Journal of Public Health | Vol. 28, No. 2, pp. 111–115 | doi:10.1093/pubmed/fdi079 | Advance Access Publication 10 March 2006
Emergency call work-load, deprivation and population
density: an investigation into ambulance services across
England
Philip J. Peacock1, Janet L. Peacock2
1
University of Bristol, Bristol, UK
School of Health Sciences and Social Care, Brunel University, London, UK
Address correspondence to Janet L. Peacock, E-mail: [email protected]
2
ABSTRACT
Demand for emergency ambulance services has risen steeply over the recent years. This study examined differences in work-load of ambulance
services across England and investigated factors linked to high demand. The number of emergency calls received by each ambulance service in
1997 and 2002 and population and area data were used to calculate call rates and population density for each of 27 service areas. Deprivation
score and proportion of the population under age 15 and over age 65 were calculated for each service area. There was wide variation in emergency call rates across England, with London having the highest rate both in 1997 (125.6 calls per 1000 persons) and in 2002 (140.1 per 1000).
Statistically significant positive associations were observed between call rates and deprivation (1997, r = 0.49; 2002, r = 0.53) and between call
rates and population density (1997, r = 0.70; 2002, r = 0.68). Following multivariable regression, the effect of deprivation score was consistently
weaker, but the effect of population density was virtually unchanged. We conclude that areas with higher population density have higher call
rates, which is not explained by deprivation. Deprivation is associated with higher usage, but its effect is partly due to population density. There is
no evidence that these relationships are confounded by age.
Keywords ambulances, population, emergency medical services, deprivation
Introduction
Department of Health (DH) statistics show that demand for
emergency ambulance services has been increasing steeply in
recent years.1 However, little has been published about factors linked to high service demand or about variations in
demand across the country. Carlisle et al. found that the use
of general practice and hospital accident and emergency services varied with deprivation,2 but their study did not examine
ambulance services and only looked at one city, Nottingham.
Wass and Zoltie reported that increased use of accident and
emergency departments is disproportionately high among
elderly patients.3
This article reports the findings of a study of ambulance
service usage in England, which has investigated the effects
of deprivation, population density and age distribution of the
population.
obtained from the DH.1 These years were chosen because
2002 was the most recent year for which government ambulance data were available at the time the study was conducted
and 1997 provided a comparison with data taken 5 years
before. Using population and area data available from the
Office for National Statistics (ONS)4,5 and information
about each service’s coverage, the area and population served
by each ambulance service were calculated. From these data,
population densities and call rates were calculated for both
1997 and 2002. Deprivation scores for each service area were
calculated by averaging the index of multiple deprivation
(IMD) scores for each of the component districts and unitary
authorities, weighted by the population. The IMD scores,
obtained from ONS, were produced in 2000 and took into
account six factors: income; employment; health and disability; education, skills and training; housing; and geographical
Methods
For each of the 32 ambulance services in England, the
number of emergency calls received in 1997 and 2002 was
Philip J. Peacock, Fourth year Medical Student
Janet L. Peacock, Professor of Health Statistics
© The Author 2006, Published by Oxford University Press on behalf of Faculty of Public Health. All rights reserved.
111
2874
8107
Bedfordshire & Hertfordshire
Berkshire; Oxfordshire;
Hereford & Worcester
3487
1559
3615
3783
Shropshire
South Yorkshire
Staffordshire; West Midlands
Sussex
3476
West Country
Wiltshire
1979
13 701
Warwickshire
West Yorkshire
13 426
7995
Northeast
Tees, East & North Yorkshire;
1580
2986
6946
Lincolnshire
Mersey Regional
3069
Lancashire
London
3735
Kent
380
3923
Hampshire; Surrey
Isle of Wight
1286
5456
Greater Manchester
3675
2653
Gloucestershire
7340
East Midlands
Essex
12 570
2653
Dorset
East Anglian
6824
Cumbria
Two Shires
1331
Avon
(sq km)
service(s)
14.30
24.19
16.92
29.44
18.47
32.88
37.34
21.98
36.78
35.52
28.84
24.94
29.97
20.19
29.34
17.93
12.80
35.61
15.95
19.34
25.69
19.45
18.22
25.59
13.61
14.66
19.34
score (IMD)
Area served Deprivation
Ambulance
Table 1 English ambulance services: summary data
