0.72 (0.50 – 1.03)
0.83 (0.62 – 1.13) 0.84 (0.62 – 1.13)
1.40 (1.20 – 1.63) 1.36 (1.16 – 1.59)
1.58 (1.38 – 1.81) 1.54 (1.34 – 1.77)
1.10 (0.95 – 1.28) 1.03 (0.89 – 1.20)
1.19 (1.02 – 1.40) 1.12 (0.96 – 1.32)
aAnalysis weighted with account for clustering at postcode area.
the OR for adults living in the poorest households was
0.74 (0.61 – 0.90) compared with those in the most affluent.
We found that lower income, measures of material
deprivation and lower socio-economic position were associated
with higher casualty use and part of this relationship could
be explained by higher levels of long-standing and limiting
illness. Conversely, lower income, material deprivation and
low socio-economic position were associated with low
household use of NHS Direct. In addition, our findings
confirm long-standing concerns that NHS Direct is
markedly underused by older people, ethnic minority households
and people born outside the UK. We believe that this is the
first national study to report on casualty or NHS Direct use
in relation to individual, rather than ecological, measures of
socio-economic status. Our ability to study casualty and
NHS Direct use in the same population allows the first
direct comparison of determinants of the use of these
complementary services. As our study is based on
secondary analysis of survey data with potential non-response and
reporting bias, it is important that they are confirmed in
Limitations of this study
Our results rely on self-reported health service utilization. It
is likely that both casualty and NHS Direct use are reliably
remembered by individuals, but we cannot exclude
incomplete or biased recall. This is likely to be a greater problem
with NHS Direct use, which required recall over a period of
1 year. However, to explain our divergent findings on
socioeconomic determinants, one would have to postulate a
different direction of recall bias for casualty and NHS Direct use.
We adjusted for health need using self-reported,
longstanding and limiting illness. This was based, a priori, on the
strong association between these measures and other health
service use including primary care and secondary care. We
also found these measures to be strongly associated with
casualty and NHS Direct use. There is concern that
selfreported measures may underestimate need in more affluent
populations compared with more objective measures and
this could partially explain our findings.6 However,
longstanding limiting illness is probably the best available proxy
in survey data and is used in the needs element of the
funding formula for primary care.15
The question on NHS Direct use in the GHS does not
distinguish use for self or for others, including children. We have
therefore analysed at household level rather than individual
level. This does not allow for control for individual
determinants of use, such as age, but we were able to control for them
indirectly through household structure. Furthermore, our main
measures of income, social and material statuses are
appropriately measured at household level.
Our measure of service use does not capture repeated
use of services by individuals and so will not equate to
actual service demand. However, for examining the
relationship between socio-economic factors and service use, our
approach may be preferable as repeated use may capture a
different aspect of need related to quality of community
support and follow-up, as well as underlying health need.
What is known already
Our findings on NHS Direct use are consistent with a
recent regional survey on NHS Direct use, which showed an
inverse relationship with measures of social position and
material circumstances.13 This contrasts with ecological
studies, which have shown inconsistent findings with either
increased use in the most deprived areas in adults but not
children or have reported a complex non-linear relationship
or no relationship with deprivation.9 – 11 All these ecological
studies utilized call rates and measures of deprivation at
ward level. Our divergent findings may be explained by our
better measures of socio-economic status at household level
or that call rates are determined by frequency of calls as well
as whether the service was used by a household.
There is very limited information on the epidemiology of
overall casualty use in the UK. One study has examined the
epidemiology of A&E use by older people using routine
data in one region.16 This reported very high utilization
rates among those aged over 80, which we did not
demonstrate. This may be explained by the relatively small number
of participants over the age of 80 in the GHS and exclusion
of older people in care homes from the sampling frame.
Most studies on socio-economic determinants of casualty
use have focused on attendances for injury or childhood
utilization in limited geographical areas.17 – 19 These studies
have reported increasing utilization with increasing
deprivation. Our study confirms this for all attendances and
further analysis has confirmed the same pattern for adults
alone. We have not been able to control for distance from
A&E, which may influence A&E attendance.3 However,
control for region, a possible proxy for dispersion of A&E
units, did not modify our findings.
What this study adds
NHS Direct provides high levels of patient satisfaction, and
a large number of calls received suggests that it is popular
and valued.20 However, evaluation has confirmed that its
impact on demand for other emergency services is limited
and it has instead increased the capacity for emergency
advice provided by the NHS rather than substituting for
other services.21 Our findings suggest that NHS Direct has
inadvertently increased age, socio-economic and ethnic
inequity in NHS provision. It is possible that NHS Direct is
meeting distinct health needs, which are differently socially
patterned from the needs met by existing health services.
We feel this is unlikely as evidence on socio-economic
variations of morbidity, accidents and self-reported health
almost invariably suggests higher need in more
socioeconomically deprived populations.22 Our findings are
consistent with the inverse care pattern described for
preventative consultations in primary care.5 We were not able to
examine reasons for calls to NHS Direct, but it is possible
that NHS Direct is meeting the demand for reassurance or
preventative advice among younger and more affluent
groups who would otherwise self-manage their concerns.
This interpretation is supported by findings on satisfaction
with NHS Direct.20
Our findings may explain why NHS Direct has had a
limited impact on use of other health services as the service
is under utilized by groups, including older people, with the
highest general practice and A&E utilization. Future
development of NHS Direct needs to address these inequities.
