The Health Effects of Earthquakes in the

Mid-1990s

DAVID ALEXANDER

This paper gives an overview of the global pattern of casualties in earthquakes which
occurred during the 30-month period from 7 September 1993 to 29 Februay 1996. It
also describes some of the behavioural and logistical regularities associated with
mortality and morbidity in these events. Of 83 earthquakes studied, there were
casualties in 49. Lethal earthquakes occurred in rapid succession in Indonesia, China,
Colombia and Iran. In the events studied, a disproportionate number of deaths and
injuries occurred during the first six hours of the day and in earthquakes with
magnitudes between 6.5 and 7.4. Ratios of death to injury varied markedly (though
with some averages close to 1:3), as did the nature and causes of mortality and
morbidity and the proportion of serious to slight injuries. As expected on the basis of
previous knowledge, few problems were caused by post-earthquake illness and disease.
Also, as expected, building collapse was the principal source of casualties: tsunamis,
landslides, debris flows and bridge collapses were the main seconday causes. In
addition, new findings are presented on the temporal sequence of casualty estimates
after seismic disaster. In synthesis, though mortality in earthquakes may have been
low in relation to long-term averages, the interval of time studied was probably typical
of other periods in which seismic catastrophes were relatively limited in scope.

Earthquake epidemiology is a young 1992) and general and theoretical discus-

science which owes much of its develop-
ment to a number of particular disasters.
These include the earthquakes which
struck Guatemala in 1976 (death toll
23,000; Glass et al., 1977), southern Italy in
1980 (death toll 2,735; De Bruycker et al.,
1985), Armenia in 1988 (24,944 deaths;
Wyllie and Filson, 1989), and the Philip-
pines in 1990 (1,600 deaths; Roces et al.,
1992). A wide range of epidemiological
studies of individual disasters has now
been published (e.g. Noji, 1989; Durkin,
1987; Durkin et al., 1991), supplemented
by detailed analyses (e.g. Armenian et al.,
sions (e.g. Jones et al., 1990; Sapir, 1993).
There is room, however, for yet more
information on the general pattern of
mortality and morbidity in earthquakes,
especially with regard to its recent and
current trends.
In the present study, the broad picture
of the health effects of earthquakes was
assessed by studying all significant seismic
disasters that could be identified over the
30-month period beginning on 1 Sep-
tember 1993 and ending on 29 February
1996. Data were compiled from several
hundred wire service news reports pub-

8 Overseas Development Institute DISASTERS VOLUME 20 NUMBER 3

232
Magnitude 5-5.9 6-6.9
FIGURE 1 World distribution of earthquakes that caused casualties, 2 September 1993-29 February 1996
lished over this period by Associated
Press, Reuters, United Press International
and other news services. Of these, 225
contained information on the epidemiolo-
gical and behavioural aspects of earth-
quakes. The information thus obtained
was cross-checked with 104 official reports
issued by the United Nations Department
of Humanitarian Affairs (DHA), which
were based on both national communi-
ques and the work of UN field representa-
tives. The study forms part of a longer-
term effort by the author to collect and
utilise current information on disasters,
which began at the start of the period
analysed and will continue into the future.
Hence the present paper is an interim
report on work in progress.
About 83 earthquakes were reported
to have had significant human effects
during the study period, and 49 of them
led to casualties. The geographical distri-
bution of these events is shown in
DISASTERS VOLUME 20 NUMBER 3
David Alexander
7-7.9 8+
Figure 1. Deaths occurred in at least 40 of
these disasters and injuries in at least 42,
while both deaths and injuries occurred in
at least 33 earthquakes.
Table 1 lists the principal characteris-
tics of those events studied that involved
casualties. Computation of annual aver-
ages gives about 20 earthquakes per year
that caused victims, 16 that caused deaths
and 17 which led to injuries. It also gives a
worldwide annual average mortality of
about 7,960 and morbidity of 35,560. Of
the 49 earthquakes, 31 (63 per cent)
occurred in Asia, including nine in Indo-
nesia, four each in China and Japan and
three in Taiwan. Nine earthquakes led to
casualties in Latin America, and 10 in the
broad area of the Middle East, Greece and
North Africa. Only two earthquakes
caused casualties in North America, and
one each in Africa (Uganda) and Oceania
(Papua New Guinea). In relation to world-
wide earthquake risk, it is possible that
The Health Effects of Earthquakes in the Mid-1990s
North and Central America, Europe and
central Asia are under-represented, but
this depends on where earthquakes
caused deaths during the study period,
not where they occurred, which gives a
different aggregate pattern.
Seven of the events studied generated
tsunamis, with a maximum recorded wave
height of 10-15 m. Six of the tsunamis
occurred in the eastern Pacific basin and
each led to fatalities, while the seventh
tsunami affected Guatemala and did not
cause deaths. Five earthquakes occurred
with landslides (four in the east Pacific and
one in Colombia). Fatalities occurred in
three of these: due to soil slips in Taiwan
in February 1995 and East Timor in May
1995, and as a result of debris avalanches
in Colombia in June 1994. The latter killed
villagers and five Red Cross workers.
BASIC FINDINGS
Figure 2(a) shows that 86 per cent of the
deaths and 97 per cent of the injuries listed
in Table 1 were caused by earthquakes in
the magnitude range 6.5-7.4, although
these constituted only 45 per cent of the
events listed. The most plausible explana-
tion for this discrepancy is that smaller
events were too weak to cause major loss
of life and that larger events were rare and
tended to occur in sparsely populated
areas. However, it is unwise to speculate
too freely without data on, for example,
the duration and acceleration of shaking,
the quality and state of maintenance of
housing stock and the occurrence of
secondary hazards such as fire: in the 1972
earthquake in Nicaragua, these factors
combined to create a death toll of more
than 4,000 on the basis of an event with a
magnitude of only 5.6 (Whittaker et al.,
1974).
Figure 2(b) shows that 94 per cent of
the deaths and 77 per cent of the injuries
occurred between the hours of midnight
and 6 a.m., local time. Even if one
233
excludes the two greatest events, the Latur
earthquake of September 1993 in India and
the Kobe (Japan) earthquake of January
1995,' the figure for deaths remains high at
69 per cent. However, the figure for
injuries is affected by the evening-time
earthquake that occurred in Yunnan Pro-
vince, China, in which 16,925 people were
non-fatally injured in the collapse of a very
large number of weak-walled masonry and
adobe buildings. Therefore, if this event is
excluded, the percentage of injuries that
occurred between midnight and 6 a.m.
rises to 96, but, surprisingly, only 43 per
cent of the earthquakes on the list occurred
during the first six hours of the day. This
suggests that the risks of injury may be
significantly higher at night in many
different physical and architectural
settings.
Previous studies examined the
hypothesis that there is an average of one
death for every three significant injuries in
seismic disasters (PAHO, 1981). On the
other hand, long-term data collected by
the Red Cross suggest a ratio of 1:1.3,
though this probably refers to a count of
only the more serious injuries (IFRCRCS,
1994, p. 150). Some years ago, in a study
of two decades of basic data on earthquake
casualties, Alexander (1985) observed that
there is no standard definition of injury,
no specific injury typology for earth-
quakes, and no threshold for injury sever-
ity to define the level at which victims are
classified as injured. Earthquake-induced
injuries are not internationally notifiable
and governments tend not to have
standard procedures for documenting
them. Despite this, Alexander concluded
that the ratio had some limited validity for
earthquakes in the magnitude range
6.5-7.4.
In the present enquiry, death-injury
ratios varied by several orders of magni-
tude. In part this is because events with
very few casualties did not generate stable
ratios. In addition, there are anomalous
DISASTERS VOLUME 20 NUMBER 3
234 David Alexander

