Cover Image
close this bookCorporal Damage as Related to Building Structure and Design: The Need for an International Survey (Centre for Research on the Epidemiology of Disasters, 1989, 16 p.)
View the document(introduction...)
View the document2. MITIGATION VS RESPONSE
View the document5. RESEARCH
View the document6. IDNDR
View the documentBIBLIOGRAPHY


Michel F. LECHAT, M.D., Dr P.H.


Earthquakes have caused some one million deaths over the last two decades. Such an event as the Tangshan earthquake in Northern China, 1976, has been responsible for several hundred of thousands in a matter of minutes. The growing urbanization of the planet, including in quake-prone areas, with as a result vulnerable megapolis of up to 20 millions, forewarns of disasters to come. Engineers on their side design anti-seismic structures - the only conceivable mitigation measure at present - at a cost of billions. This “International Workshop on Earthquake Injury Epidemiology” intends to tackle the problem of mitigation or and response to earthquake, bringing together widely diversed expertise, from the structural engineer conceiving safe designs to the epidemiologist scoring the end-results.


For the purpose of preparedness and management, disasters in general, including earthquakes, may be considered as a time-sequence in five phases:

* the anticipatory phase
* the pre-disaster phase
* the impact phase
* the relief phase
* the rehabilitation phase

The anticipatory phase is the time for planning and for such preparedness activities as hazard mapping, preventive technology, disaster training and public education. The pre-disaster phase commences when there are indications of an impending disaster and a prediction based on a probability judgment can be made. It involves the activation of preset warning systems identified during the non-disaster phase which are appropriate for key groups in disaster response and for the general public. The impact phase involves a major disruption of the local community. Though it is characterised by initial chaos, it often includes a remarkable degree of self organisation whose effectiveness can be increased by suitable programmes during the anticipatory phase. The relief phase begins when assistance arrives from outside the disaster area. The effectiveness of such aid may be increased by preparedness and planning. In the rehabilitation phase, short-term relief is progressively replaced by more specific assistance based on an assessment of local conditions. It ends with the restoration of, or an improvement on the conditions prevailing before the disaster. This framework is common to all natural disasters, and to some extent can also be applied to man-made disasters. Intervention can thus be envisaged at various points of entry in the system.

In recent years, the emphasis in disaster management has definitely shifted from post-facto improvisation to pre-disaster planning and preparedness, including prevention and prediction whenever possible. Floods for example are amenable to prevention by appropriate land use, soil conservation, construction of dykes. While hurricanes, typhoons and other strong wind phenomena cannot at the moment be prevented, prediction is possible. Provided the technology is applied and the population is educated to response to warning signals, evacuation or other preparedness measures can be implemented.

It may however happen, and it often happens, that even when the technology for prevention and/or prediction is available, it cannot be applied. Cost and social factors have to be taken into account. It might be too expensive to build dykes against tsunamis. There are also very good reasons for people living in flood-prone areas or on the slopes of volcanoes. Some countries can support a large population because they have floods which irrigate the lands; in countries with an accelerating population growth volcanic soil is good for agriculture, and anyway people may have no other places to go. Then, often, rescue and relief, that is response once the disaster has occurred, is still essential to save what remains to be saved.

Regarding earthquakes, prediction cannot be operational at present. Earthquakes prone areas are well mapped, but neither the time of occurrence nor the exact place can be predicted with sufficient accuracy. Long-term prediction is neither sensitive nor specific. Would less unreliable prediction become possible, it would raise excruciating dilemmas for the decision maker, apart of the perspective of horrendous lawsuits in case of false positive or false negative warnings. Major earthquakes may also occur in low-risk areas where for centuries no disaster has occurred, the more damageable so that the population is not prepared for it. The Eastern United States (Charleston, South Carolina, 1886) is such an area. Short-term prediction is the subject of much talk. Some success has been claimed in China. It is however doubtful that a prediction with a lag time of a few minutes can reach the population in time to prompt effective escaping behaviors, especially in high buildings and in large cities. It also requires an efficient system of communication. Not every country, especially in rural areas, can afford warning system based on computers linked to the national telephone network, like Iceland, or radios and televisions systems delivering emergency messages even when the units are turned off, like Japan.