595 100
2 019 100
497 300
3 908 100
1 463 700
3 634 500
1 277 300
426 500
2 008 100
2 365 600
6 927 800
933 600
1 411 300
1 548 100
127 000
702 500
2 661 900
2 502 500
554 900
1 579 300
2 883 900
2 158 300
681 100
489 200
2 665 200
1 558 700
967 000
Population
1997
Population density
171.2
147.4
251.3
291.1
386.9
1005.4
819.3
122.3
251.2
792.2
4384.7
134.4
459.9
414.5
334.2
179.1
487.9
1946.0
209.2
429.7
392.9
171.7
256.7
71.7
328.8
542.3
726.5
(persons per sq km)
22 700
105 800
23 300
282 900
120 200
294 500
78 200
19 800
151 000
193 700
869 900
39 500
107 700
88 400
6600
29 800
196 800
222 500
23 600
103 000
191 700
96 900
44 900
20 700
137 700
99 700
53 400
Calls
Call rate (calls per
38.1
52.4
46.9
72.4
82.1
81.0
61.2
46.4
75.2
81.9
125.6
42.3
76.3
57.1
52.0
42.4
73.9
88.9
42.5
65.2
66.5
44.9
65.9
42.3
51.7
64.0
55.2
1000 persons)
616 900
2 090 200
512 700
3 944 700
1 500 100
3 623 400
1 267 300
445 800
1 972 100
2 347 100
7 355 300
969 500
1 421 900
1 589 200
134 900
722 400
2 713 500
2 491 200
567 000
1 622 400
2 919 500
2 191 000
696 300
488 500
2 737 800
1 607 000
987 000
Population
2002
Population density
177.5
152.6
259.1
293.8
396.5
1002.3
812.9
127.8
246.7
786.0
4655.3
139.6
463.3
425.5
355.0
184.1
497.3
1937.2
213.7
441.5
397.8
174.3
262.5
71.6
337.7
559.2
741.5
(persons per sq km)
84 100
91 500
132 500
144 800
11 100
43 500
258 500
292 000
47 900
141 800
301 600
156 800
67 300
33 900
207 900
137 700
47 200
157 600
43 000
388 600
159 400
410 300
127 600
35 100
197 800
247 400
1 030 600
Calls
Call rate (calls per
76.5
75.4
83.9
98.5
106.3
113.2
100.7
78.7
100.3
105.4
140.1
94.4
93.2
91.1
82.3
60.2
95.3
117.2
84.5
87.4
103.3
71.6
96.7
69.4
75.9
85.7
85.2
1000 persons)
112
JOURNAL OF PUBLIC HEALTH
EMERGENCY AMBULANCE WORK-LOAD ACROSS ENGLAND
access to services [http://www.statistics.gov.uk accessed
February 2004]. A higher IMD score indicates greater
deprivation.
In some cases, the division between two ambulance
services cut across a district. As population data were not
available below district level, in these instances, data from
adjacent ambulance services were combined. In three cases,
data from two ambulance services were merged together,
and in one case, three adjacent services had to be combined.
This gave 27 ambulance services to be analysed.
For both 1997 and 2002, the relationship between call rate
and other variables was tested using Pearson’s correlation.
For the correlation tests, the natural logarithm of the call
rates was used, as this best fitted a Normal distribution. All
variables that were significantly related to call rate were fed
into a multivariable linear regression to attempt to disentangle
the relationships.
Statistical analyses were performed using Stata version 7.
113
In 2002, London was again the highest (140 calls/1000), with
Hereford & Worcester having the lowest call rate (60 calls/
1000 people). In both 1997 and 2002, London’s call rate was
far higher than the service with the second highest rate.
Table 3 summarizes the correlations between call rate and
deprivation, population density and two age categories. Fig. 1
shows scatter plots of call rates by population density and
deprivation. As London was a major outlier, correlations
were repeated with this area excluded. There was a moderately
Table 2 Summary statistics
Variable
Mean
SD
Min
Max
125.6
1997 call rate*
62.8
19.4
38.1
2002 call rate*
91.6
16.8
60.2
140.1
1997 population density†
581.8
850.5
71.7
4384.7
2002 population density†
596.7
895.2
71.6
4655.3
23.7
7.6
12.8
37.3
IMD‡
Results
*Call rates are calls per 1000 population.
Call rates varied widely across the country (Tables 1 and 2).
In 1997, London had the highest call rate (126 calls/1000
people per year) and Wiltshire had the lowest (38 calls/1000).
Fig. 1 Scatter plots of call rates by population density and deprivation.
†Population densities are persons per sq km.
‡IMD (index of multiple deprivation) is a measure of deprivation; the higher
the number, the greater the deprivation.
114
JOURNAL OF PUBLIC HEALTH
Table 3 Correlations between call rate* and other variables
1997
2002
All services
Variable
London excluded
Correlation
P-value
Correlation
All services
P-value
London excluded
Correlation
P-value
Correlation
P-value
Deprivation score (IMD)†
0.49
0.01
0.49
0.01
0.53
0.0046
0.54
0.0048
Population density
0.70
<0.0001
0.64
0.0005
0.68
0.0001
0.63
0.0006
−0.21
0.30
−0.06
0.76
−0.30
0.13
−0.12
0.58
0.38
0.05
0.34
0.09
0.22
0.27
0.24
0.23
(persons per sq km)
Proportion of over 65s
Proportion of under 15s
*A logarithmic transformation was applied to call rate to best fit a Normal distribution; call rates are calls per 1000 population.
†IMD (index of multiple deprivation) is a measure of deprivation; the higher the number, the greater the deprivation.