TABLE 1
Earthquakes that caused deaths and injuries, 1 September 1993-29 February 1996

No. Date Location Local Magnitude Deaths Injuries Deaths1 Remarks

time Iniu ries

1 19 Sep 93 Oregon, USA night 5.7 2 0 -

(Klamath Falls)
2 30 Sep 93 Maharashtra, 3.58 6.5 9,475 10,500 0.90
India (Latur)
3 12 Oct 93 N. Japan 12.50 7.1 1 0 -
4 13 Oct 93 Papua New Guinea 12.00 7.2 60 - -
5 20 Oct 93 W. Himalayas, early 5.4 0 6 -
Nepal morning
6 01 Dec 93 China early 6.0 0 4 -
(Xinjiang) morning
7 19 Jan 94 Indonesia 10.54 6.8 7 40 0.17 2 m tsunami
(Irian Jaya)
8 06 Feb94 Uganda 2.45 6.2 7 0 -
9 17 Jan 94 California, USA 4.31 6.6 60 9,202 0.01
(Northridge)
10 16 Feb 94 Indonesia (Liwa) 0.08 6.5 207 2,389 0.09 Landslides
11 23 Feb 94 S.E. Iran 11.34 6.1 9 0 -
12 01 Mar 94 S. Iran 7.19 5.6 3 31 0.10
(Fars Province)
13 01 May 94 N.W. Afghanistan 16.31 6.3 167 300 0.55
14 02 Jun 94 Indonesia 3.17 7.2 222 440 0.50 Tsunami
(E. Java)
15 06 Jun 94 S.W. Colombia 15.47 6.4 271 158 1.72 Debris flows
killed 5 relief
workers
16 20 Jun 94 S. Iran 13.40 6.0 2 loo 0.02
17 18 Aug 94 W. Algeria 2.10 6.0 171 289 0.59
18 16 Sep 94 Straits of Taiwan 10.20 6.4 1 402 0.00
19 04 Oct 94 Kuril Islands, 22.23 8.2 11 422 0.02 1.5-3 m
N. Japan tsunami
20 09 Oct 94 E. Indonesia 6.44 6.9 1 50 0.02
21 14 Nov 94 Philippines 3.15 7.1 74 171 0.43 10-15 m
(Mindoro) tsunami
(1.5 m min.)
22 20 Nov 94 Indonesia - 5.7 0 39 -
(Irian Jaya)
23 17 Jan 95 Japan (Kobe) 5.46 7.2 6,308 43,177 0.15
24 19 Jan 95 N.E. Colombia 10.05 6.5 8 10 0.80
25 08 Feb 95 W. Colombia 13.40 6.5 41 260 0.15
(Pereira)

DISASTERS VOLUME 20 NUMBER 3

The Health Efects of Earthquakes in the Mid-2990s 235

TABLE 1
(Continued)

No. Date Location Local Magnitude Deaths lnjuries Deaths1 Remarks

time Injuries

26 23 Feb 95 Cyprus 23.00 5.8 2 12 0.16

27 23 Feb 95 Taiwan - 5.2 2 10 0.20 Landslides,
with fatalities
28 05 Mar 95 S.W. Colombia 8.23 5.1 a 11 0.72
29 01 Apr 95 Sea of Japan - 6.0 0 10 -