It means that with respect to earthquakes, we are left with two points of entry into the system, i.e. mitigation of effects, through anti-seismic buildings, and efficient rescue after the impact, that is response and rescue.


Anti-seismic building aims at modifying the interaction between the habitat and its occupants. It is reported to be most effective at least in saving lives and preventing casualties, provided the building codes are actually adhered to. It is reported that in Japan, the number of deaths due to earthquakes has steadily decreased since the ediction of the Disaster Countermeasures Basic Act in 1966. Time series however are generally not matched for the number and magnitude of earthquakes.

Anti-seismic building techniques have made tremendous progress over the last few decades. Buildings of 50 or more stories sheltering thousands of people are build in highly hazardous areas. They are reputed safe in case of earthquakes of high magnitude. Anti-seismic design have been tested through computer simulation, on mechanical models, and to some extent, in real life situations. The experience of Mexico City, 1985, has clearly demonstrated that such buildings, when the coding codes are applied, can resist an earthquake of magnitude 8.4.

That is a remarkable achievement. One may however wonder whether the primary purpose of these engineering developments is to protect people, or rather to overcome the challenge of pursuing the necessary modernization of very large cities which happen to be constructed in hazardous areas. Thus, whereas there is no doubt that anti-seismic design confers protection on an individual building basis, what the protection is in terms of population-coverage remains questionable.

Three types of situations should be considered:

(1) The enforcement of anti-seismic building code should be enforced for all new structures, especially public facilities, offices spaces and housing, in earthquake prone areas. Can it be afforded everywhere, especially in poor countries, while it is known to increase the building cost by 10 or 15 percent ? Often, antiseismic building measures are enforced for the reconstruction of a town which has suffered a recent earthquake or has been repeteadly destroyed in the past. There is no guarantee however that the next earthquake will occur at the same place (although it may be the case as in Orlsville/El Asnam, Algeria, 1954 and 1980). One would like to know, world-wide, and by country or disaster-prone areas, (a) the proportion of people who are effectively protected at the moment by anti-seismic buildings, both at home and in the working place; (b) what would be the cost to afford complete protection. One could of course conceive of a scheme which should make foreign loans or grants conditional to the implementation of anti-seismic measures, on the model of what is now recommended for environmental protection. This conditionality is however pure anathema for a number of developing countries. In conclusion, in many places, in the next future, anti-seismic building will be reserved to a limited number of major buildings, the more the better.

(2) In most earthquake exposed cities, structures, modern or ancient, built with no antiseismic techniques whatever, represent the majority. Antiseismic fitting of old buildings raise additional issues. Beyond the cost, the mere number of buildings to be renovated may require setting priority, which in turn could lead to delicate ethical and political choices.

What about the existing old houses? Should they be renovated, destroyed, replaced? The question is particularly relevant for the Mediterranean countries. Cost of renovation for making them anti-seismic would be considerable. It could be carried out on an individual basis, and probably not on a large scale.

The most extreme irremediable situation consists of those buildings which are supposed to be antiseismic while actually the antiseismic codes were not adhered to.

(3) The earthquake resilience of traditional housing varies according to the style and shows large geographical variations. In most developing countries, traditional housing is gradually replaced by new styles. The study conducted by R. Glass in Santa Maria Cauque, Guatemala, following the 1976 earthquake, provides a well-documented example of the relationship between deaths and casualties and the type of housing (1). In a 1918 earthquake, at a time when all houses were built of cornstalk and mud-covered, no death occurred. The first adobe houses, with dried bricks held by weak mud mortar, were built around 1925. In 1971, 85 percent of the houses were built in adobe. In the 1976 earthquake, 5 per cent of the village's population were killed. Clearly, in 50 years, the risk of quake-related trauma has gone from minimal to maximum.