Table 4 Regression between call rate, deprivation and population density
1997
Outcome : call rate*
Predictor variable
2002
Unadjusted regression
Adjusted regression†
Regression
Regression
P-value
coefficient (SE)
Deprivation
Adjusted regression†
Unadjusted regression
P-value
P-value
coefficient (SE)
Regression
P-value
coefficient (SE)
0.14 (0.05)
0.010
0.08 (0.04)
0.059
0.10 (0.03)
0.005
0.06 (0.03)
0.021
0.20 (0.04)
<0.001
0.17 (0.04)
<0.001
0.12 (0.03)
<0.001
0.10 (0.03)
0.001
score (IMD)
Population density
(persons per sq km)
*A logarithmic transformation was applied to call rate to best fit a Normal distribution; call rates are calls per 1000 population.
†The model included deprivation score and population density as predictors of call rate and hence each adjusted coefficient is adjusted for the other
predictor in the model.
strong positive correlation between call rate and deprivation
for both 1997 and 2002 (r = 0.49 and 0.53 respectively).
There was a stronger positive correlation with population
density for both years (r = 0.70 and 0.68). These associations
were slightly weaker when London was excluded from the
analysis.
For both years, there was a negative correlation between
call rate and the proportion over age 65, but the relationship
was weak and non-significant. The proportion under age 15
was positively related to call rate, but all correlations were
weak and mostly non-significant (Table 3).
Table 4 summarizes the results of multiple regressions on
call rate. Predictor variables were deprivation score and
population density; age was not included in the models as it
was non-significant in the unifactorial analyses. For the
1997 data, the relationship between call rate and deprivation
was weaker (regression coefficient halved) and became
non-significant (P = 0.06) after adjusting for population density, whereas the relationship with population density was
hardly changed. A similar pattern was seen in 2002, with the
relationship with deprivation becoming weaker after adjustment and that with population density not changing although
both factors remained statistically significant.
Discussion
Main finding of this study
This study has shown that there is considerable variation in
call rates between areas and that London has a far higher
call rate than any other area of the country. The variation in
call rates is partly explained by deprivation and population
density, which are both positively related to call rates. The
effect of deprivation was partly due to the relationship with
EMERGENCY AMBULANCE WORK-LOAD ACROSS ENGLAND
population density, but the effect of population density
is not explained by deprivation. There is no evidence
that these effects are confounded by differences in age
distributions.
What is already known on this topic
Demand for ambulance services has increased substantially
in recent years, with call rates doubling between 1989 and
1999 in London.6 Government statistics show that this trend
is continuing.1 There is evidence that utilization of health
services is linked to levels of deprivation.2
What this study adds
To our knowledge, there have been no studies of English
ambulance services, which have investigated variation in
demand across the country and factors related to the variability. This study explores and quantifies the variability in call
rates across England at two time points, five years apart, and
shows consistent positive relationships with deprivation and
population density.
As ambulance services across England are faced with
increasing demand for emergency ambulance services, it is
important to know what factors are linked to high service
demand. Further research is needed to investigate why
people living in high population density areas and those with
greater levels of deprivation are making greater use of ambulance services. A clear understanding of reasons for high
usage of emergency ambulance services is necessary to best
direct attempts to meet genuine needs and to reduce inappropriate usage of the English ambulance services.
Limitations of this study
This study has certain limitations. The call rates were not age
standardized, as age-specific call data could not be obtained
from either the DH or individual ambulance services. As a
substitute, we used both the proportion of the population
under 15 and proportion over 65 to investigate differences in
age distribution between service areas. Since there was no
evidence of a relationship between the age distribution and
call rate, it is unlikely that the relationships between deprivation, population density and call rate are because of children
or the elderly using services more.
115
Another limiting factor is that the call rates were based
solely on the resident population in each area. The number
of non-resident workers was not taken into account, a factor
that could possibly explain London’s high call rate. However,
previous research has shown that use of ambulance services
in London has increased most outside of working hours, when
non-resident workers would be unlikely to be service users.6
This study used available routine data in an ecological
analysis investigating effects of deprivation and population
density. However, these two factors only explained
56% of the total variability in call rates between areas, and
hence a substantial proportion of the variability remained
unexplained.
Acknowledgements
This study was originally conducted to fulfil a ‘student
selected component’ for the MBChB course at the University
of Bristol.
We thank Professor Yoav Ben-Shlomo from the Department of Social Medicine at Bristol, for his help throughout
the original project.
The authors declare there are no conflicts of interest.
References
1 Department of Health. Statistical bulletin 2003/13: Ambulance services, England: 2002–3. 2003. London.
2 Carlisle R, Groom LM, Avery AJ, Boot D, Earwicker S. Relation of
out of hours activity by general practice and accident and emergency
services with deprivation in Nottingham: longitudinal survey. Br Med J
1998;316:520–3.
3 Wass A, Zoltie N. Changing patterns in accident and emergency
attenders. J Accid Emerg Med 1996;13:269–71.
4 Office for National Statistics. Mid-1997 Population Estimates; Quinary age groups and sex for health authorities in England and Wales;
estimated resident population revised in light of results of the 2001
Census. 2002. London.
5 Office for National Statistics. Mid-2002 Population Estimates; Quinary age groups and sex for health authorities in England and Wales.
2002. London.
6 Peacock PJ, Peacock JL, Victor CR, Chazot C. Changes in the emergency workload of the London Ambulance Service between 1989 and
1999. Emerg Med J 2005;22:56–9.