30 13 May 95 N. Greece - 6.6 0 20

31 14 May 95 Indonesia 19.33 6.8 6 26 0.23 Tsunami and
(East Timor) landslides,
with fatalities
32 20 May 95 Indonesia 5.29 5.7 0 53
(Sulawesi)
33 28 May 95 Russia (Sakhalin) 0.04 7.6 1,989 406 4.89
34 15 Jun 95 W. Greece 3.00 6.1 26 59 0.44
(Egion)
35 25 J u n 95 Taiwan 2.59 6.5 1 0 -
36 13 Jul 95 S. China 5.55 7.3 11 136 0.08
37 30 Jul 95 N. Chile 1.00 7.8 3 58 0.05
38 01 Oct 95 Turkey (Dinar) 17.57 6.1 101 348 0.29
39 03 Oct 95 S.E. Ecuador 20.51 6.9 2 5 0.40
40 07 Oct 95 Indonesia 1.10 7.0 84 1,868 0.04 Landslides
(Sumatra)
41 09 Oct 95 W. Mexico 9.37 7.6 66 100 0.66
(Manzanillo)
42 25 Oct 95 China 6.47 6.5 45 350 0.12
(Yunnan Province)
43 03 Nov 95 Bolivia - 5.2 0 6 -
44 22 Nov 95 Egypt, Israel, Jordan 8.15 6.2 10 69 0.14
45 05 Dec 95 S. Turkey 20.51 5.6 0 5 -
46 01 Jan 96 Indonesia 16.05 7.0 9 0
(Palu, Sulawesi)
47 09 Jan 96 N.E. Guatemala 16.25 4.6 0 9 - 2 m tsunami
48 03 Feb 96 China 19.14 7.0 322 16,925 0.02
(Yunnan Province)
49 17 Feb 96 Indonesia 14.59 7.0 108 423 0.26 2-4 m
(Irian Jaya) tsunami

Means G. totals 6.4 19,903 88,899 0.31'

*Does not include Sakhalin (Russia) earthquake of 28 May 1995 (deathslinjuries = 4.89)

DISASTERS VOLUME 20 NUMBER 3

236 David Alexander

(a)
Deaths: n = 39, Injuries: n = 41

70,000
u)
a, 60,000
2
'q
-550,000
40,000
r.
(0 30.000
u)
s, 20,000
& 10,000
(D

n
"
5.25 5.75 6.25 6.75 7.25 7.75 8.25
Magnitude cfass mid-point

(b)
Deaths: n = 37, Injuries, n = 34

70,000
Q, 60,000
u)

*g50,000
*Z

U 40,000
e
@ 30,000
u)
s(D 20,000
8 10,000
n
U
01.30 04.30 07.30 10.30 13.30 16.30 19.30 22.30
Time period (3-hour blocks)

FIGURE 2 (a) Deaths and injuries by magnitude; (b) Deaths and injuries by time of day

DISASTERS VOLUME 20 NUMBER 3

The Health Effects of Earthquakes in the Mid-1990s
data, such as the very high mortality to
morbidity ratio in the Sakhalin Russian
earthquake of May 1995, where victims
had little opportunity to survive the catas-
trophic collapse of buildings. Neverthe-
less, if Sakhalin is excluded, the mean
ratio is very close to 1:3, which implies
either that the hypothesis has some resi-
dual validity or that it is a very enduring
spurious correlation. Moreover, the mean
ratio is 1:3.22 for earthquakes in the
magnitude range 6.0-7.4, which tends to
confirm my earlier findings (Alexander,
1985). In synthesis, despite the allure of
the death to injury ratio, it could not be
correlated more precisely with magnitude.
Nor, for the earthquakes studied, did it
correlate with time of day or number of
buildings damaged.
THE NATURE, CAUSES AND SETTINGS
OF INJURY
Not all reports gave details of the nature or
severity of injuries sustained, and hence a
full injury epidemiology could not be
constructed. However, the nature of injur-
ies was amply referred to, which enabled a
fairly comprehensive picture of injury
types to be assembled. Physical trauma
included head and back injuries, leg frac-
tures, broken ribs, multiple fractures of
limbs, clavicle fractures, spinal damage,
paraplegia, cuts, bruises, lacerations,
burns and crush injuries. Alexander (1993,
p. 470) described similar injury types in
other earthquakes but noted that the
proportions varied considerably from case
to case. In the Latur (India) earthquake of
1993, upper limb injuries accounted for
nearly one-quarter of the very small
number of people who were hospitalised
(IFRCRCS, 1994, p. 120). Crush injuries
were widely reported in earthquakes in
the Philippines, Japan and Russia. In
addition, at least one case during the
Sakhalin disaster of May 1995 involved
crush effects that necessitated amputation
237
before the victim could be removed from
the rubble. Severe and fatal burn injuries
were also widely reported, especially as a
result of events in India, Indonesia, Japan
and Russia: reports were about equally
distributed between the effects of domestic
fires (burns while cooking, scalding with
boiling water and so on) and more general
urban post-earthquake fire. The latter was
a particularly important cause of death and
injury in the Kobe earthquake of January
1995, which gave rise to 531 urban fires
(AGU, 1995).
After the Kobe earthquake, full statis-
tics on injury were published by the
Medical Examiner’s Office of Hyogo Pre-
fecture. About 71 per cent of 3,649 victims
autopsied died within 14 minutes of the
earthquake, and a further 10.7 per cent
may have died within six hours of the
tremors. Some 54 per cent died as a result
of crush injuries or other physical trauma
sustained in the collapse of buildings,
while 15.3 per cent suffered severe burns
(Anon., 1995). More comprehensive statis-
tics published a year later reduced this
figure to 10 per cent (Disasfer Research,
1996).
In the earthquakes studied, death also
resulted from dehydration, choking and
suffocation. The last of these was a very
significant cause of mortality in the Kobe
earthquake (ibid.). In the Pereira (Colom-
bia) earthquake of February 1995, dust
generated by building collapse led to the
death of some trapped victims, while
others suffered dehydration before they
could be rescued from under the rubble of
their homes. In nine other events, and in a
wide variety of settings, a small number of
deaths occurred as a result of acute myo-
cardial infarction (heart attack); forexam-
ple, five such fatalities were recorded in
the January 1994 Northridge, California,
earthquake. Victims were predominantly
over the age of 50 (cf. Katsouyani et al.,
1986). Three sets of reports gave details of
suicides that were apparently linked with
DISASTERS VOLUME 20 NUMBER 3
238
the catastrophes. Self-poisoning led to
fatalities in the Latur (Indian) earthquake
of September 1993 and the Liwa (Indone-
sian) disaster of February 1994, while
hanging appears to have been the pre-
ferred method after the Kobe earthquake
of January 1995. In each case there seem to
have been few victims (it was stated in the
Internet bulletin, Disaster Research no. 187,
31 January 1996, that there were at least
eight in Kobe, though other cases may
have gone unrecorded). My own studies
from other settings suggest that mortality
rates associated with suicides are practi-
cally unaffected by earthquakes (see Alex-
ander, 1982a). Nevertheless, the sense of
crisis engendered by earthquakes can
undoubtedly lead people to take their own
lives. This may be related to the so-called
‘survivor syndrome’, a state of helpless-
ness and despair that was amply docu-
mented after the 1973 Buffalo Creek flood
disaster (Tichener and Kapp, 1976). How-
ever, very little is known about the timing
and motivation of suicide after
earthquake.
In informed circles, post-earthquake
infection and disease are seldom regarded
as major problems, even though they do
occasionally break out (Blake, 1989).
Among the events studied here, illness
appeared to have been a significant
problem only in events with very large
numbers of survivors who were then
constrained to live in precarious con-
ditions. Survivors often had to wait for
adequate health care and shelter to be
improvised for them. Thus, after the Latur
(India) earthquake of September 1993,
scabies, skin diseases, diarrhoea and
dysentery were noted. An outbreak of
malaria was recorded following the south-
ern Chinese catastrophe of July 1995;
while after the Kobe (Japan) disaster of
January 1995, influenza and incipient mal-
nutrition were diagnosed. Rather than lack
of food, the latter probably resulted from
depression, a psychological state that was
DISASTERS VOLUME 20 NUMBER 3
David Alexander
widely identified among survivors in
Japan and also after various Indonesian
earthquakes. Shock and the ‘disaster syn-
drome’2 (Wallace, 1956) were described
after many of the events considered in the
present study; for example, the Liwa
(Indonesia) and Kobe (Japan)earthquakes.
As noted above, injuries, some of
them fatal, were caused by landslides in
Indonesia and Taiwan, debris flows in
Colombia, and tsunamis in various
locations. However, in most cases these
accounted for only a minority of the
victims and building collapse was as usual
the principal cause of death and injury. In
many instances, the victims tended to be
concentrated in relatively few buildings.
For example, in the Sakhalin (Russia)
earthquake of May 1995, the majority of
the nearly 2,400 casualties came from only
17 five-storey housing blocks, each of
which contained 80 apartments and all of
which collapsed during the tremors. Many
victims also came from apartment blocks
in the Northridge (California) earthquake
of January 1994 (at least 16 out of 60
deaths) and the Turkish earthquake of
October 1995. Instead, they were extracted
from hotels in earthquakes in Egion
(Greece) in June 1995 and Manzanillo
(Mexico) in October 1995. Schools col-
lapsed on pupils in two Chinese earth-
quakes, and bridges were a significant
source of mortality in two Japanese
disasters and in the Northridge, Califor-
nia, earthquake. In contrast, casualties
mainly occurred in vernacular housing
during the Latur (India) earthquake of
1993, the Liwa (Indonesia) tremors of 1994,
the Pereira (Colombia) disaster of 1995,
and the Yunnan Province (China) catas-
trophe of February 1996.
As building collapse is generally held
responsible for most casualties in earth-
quakes (Page et al., 1975), it is reasonable
to suppose that there will be a relationship
between the number of buildings
damaged and the number of victims
The Health Effects of Earthquakes in the Mid-1990s 239