The unregulated wild mix of traditional building methods and new materials is a characteristic of economic development in poor countries. In Turkey, people are extending their houses by building additional floors, placing a large concrete slab on insufficiently supported walls. Such procedure is calling for disaster at the first tremor.

The problem is an usual one not dissimilar to the Green Revolution in which irrigation without proper sanitary engineering brings water borne diseases, and insecticide poisoning is the price paid to protect the crops. Introduction of new technologies must go together with corresponding education as a package. For earthquake protection, it calls for research on the housing factors as related to traumas, in view of edicting simple minimal building codes at no or minimal cost, and education at the community level.

One conclusion is that, whatever the great success achieved in anti-seismic design, in many areas, emergency and rescue will most likely remain a prime component of disaster management.


At the moment, rescue is most often unplanned. The population is generally ill-prepared, yet ready to help. It is only recently that the potential of the local community for self help has been fully realized by health managers, with its implication for preparedness of the population and training of personnel. The concept is not yet fully accepted. Too often too large a reliance is still placed on external help.

Three sets of observation points to the central role that the disaster stricken community can play in search and rescue.

(1) Social scientists have been much more eager than health managers in studying the reaction of disaster-struck populations. Observations suggest that the response of the local community is quite effective, if not always efficient. In Japan, it was observed that within half an hour after an earthquake (Niigata, 1964), 75 percent of the non-affected survivors were engaged in some kind of rescue activity. In another more recent earthquake, Southern Italy, 1980, the proportion of survivors who remembered participating to rescue was lower, around 20 percent (3,4).

In earthquake prone areas, this capacity for search and rescue should be improved by adequate preparedness and training.

(2) Studies on the survival of trapped victims according to the delay in rescue show that immediate rescue is what counts. It has led to the concept of the “Golden twenty-four hours”. In the Tangshan, China, 1976 earthquake, the proportion of trapped people who survived declined from 99 percent within half and hour to 81 percent for those extricated later but before 24 hours after the impact, and 37 percent for those extricated on the second day (Fig. 1). Similar figures were observed in the Campania-Irpinia earthquake, Italy, 1980 where a retrospective survey of some 3.600 survivors in the 7 worst affected villages (overall casualty rate 19.7 percent) has shown that the proportion of people extricated alive among all extricated was respectively 88 percent for the first day, 35 percent for the second day, and 9 percent on the two following days. Of all the trapped victims extricated alive, 94 percent were rescued within the first 24 hours (3,4).

For all trapped victims which were extricated, dead or alive, 471 in total, 25 percent were extricated within half an hour, 44 percent within 3 hours, and 56 percent within 12 hours (Fig. 2). Within two days, 80 percent of all the trapped people were extricated. The time within which 50 percent of the victims had been extricated, dead or alive, (Time Lib50), was 8 hours.

Such studies however have severe limitations. The observations nay be partly due to a selection bias, since the less affected victims screaming for help or easily located are likely to be extricated first.

In addition, in Italy, 95 percent of the trapped survivors later evacuated to a medical center, were extricated with local means such as shovel, spikes, or bare hands. While it does not mean that more people could not have been saved would adequate equipment be available, this observation shows at least that the local population can play a determinant role in rescue. (10)

This was confirmed by comparing death rates, ratios of injured to death, and delays for extrication in single and multiple households. People living in single households had a death rate 2,4 times higher than those living in households with one or more members of the household present. When trapped, their death rate was approximately 1.5 times higher. Injured to death ratio was 2.6 for single households victims vs 1.8 people in multiple households. Proportion of people trapped who were extricated, dead or alive, within 24 hours was 46 percent only in simple households as compared to 61 percent in multiple households (Table 1).

Whatever the limitations of these studies, they indicate that the local community is quite effective in search and rescue, which as an essential component of disaster preparedness should be part and parcel of primary health care.