This is likely to require a more detailed understanding of the
reason and social patterning of calls and an understanding
of the barriers to use of NHS Direct.
The UK government is currently developing its strategy for
urgent and emergency healthcare. This will address
pressures for centralization of emergency services to meet
requirements for sub-specialization and access to new
technology interventions. It is likely the strategy will support
centralization of major A&E services with provision of local
care alternatives through emergency care networks.23
Our findings suggest that the utilization of current A&E
services does reflect the expected socio-economic pattern of
health needs in the population, although we cannot say
whether the services adequately or appropriately meet the
needs of all groups. This compares starkly with NHS
Direct, which seem to disproportionately serve populations
with the lowest expected need. Furthermore, our study only
included individuals in households and it is likely that
groups living outside private household, including those
without homes and migrants, will be more reliant on direct
access rather than telephone or booked services.
Our findings need to be confirmed in specifically designed
studies, which address our study limitations, particularly in
relation to the measures of service utilization and coverage of
the whole population. If confirmed, our findings suggest that
proposals for telephone-based triage to optimize the
utilization of A&E may introduce inequity into emergency care
utilization and further investment in telephone-based
emergency services risks increasing inequity in NHS provision. In
addition, our work suggests that alternatives to A&E should
continue to provide the model of high-profile direct walk in
access provided by A&E, rather than telephone access, to
ensure access to care for those with greatest need.
This research is based on the data from the British General
Household Survey, which were produced by the Office for
National Statistics and made available by the Economic and
Social Data Service through the UK Data Archive. The
GHS data are crown copyright. Neither the Office for
National Statistics, Social Survey Division, nor the Data
Archive, University of Essex, bears any responsibility for the
analysis or interpretation of the data described in this paper.
Downing A, Wilson R. Regional surveillance of Accident and
Emergency Department attendances: experiences from the west
Midlands. J Public Health 2005;27:82 – 4.
George S. NHS Direct audited. BMJ 2002;324:559 – 60.
1 Boyle R. Mending hearts and brains - Clinical case for change . London, UK: Department of Health , 2006 .
3 Carlisle R , Groom LM , Avery AJ et al. Relation of out of hours activity by general practice and accident and emergency services with deprivation in Nottingham: longitudinal survey . BMJ 1998 ; 316 : 520 - 3 .
4 Reid FDA , Cook DG , Majeed A. Explaining variation in hospital admission rates between general practices: cross sectional study . BMJ 1999 ; 319 : 98 - 103 .
McCormick A , Fleming D , Charlton J. Morbidity statistics from general practice: fourth national study 1991 - 92 . London, UK: HMSO, 1995 .
Dixon A , Le Grand J , Henderson J et al. Is the NHS equitable? A review of the evidence . London, UK: London School of Economics, 2003 . ( Health and Social Care discussion paper No . 11).
7 The Comptroller and Auditor General . NHS Direct in England . London, UK: Stationery Office , 2002 .
9 Burt J , Hooper R , Jessopp L. The relationship between use of NHS Direct and deprivation in southeast London: an ecological analysis . J Pub Health Med 2003 ; 25 : 174 - 6 .
10 Cooper D , Arnold E , Smith G et al. The effect of deprivation, age and sex on NHS Direct call rates . Br J Gen Pract 2005 ; 55 : 287 - 91 .
11 Ring F , Jones M. NHS Direct usage in a GP population of children under 5 years: is NHS Direct used by people with the greatest need? Br J Gen Pract 2004 ; 54 : 211 - 13 .
12 Ullah W , Theivendra A , Sood V et al. Men and older people are less likely to use NHS Direct . BMJ 2003 ; 326 : 710 .
13 Knowles E , Munro J , O' Cathain A et al. Equity of access to health care. Evidence from NHS Direct in the UK . J Telemed Telecare . 2006 ; 12 : 262 - 5 .
14 Office for National Statistics. Social and Vital Statistics Division, General Household Survey , 2004 - 05 [computer file]. Colchester, Essex: UK Data Archive [distributor], May 2006 . SN: 5346.
15 Department of Health. Global sum allocation formula . London, UK: Department of Health , 2004 .
16 Downing A , Wilson R. Older peoples use of Accident and Emergency services . Age Ageing 2005 ; 34 : 24 - 30 .
17 Beattie TF , Gorman DR , Walker JJ . The association between deprivation levels, attendance rate and triage category of children attending a children's accident and emergency department . Emerg Med J 2001 ; 8 : 110 - 11 .
18 Laing GJ , Logan S. Patterns of unintentional injury in childhood and their relation to socio-economic factors . Public Health 1999 ; 113 : 291 - 4 .
19 Brown CE , Chishti P , Stone DH . Measuring socio-economic inequalities in the presentation of injuries to a paediatric A&E department: the importance of an epidemiological approach . Public Health 2005 ; 119 : 721 - 25 .
20 O ' Cathain A , Munro JF , Nicholl JP et al. How helpful is NHS Direct? Postal survey of callers . BMJ 2000 ; 320 : 1035 .
21 Munro J , Nicholl J , O' Cathain A et al. Impact of NHS Direct on demand for immediate care: observational study . BMJ 2000 ; 321 : 150 - 53 .
22 Acheson D. Independent inquiry into inequalities in health. Report . London, UK: The Stationery Office, 1998 .
23 Alberti G. Emergency care networks . BMJ 2004 ; 329 : 1057 - 58 .