TABLE 2
Casualties in relation to building damage for selected earthquakes, 1994-1996

Date Locution Magnitude Deaths Injuries Buildings Deaths1 Deaths +

Bfdgs InjurieslBfdgs
Destroyed Severely Slightly destroyed damaged b
damaged damaged destroyed

17 Jan 94 California 6.6 60 9,202 8,938 12,493 23,887 0.01 0.20

16 Feb 94 Indonesia 6.5 207 2,389 2,066 3,483 4,549 0.1 0.26
01 Mar 94 Iran 5.6 3 31 3,007 0.01
02 Jun 94 Indonesia 7.2 222 440 1,355 102 15 0.164 0.45
06 Jun 94 Colombia 6.4 271 158 1,664 3,160 0.163 0.09
04 Oct 94 Russia 8.2 11 242 111 248 0.1 0.7
14 Nov 94 Philippines 7.1 74 171 797 3,288 0.093 0.06
17 Jan 95 Japan 7.2 6,308 43,177 83,767 131,233 4,700 0.075 0.23
08 Feb 95 Colombia 6.5 41 260 250 750 0.164 0.3
20 May 95 Indonesia 5.7 0 53 24 125 407 - 0.1
01 Oct 95 Turkey 6.1 101 348 1,000 3,500 0.101 0.1
03 Oct 95 Ecuador 6.9 2 5 83 0.08
07 Oct 95 Indonesia 7.0 84 1,868 7,137 10,533 0.11
01 Jan 96 Indonesia 7.0 9 0 211 - 0.04
03 Feb 96 China 7.0 322 16,925 358,174 490,000 164,000 0.001 0.02
17 Feb 96 Indonesia 7.0 108 423 1,018 3,486 845 0.106 0.1