(3) External rescue teams, with an arrival delay on the spot of 2 to 4 days if not more, come generally too late to have a significant impact on the saving of life. While they may achieve isolated brilliant results well publicized by the media, their impact population wise needs to be properly evaluated. It is likely to be minimal and should be properly evaluated. This is not to say that external aid is not needed, far from it, since one life saved is worth every effort. It just does mean that investments are likely to be more productive in terms of the total number of lives saved when the local population is properly educated in search and rescue. That technologically advanced equipment can achieve what the local population is not able to perform with the local tools concerns only a small fraction of the victims surviving long enough to benefit from it.


Research is therefore needed in two directions, i.e. (1) to identify the hazards of different types of housing structures as related to earthquakes, in view of providing planners, architects and structural engineers with a quantitative assessment of risks; (2) to determine the response of trapped victims.

These studies are complementary. The dual objective is to provide a basis for the design of safer structures, and to contribute to a better planning of emergency and rescue services. This approach does not have to be restricted to earthquake, and could be extended to other types of natural disasters where the interaction of houses and occupants play a major role, mainly atmospheric hazards such as wind storms, tornadoes, hurricanes and tropical cyclones.

Such studies have been carried out extensively for automobile and traffic accidents, aircraft crashes, and industrial accidents, but to no significant extent for habitation.

(1) Regarding the identification of hazards associated with building structure, some data are available from the literature.

An interesting observation is that previously, following earthquakes, the ratio of casualties requiring medical care to deaths was approximately (within a large range) 3 to 1. Now, as shown in Armenia, it could come close to 1 to 3 (5). It seems that earthquakes in the ecology of large industrialized or developing urban areas is becoming a killer, with few casualties and many deaths.

In Guatemala, 1976, the adobe houses especially when overcrowded were the most lethal. No relation was observed with the type or roof or the number of doors and windows. The worst houses were those with heavy adobe block held by weak mud mortar. The best were chose with non-adobe lightweight cornstalk walls supported on a wooden frame; if the walls collapse, the frame and the roofs remain intact. Houses older that 7 years had a higher risk, with 1.6 times more casualties, than newer ones, since age tends to dry out the bricks and make them brittle. The size of the adobe blocks had no relation with the severity of injuries (1).

In Jalapa, Guatemala, after the same earthquake in 1976, a survey of the survivors showed that 11.5 of the admitted in the hospital casualties had fractures of the clavicle, suggesting that they were hit by falling roofs, walls, or beams (6).

In China, a large number of pelvic fractures were reported after the Hsing Tai earthquake, totalizing 30 % of all cases admitted to hospital (7), while only 32 cases of such fractures were noted in Tangshan estimated at 240.000 dead.

High rates of crush syndrome in the rescued survivors are at times reported (16 % in Agadir, 5.5 % in Skoplje) (8,9), whereas in other occasions none is mentioned, what suggests that either rescue was quite fast or more probably came too late.

Such data are scarce, and most of the time they were collected in a haphazard manner or are just chance observation. No systematic survey of deaths and casualties associated to building structure have been carried out. There are of course good reasons for this lack of data, since at the site of an earthquake all efforts are directed to immediate rescue.

Two research programs should be envisaged, one on a macrolevel, the other more on a more detailed scale.

The first research should assess the probability of death and casualty according to the type of housing (style, type of material, number or floors, age).

The second research should aim at identifying individual building characteristics responsible for deaths and casualties (ornaments, material, design). It will require a careful evaluation of the body damages, and to the extent possible should include forensic determination of the immediate cause of death.

Surveys on body damage and structures will have to take into account other major variables, mainly behavior at the tine of impact, and rescue procedures. Such data will incidentally have a bearing for defining the appropriate protective behaviors that the population should be taught, as well as improving training for rescue and designing appropriate non counterproductive rescue procedures.