(Table 2). Hence, in a field study in Italy,
De Bruycker et al. (1985) found that being
trapped under rubble increases a victim’s
chances of being injured fivefold. Data on
damage were reported for 16 of the earth-
quakes studied, but little pattern is evident
in the ratio of total casualties to the total
number of buildings damaged (last col-
umn, Table 2, mean 0.178, standard devi-
ation 0.182). The number of casualties
varied between one and 45 per 100 build-
ings damaged, with a mean of 18 per 100.
In contrast, there is slightly more regular-
ity in the average ratio of deaths to
buildings destroyed, which suggests a
fatality rate of 10-16 people per 100 build-
ings suffering total or partial collapse.
Apart from the causes of mortality and
morbidity, one other question concerns
the classification of non-fatal injury.
Official statistics rarely differentiate
between injuries sustained according to
their seriousness and, if they do, it is
highly unusual to find any explanation of
-
criteria used in the compilation of statis-
tics. However, 10 of the disasters studied
did end with the dissemination of a crude
classification of injury severity into serious
(requiring hospitalisation or otherwise
requiring protracted medical attention)
and slight (needing only limited outpatient
status). Table 3 shows that for these
events the proportions varied widely from
84 per cent serious to 99 per cent slight
(though, of course, the number of casual-
ties varied by about 2 orders of magni-
tude). With regard to the latter figure, the
tiny proportion of people who remained in
hospital after the Latur earthquake of 1993
in India (only 125 out of 10,500 injured, see
IFRCRCS, 1994, p. 120) recalls the 1968
Khorasan tremors in Iran, in which only
3.3 per cent of 11,000 injured people
required prolonged inpatient care (Rennie,
1970).
In sum, no pattern is evident in the
small number of available data, which is
perhaps hardly surprising given the wide
DISASTERS VOLUME 20 NUMBER 3
240 David Alexander

variety of physical settings in which injur- cribed in 13 different earthquakes that

ies occurred. This lack of regularity is
reinforced by the high variability of serious
in relation to slight injuries reported in
other earthquake disasters, where the
ratios varied from 1:9 to 1:30 (Alexander,
1993, p. 470). It is probable that no simple
pattern underlies the statistics, not least
because the proportions must depend in
part on the speed and efficacy with which
urban heavy rescue is conducted and lives
are either saved or lost (Olson and Olson,
1987; Noji, 1991).
BEHAVIOURAL FACTORS
It was reported in 23 of the events studied
that people ran out of doors during the
tremors. This common and widespread
reaction to earthquakes has profound
implications for safety: in the Colombian
disaster of 1995 the streets of Pereira filled
with collapsed masonry, shattered
window glass and fallen roof tiles. The
most extreme stimulus to flight, panic, is
regarded as a phenomenon whose occur-
rence has been exaggerated, whose defini-
tion is controversial, and whose identifica-
tion presents numerous problems3
(Johnson, 1987; Alexander, 1995).
Nevertheless, panic was specifically des-
occurred during the study period in places
as diverse as Ecuador, Ethiopia, Italy,
Lebanon, Taiwan and the USA. Moreover,
competitive behaviour engendered by
panic-induced flight was reported to have
occurred in schools in Guangdong, China,
in September 1994, and Manzanillo, Mex-
ico, in October 1995, as well as in an
Egyptian youth hostel in November 1995.
The last of these cases recalls the 1992
Cairo earthquake (Alexander, 1995, p. 171;
Degg, 1993) in which, according to eye-
witness reports, up to a quarter of the 480
fatalities may have resulted from injuries
sustained during mass panic. With respect
to the period studied here, panic-induced
flight also led to casualties in various
earthquakes in China, Greece, Israel and
Mexico. The victims suffered injuries that
ranged from sprains and broken bones to
death by falling or being crushed.
When viewed as a temporal sequence,
official statistics give an indication of the
rise in casualties as search-and-rescue
teams extract the injured and account for
the dead. Victims were extracted alive
from the rubble up to six days after the
earthquakes at Latur (India) in September
1933 and Sakhalin (Russia) in May 1995,
but no later than this. In the latter case, 33

TABLE 3
Proportion of serious and slight injuries in selected earthquakes, 1993-1996

Date Locat ion Local time Magnitude Total injured Seriously Slightly
of day injured (TO) injured (%)

03 Feb 96 China 19.14 7.0 c. 16,925 3,925 (23) c. 13,000 (77)

30 Sep 93 India 3.58 6.5 c. 10,500 125 (1) c. 10,375 (99)
17 Jan 94 Catifornia 4.31 5.6 9,202 1,495 (16) 7,707 (84)
07 Oct 95 Indonesia 1.10 7.0 1,868 558 (30) 1,310 (70)
16 Feb 94 Indonesia 0.08 6.5 1,439 464 (32) 975 (68)
17 Feb 96 Indonesia 14.59 7.0 423 56 (13) 367 (87)
28 May 95 Russia 0.04 7.6 406* 341 (84) 65 (16)
01 Oct 95 Turkey 17.57 6.1 348 210 (60) 138 (40)
04 Oct 94 Japan 22.23 8.2 242 32 (13) 210 (87)
20 May 95 Indonesia 5.29 5.7 53 8 (15) 45 (85)

*30 per cent were crush injuries.

DISASTERS VOLUME 20 NUMBER 3

The Health Effects of Earthquakes in the Mid-2990s
people were rescued alive nearly 48 hours
after the event, but by and large the
earthquakes studied confirm the well-
known rule that most living victims are
rescued within the first 24 hours following
an earthquake (Noji, 1991). Rescue ope-
rations then largely concentrate on reco-
vering bodies and accounting for people
who have been reported missing. The
Kobe earthquake of January 1995 is a
typical case (Figure 3). Although it took 25
days to establish preliminary totals, and
final totals were eventually published a
full year later, most of the dead, missing
and injured were accounted for within a
week of the catastrophe and thereafter the
curves home in asymptotically upon their
final values. By contrast, the total of
houses damaged and destroyed rose
steadily for weeks after the disaster as
buildings and their sites were methodi-
cally surveyed. However, it should be
noted that over the year following the
disaster, official death tolls were increased
by 15 per cent and injury totals by 30 per
cent as the definitions of casualties were
gradually altered to include people who
had died of their injuries and those who
were treated some time after the
earthquake.
In the light of previous studies, Alex-
ander (1993, p. 18) observed that the
official total of casualties in a disaster rises
steadily for some days on the basis of
head- and body-counts, but officials who
are involved in search-and-rescue ope-
rations tend to overestimate the totals. The
‘overshoot’ of estimated figures then con-
verges on the official figures until a final
total is arrived at about 2-3 weeks after the
disaster. Data from the Sakhalin earth-
quake of May 1995, derived from both
official Russian government communiques
and interviews with rescue workers, con-
firm this model for non-fatal injuries but
suggest that the death tolls were not only
overestimated, but also underestimated.
The result is an ‘envelope’ of estimates
241
(Figure 4) that eventually converges upon
the official totals. The latter rose with a
regularity not found in other earthquake
disasters (for example, see Alexander
1982b, p. 82), which is testimony either to
the orderliness of the Russian authorities’
handling of the emergency, or to the
unvarying difficulty of recovering the
bodies. As it is common to find that the
numbers of casualties are over- or under-
estimated in the period immediately after
an earthquake, it would be helpful to
study this aspect in more disasters so as to
arrive at better early predictions of death
and injury tolls. Such estimates have a
considerable bearing on the number and
quality of medical and logistical resources
that are sent into disaster areas.
EVALUATION
According to the International Federation
of Red Cross and Red Crescent Societies,
mortality and morbidity in earthquake
disasters averaged 22,956 and 30,003, res-
pectively, over the period 1968-92
(IFRCRCS, 1994, pp. 148-50). Though the
figure for morbidity (35,560) is slightly
higher in this study, the rather lower
figure reported here for mortality (7,960)
results from the fact that the period stu-
died was free from major catastrophes
such as the magnitude 7.8 earthquake that
struck Tangshan, China, in 1976 and killed
242,469 people (Coburn and Spence, 1992,
p. 5), and the magnitude 7.3 disaster in
Iran in 1990 that killed 40,000. In such
events, the scale and severity of destruc-
tion were so great that mortality tended to
be high in relation to morbidity, hence the
discrepancy in death to injury ratios
between the Red Cross totals and those
reported in this study (1:1.31 and 1:3.22,
respectively). It should be noted also that
overall mortality in earthquakes is concen-
trated heavily in the largest events and in
particular places (Coburn and Spence,
1992). The Red Cross’s long-term data
DISASTERS VOLUME 20 NUMBER 3
242 David Alexander

6,000 I
5,000 -
Final totals published
Dead
one year later:
6,308dead,
2 missing

o " l ' l "

*:---? ...,-.-
* ....-.-, pL . 1

17 18 20 22 24 26 28 30 I 3 5 7 9
JanlFeb 1995

"17 18 20 22 24 26 28 30 1 3 5 7 9
JanlFeb 1995

Final total published

one year later:
219,500

Q 1 ~ ' ~ ' l ' i ' ~ ' ~ ' ' ' ~ ' ~ ' " ' ' ~
17 18 20 22 24 26 28 30 1 3 5 7 9
JanlFeb lSg5

FIGURE 3 Temporal pattern of accounting for deaths, missing persons, injuries and damage to housing in
the Kobe, Japan, earthquake of 17January 1995. Data were supplied by the lapanese Government and
disseminated widely on the Internet.

DISASTERS VOLUME 20 NUMBER 3

The Health Effects of Earthquakes in the Mid-3990s 243

(IFRCRCS, 1994) suggest that 35 per cent events each year from 1900 until 1992,
of earthquake disasters occur in Asian which suggests that perhaps only half the
countries, which makes the 63 per cent former total may have involved fatalities.
recorded during this study period (and the This is corroborated by Alexander (1985),
96 per cent of deaths) anomalously high, who studied earthquakes over 21 years
though perhaps less so when one takes and found that seven to 11 of them
into account the observation (Coburn et resulted in significant numbers of deaths.
al., 1989) that half of all recorded deaths in In terms of mortality per earthquake, the
earthquakes have occurred in China. Red Cross's annual average figure of
In synthesis, the period studied here 22,956 is higher than that quoted by Alan
may well be typical of intervals of time Feuerbacher in an Internet communication
during which earthquake casualties are dated 1994. He suggested a mean of 18,672
relatively limited. However, this assertion seismic deaths per annum for the first nine
is difficult to back up with quantitative decades of the present century; however,
information, in particular because histori- it appeared that this figure was based on
cal earthquake data tend to be under- an average of only 2.8 events each year,
reported and unreliable, and hence long- which implies considerable under-report-
term averages may be suspect. While the ing of earthquakes that involved relatively
Red Cross gave an average of 25.7 signifi- few deaths. Feuerbacher's data for the last
cant earthquakes per annum over the five centuries showed a drop of nearly
period 1968-92 (IFRCRCS, 1994), Coburn three-quarters in seismic mortality per
and Spence (1992) listed only 12.7 lethal million of the world population in the

'.. ....
, ...-....
: o -........ -..._ Envelope of estimates
-.. .-.._.....................
.......... / of number of dead
....
.K-... .
.-..........
_ _ _ _.-O...---o--
_- 0 ___. U". 0. .I -5, -- ...........
0 0 0
. . .. . .. .
_-....
_.__.-
-.
----.-.- . . .
.. ,-_._-..
0 -.......... -.,_.-.- 0 -
0. ~ ~

. -.4. -\
,._.-..
....__.-
'., /..'
,.."
_/-
Official figures
'. ,. -..'_,_.*'. - - b
-i. (no. of dead)

28 29 30 31 1 2 3 4 5 6 7 16
May June 1 995

FIGURE 4 Estimates and official figures for the number of deafhs and ifijuries in f h e Sakhalin, Russia,
earthquake of 28 M a y 1995