Studies should also be conducted on the hazards associated with anti-seismic structures. It may happen that people survive well in building damaged beyond any repair, and conversely. The correlation of damage to antiseismic structures and casualties should therefore be further investigated, in order to develop failure models compatible with prevention of human losses.

(2) With respect to search and rescue, the response of trapped victims surviving the impact determine their chance to be extricated alive in a given period of time. Since it is obviously impossible to conduct cohort survival studies from the time of impact, chances of survival can only be estimated indirectly, by comparing the proportion of victims extricated death or alive at different times after impact. Data should be collected on the absolute number extricated dead or alive by periods of time (hours or days), or as cumulative figures over time after impact. Different patterns may emerge, which will give indication on the effectiveness of rescue and on the efficiency of extrication procedures.

Such observations however do not provide indication on when the dead victims actually died, and how much time the ones alive would have survived further. Still included among the deaths will be those victims who were killed at impact, and for whom earlier rescue would have made no difference. Determination of the cause of death, including wherever possible autopsies in at least a subsample, could provide some indication on the size of this group.

Since the objective of such studies is to determine delay in rescue for effective emergency health care, it should take into account the long term survival of the victims which were finally extricated alive, making provision for those who are beyond chance of survival at time of rescue.


A study on “Earthquake Injury Epidemiology and Search and Rescue in Collapsed Buildings” is presently undertaken under the sponsorship of NCEER, the National Center of Earthquake Engineering Research.

IDNDR, the International Decade for Natural Disaster Reduction, with its large international interdisciplinary research programs, provides an unique opportunity to extend this study in different geographical contexts, especially in earthquake prone areas in developing countries (10).

It should be one of the outcome of this workshop to prepare a preliminary protocol and set up a task force for such a study.


One person of household present

Two or more persons or household present


number of people



death rate

11.2 %

5.1 %

(p < 0.05)

injury rate

20.7 %

13.6 %

(p < 0.05)

injured/death (ratio)



proportion of persons trapped

23.9 %

14.4 %

(p < 0.001)

death rate amongst trapped

45.6 %

33.6 %

(p < 0.05)

proportion of trapped extricated (dead or alive) within 24 hours

46.5 %

61.0 %

(p < 0.05)




(1) GLASS, R.I. et al. - “Earthquake Injuries Related to Housing in a Guatemalan Village”: Science, 197; 1977: 638-643.

(2) WEN-KUI-MAI (personal communication).

(3) DE BRUYCKER, M. et al. - “The 1980 Earthquake in Southern Italy, Rescue of Trapped Victims and Mortality”: Bull. WHO, 61; 1983: 1021-1025.

(4) DE BRUYCKER, M., GRECO, D., LECHAT, M.F. - “The 1980 Earthquake in Southern Italy - Morbidity and Mortality”: Int. J. of Epid., 14; 1985: 113-117.

(5) LINDLEY, D. - “US Team Returns with Insights into Armenian Earthquake”: Nature, 107; 1989: 337.

(6) de VILLE de GOYET, C. et al. - “Earthquake in Guatemala: Epidemiologic Evaluation of the Relief Effort”: Bull. PAHO, 10; 1976: 95-109.

(7) PAN, T.-T.; TSUNG, S.-L.; SUN, C.-Y. - “The Mechanism and Treatment of Pelvic Fractures Encountered in Earthquakes”: Chinese Med. J., 4; 1978: 166-170.

(8) OURADOU, J. et al. - “Etude sous l'Angle Chirurgical des Blessd'Agadir Re au Centre Hospitalier de Casablanca”: Maroc M, 429; 1961: 166-170.

(9) ANDSELSKI, A. and SIVIC, A. - “Le Sme de Skoplje du Point de Vue Mcal”: Rev. Int. Serv. Santrm., 45; 1972: 403-410.

(10) International Ad Hoc Group of Experts - “Implementing the International Decade for Natural Disasters Reduction”: Report prepared for the Secretary General of the United Nations, ECOSOC, Geneva; 1989.