DISASTERS VOLUME 20 NUMBER 3

244
twentieth century, but as the number of
deaths remained stable over the years, this
implies that the fall is largely accounted for
by increasing numbers of people, not by
decreasing death tolls. As the present
century’s 20 most lethal earthquakes
involved a mean mortality of 52,900
(Coburn and Spence, 1992, and other
sources), all that can be said with certainty
about the mid-1990s is that no event then
led to loss of life on such a scale.
Discrepancies between long-term data
and the results of the present study call
into mind a series of criticisms that can
possibly be levelled against macroscopic
studies such as the present one. For
instance, it is well known that death and
injury during earthquakes depend on a
wide variety of factors (Coburn and
Spence, 1992), including physical variables
(hypocentral depth, duration of shaking,
seismic acceleration, geology of the local
terrain, distance from epicentre), architec-
tural variables (construction type, state of
maintenance of buildings, enactment and
enforcement of anti-seismic building
codes) and social variables (patterns of
human activity, level of popular aware-
ness of earthquakes, counter-disaster
training, efficiency of rescue operations),
none of which is explicitly considered
here. Given these complexities, one may
ask whether global surveys have any
validity at all. However, the presence of
regularities in the data analysed here
indicates that the exercise was indeed
worthwhile, and some broad generalisa-
tions can be made about earthquakes,
even if these tend to be more akin to
partially tested hypotheses than to phy-
sico-social laws.
Accordingly, this article ends with a
series of provisional conclusions that may
serve as hypotheses for further investi-
gation. First, earthquakes are highly repe-
titive events that tend to cluster in both
time and space. In Indonesia and Iran, for
instance, casualties occurred in seismic
DISASTERS VOLUME 20 NUMBER 3
David Alexander
disasters that were separated from one
another by only a few weeks, and in
Yunnan Province, China, by only a few
months. But despite the frequency of
events, only about three-fifths of the
damaging earthquakes that occur around
the world generate mortality or morbidity.
Not only do these vary widely in earth-
quake disasters, but so do the proportions
of serious to slight injuries. The pattern of
earthquake casualties in single events can
vary from a few people with minor injur-
ies, or one or two heart-attack victims, to
major disasters in which loss of life is
heavier than the toll of non-fatal injuries.
Leaving aside the question of very large
earthquake catastrophes (with death tolls
of at least 30,000), none of which occurred
during the study period, casualties may be
greatest in the early hours of the day and
in events of intermediate magnitude
(6.5-7.4).
In earthquakes that generate only
moderate numbers of casualties (no more
than a few score victims), mortality and
morbidity tend to be clustered in particular
locations, such as individual schools,
hotels, apartment complexes or groups of
vernacular houses. Despite the publicity
given to landslides and tsunamis, these
tend to cause far fewer injuries than does
building collapse. Moreover, the data ana-
lysed here suggest that seismic effects
upon transport (bridge collapse and vehi-
cular crashes) may be much less important
sources of casualty than the crushing and
trapping of victims under collapsed
buildings.
Physical trauma tends to receive more
attention than psychological impairment,
though it appears that Wallace’s ‘disaster
syndrome’ (Wallace, 1956 - and see
above) could well be a universal result of
seismic disaster, at least among a certain
proportion of survivors. Internationally,
the phenomenon is not as well studied as
panic, though it may be at least as wides-
pread. The present data suggest that a
The Health Effects of Earthquakes in the Mid-1990s
significant proportion of earthquake
disasters, perhaps at least one-quarter,
will result in panic, flight or a combination
of these. In this context, one of the most
common reactions is running out of doors:
events that occurred during the study
period indicated that it can be a highly
dangerous, and occasionally fatal, reaction
(but, on the other hand, virtually nothing
is known about the ability of flight to save
lives when buildings start to collapse upon
their occupants). Finally, the present
study confirms that it usually takes at least
two weeks, and sometimes much longer,
to establish the full nature of casualties
caused by a major earthquake disaster,
and during this period the figures are
subject to predictable forms of
mis-estimation.
These observations represent an
attempt to generalise worldwide tenden-
cies in seismic disasters over a limited
interval of time. The period covered falls in
the middle of the UN’s International
Decade for Natural Disaster Reduction.
The Decade has stimulated considerable
research on the human consequences of
earthquakes, but much more needs to be
done before epidemiology can be used
reliably to predict casualty patterns in
future earthquakes.
Acknowledgements
This paper was written as a briefing document
for the US Second National Workshop on
Modelling Earthquake Casualties for Planning
and Response, which took place at Los Altos,
California, during February 1996. For their help
and support I thank the organiser of the
meeting, Dr Robert Olson, and our sponsors,
the US National Science Foundation, the
Federal Emergency Management Agency, the
National Institution of Standards and Techno-
logy and the California Emergency Medical
Services Authority.
Notes
1. Allowing for the fact that reporting rates
245
have risen by about one-third per century,
the long-term average recurrence interval of
earthquakes that cause 6,000 or more deaths
is approximately 31 months (Canse and
Nelson, 1981; and elsewhere). This means
that, in purely statistical terms, a period
such as the one studied here should contain
one event with a death toll at least as high as
that found in Kobe.
2 . Wallace’s ’disaster syndrome‘ is a psycholo-
gical defensive reaction pattern involving
the spontaneous withdrawal of a patient’s
contact with his or her external environment
and three stages of its gradual re-establish-
ment (Alexander, 1993, pp. 565-6). In
general terms, ‘shock’ is a state of profound
psychological impairment leading to apathy,
introversion or, conversely, to agitated
extrovert beh aviour .
3. According to some definitions, panic is an
instinctive, asocial and non-rational reaction
to a tangible, immediate threat. It leads an
individual to undergo spontaneous with-
drawal from social relationships and to focus
intensely on what will happen next (ibid.,
pp. 559-60).
References
AGU (1995) Kobe Earthquake: An Urban
Disaster. EOS: Transactions of the American
Geophysical Union 76(6), 49-51.
Alexander, D.E. (1982a) Disease Epidemiology
and Earthquake Disaster: The Example of
Southern Italy after the November 23, 1980,
Earthquake. Social Science and Medicine
16(22), 1,959-69.
Alexander, D.E. (1982b) The Earthquake of 23
November 1980 in Campania and Basilicata,
Southern Italy. International Disaster Insti-
tute, London.
Alexander, D.E. (1985) Death and Injury in
Earthquakes. Disasters 9(1), 57-60.
Alexander, D.E. (1993) Natural Disasters. UCL
Press, London, and Chapman & Hall, New
York.
Alexander, D.E. (1995) Panic During Earth-
quakes and Its Urban and Cultural Contexts.
Built Environment 21(2-3), 171-82.
Anon. (1995) Most Kobe Victims Died Quickly.
Associated Press Report, Los Angeles,
15 March 1995.
DISASTERS VOLUME 20 NUMBER 3
246
Armenian, H.K., E.K. Noji and A.P. Oganes-
sian (1992) Case Control Study of Injuries
Due to the Earthquake in Soviet Armenia.
Bulletin of the World Health Organization 70,
251-57.
Blake, P.A. (1989) Communicable Disease
Control. In M.B. Gregg (ed.) The Public
Health Consequences of Disasters. Federal
Centers for Disease Control, Atlanta, 5-20.
Coburn, A.W., A. Pomatis and S. Sakai (1989)
Assessing Strategies to Reduce Fatalities in
Earthquakes. lnternational Workshop on Earth-
quake l nj uy Epidemiology for Mitigation and
Response. Johns Hopkins University Press,
Baltimore, 107-32.
Coburn, A. and R. Spence (1992) Earthquake
Protection. Wiley, New York.
De Bruycker, M., D. Greco and M.F. Lechat
(1985) The 1980 Earthquake in Southern
Italy: Mortality and Morbidity. lnternational
Journat of Epidemiology 14, 113-17.
Degg, M. (1993) The 1992 Cairo Earthquake:
Cause, Effect and Response. Disasters 17(3),
226-38.
Disaster Research, 31 January 1996. Internet
Bulletin of the Natural Hazards Research
and Applications Information Center,
University of Colorado, Boulder.
Durkin, M.E. (1987) Casualties, Search and
Rescue, and Response of the Health Care
System. (The San Salvador earthquake of
October 10, 1986.) Earthquake Spectra 3(3),
621-34.
Durkin, M.E., C.C. Thiel Jr, J.E. Schneider and
T. De Vriend (1991) Injuries and Emergency
Medical Response in the Loma Prieta Earth-
quake. Bulletin of the Seismological Society of
America 81(5), 2,143-66.
Ganse, R.A. and J.B. Nelson (1981) Catalog of
Significant Earthquakes, 2000 B.C.-1979.
National Geophysical Data Center, National
Oceanographic and Atmospheric Administ-
ration, Boulder.
Glass, R.I., J. J. Urrutia, S. Sibony, H. Smith, B.
Garcia and L. Rizzo (1977) Earthquake Injur-
ies Related to Housing in a Guatemalan
Village. Science 197, 638-43.
IFRCRCS (1994) Wortd Disasters Report 1994.
International Federation of Red Cross and
Red Crescent Societies, Geneva, and
Kluwer, Dordrecht.
Johnson, N.R. (1987) Panic and the Breakdown
DISASTERS VOLUME 20 NUMBER 3
David Alexander
of Social Order: Popular Myth, Social
Theory, Empirical Evidence. Sociological
FOCUS 29(3), 171-83.
Jones, N.P., E X . Noji, F. Krimgold and G.S.
Smith (1990) Considerations in the Epide-
miology of Earthquake Injuries. Earthquake
Spectra 6, 507-28.
Katsouyani, K., M. Kogevinas and D. Tricho-
poulos (1986) Earthquake-related Stress and
Cardiac Mortality. International Journal of
Epidemiology 15, 326-30.
Noji, E.K. (1989) Medical and healthcare
aspects of the 1988 earthquake in Soviet
Armenia. In L.A. Wyllie, Jr and J.R. Filson
(eds.) Armenia Earthquake Reconnaissance
Report. Earthquake Spectra 5 (supplement),
101-7.
Noji, E.K. (1991) The Medical Consequences of
Earthquakes: Co-ordinating the Medical and
Rescue Response. Disaster Management 4(1),
3-11.
Olson, R.S. and R.A. Olson (1987) Urban
Heavy Rescue. Earthquake Spectra 3(4),
645-58.
Page, R.A ., J.A. Blume and W .B. Joyner (1975)
Earthquake Shaking and Damage to Build-
ings. Science 189, 60148.
PAHO (1981) Emergency Health Management
After Natural Disasters. Scientific Publication
407, Pan American Health Organization,
Washington.
Rennie, D. (1970) After the Earthquake. The
Lancet 2, 704.
Roces, M.C., M.E. White, M.M. Dayrit and
M.E. Durkin (1992) Risk Factors for Injuries
Due to the 1990 Earthquake in Luzon,
Philippines. Bulletin of the World Health Orga-
nization 70, 509-14.
Sapir, D.G. (1993) Health Effects of Earth-
quakes and Volcanoes: Epidemiological and
Policy Issues. Disasters 17(3), 255-62.
Titchener, J.L. and F.T. Kapp (1976) Family and
Character Change at Buffalo Creek. American
Journal Of PsychiatT 133(3),295-9.
Wallace, A.F.C. (1956) Human Behaviour
During Extreme Situations. Disaster Study
No. 1, National Academy of Sciences,
Washington.
Whittaker, R., D. Fareed, P. Green, P. Barry
and A. Borge (1974) Fletes-Barrios Earth-
quake Disaster in Nicaragua: Reflections on
the Initial Management of Massive Casual-
The Health Effects of Earthquakes in the Mid-1990s 247

DISASTERS VOLUME 20 NUMBER 3