Cover Image
close this bookDisaster Chronicles Number 3: Earthquake in Mexico, September 19 and 20, 1985 (PAHO-OPS, 1985, 76 p.)
View the document(introduction...)
View the documentPrologue
View the documentList of authors who presented papers for this publication
View the document1. Introduction
Open this folder and view contents2. Magnitude of the disaster
Open this folder and view contents3. Public health and preventive medicine
View the document4. International cooperation
View the document5. Perspectives
View the documentAnnex : The survival of people in collapsed buildings
View the documentAcknowledgements
View the documentReferences

Annex : The survival of people in collapsed buildings

Michael E. Durkin (1), Anne H. Coulson (2), Martha Hijar (3), Jess Kraus (4), and Hitomi Ohashi Murakami (5)

Introduction

Complete collapse is the most serious outcome of earthquakes for buildings and occupants. However, despite its obvious implications for life safety, collapse does not result in death for all building occupants. The survival of some occupants raises important research questions. Why did some occupants survive while others perished? Did such factors as location of occupants, actions of occupants, injury type and severity, nonstructural cements and building contents, nature and time of entrapment, and method of rescue play a role in survival - factors which we can learn from and train people about? Or were the survivors mere lucky beneficiaries of fortuitous circumstances? These questions have an obvious bearing on how we prepare for future earthquakes.

Following the 1985 Mexico earthquakes, we surveyed eighteen medical residents who survived the collapse of the Medical Residents Dormitory at General Hospital in Mexico City; in addition we surveyed six other residents who lived in the dormitory, but were in other structures at the time of the tremor: At the earthquake's onset, four residents were in the neighboring Obstetrics and Gynecology Hospital building (that also collapsed) and two were in a basement corridor connecting the dormitory with a cafeteria (Table 27). The dormitory and hospital buildings became a major focus of search and rescue efforts lasting for several weeks.

The following paper presents results of these two case studies. Subsequently, other collapsed buildings in Mexico City and San Salvador were studied and these findings are presented in the discussion section. While these buildings cannot help us generalize about collapse patterns, they can help us explore a situation that has received little systematic attention to date. So far we have derived preliminary results on the role of location and occupant actions, the severity of injury, nonstructural elements and building contents, and method and timing of rescue. In order to derive these results, we used a cohort approach.

Table 27. Location of Interviewed Survivors (N=24)

Location

Number of Survivors

Dormitory Building

18

Obgyn Hospital Building

4

Underground Corridor

2

TOTAL

24

Background

The Buildings

The Medical Residents Dormitory was an eight-story, reinforced concrete structure. The building had a rectangular configuration. Seven of the eight floors were allocated to domiciliary space. Each floor in the repeated floor plan had eight bedrooms, four on either side of a double loaded corridor. An elevator, fire stair, and multi-purpose room were located in the center of each floor.

Four of the residents we interviewed were, at the time of the tremor, located in the Obstetrics and Gynecological Hospital building, a six-story reinforced concrete structure that also collapsed. Two others were in a basement corridor connecting the dormitory to a cafeteria.

The Occupants

On the morning of September 19th, 1985, this dormitory was home to seventy-six medical residents. Of this number, forty residents ultimately survived, and thirty-six perished from earthquake-related injuries (44)(45). The residents were in different years of the four years' residency training. They ranged in age from 27 to 32. Also, at least ten other students were living in the dormitory illegally when the temblor struck.

Research Approach and Methods

Cohort Approach

In this application, the cohort approach considers a population group that experienced an earthquake. It attempts to document injury occurrence as related to different types and levels of exposure. A promising variation of this approach focuses on the performance of a specific building and the fate of its occupants (42). A useful approach for exploratory studies, the building-specific case study is especially appropriate for investigations of the survival of occupants of collapsed buildings - an area receiving little systematic investigation. In cases where most building occupants are available for interview and where detailed damage information is available, this strategy permits a precise measurement of the exposure with respect to specific people to be studied, and enables the calculation of injury rates and severity attributed to specific elements and situations.

The disadvantage of this approach is that generalization is limited. For one, the results may be influenced by unique features of the situation not found elsewhere; for another, it is difficult to combine or generalize from other case studies that use different methods and/or data gathering instruments (43).

Specific Methods

We used a semi-structured, self-administered questionnaire to document the residents' experience during and following the September 19 earthquake. The questionnaire covered a number of topics, including personal characteristics, location at the beginning of the earthquake, actions and movements (from the beginning of the shaking until exiting from the building; or where applicable, actions from time of entrapment until rescue), the cause, type, and severity of injury, the search-and-rescue process, the receipt of medical care, and the longer term impacts of the injury.

We subsequently made contact with twenty-eight of the surviving residents, interviewing twenty-four. Of the 40 surviving residents, 12 had graduated and could not be located, twenty-four agreed to be interviewed for this study (including six who were not actually in the building at the time of the collapse), and four declined, finding the remembrance of their experience too traumatic. Of the twenty-four people interviewed, twenty-two were male and two female. The surviving residents ranged in age from 27 to 32. Since we had no knowledge of the exact number of "illegal" residents, or the whereabouts of survivors from this group, we excluded them from this investigation.

Results

General Factors Timing and Location

Timing as a Factor

Other earthquake injury investigations show that the time of earthquake occurrence is a key factor in determining occupant exposure to specific hazards (42). For example, when the September 19 earthquake occurred, at least thirty-two of the seventy-six residents were in the dormitory (which collapsed): Eleven were in the Obstetrics and Gynecology Hospital building (that also collapsed), and two were in a basement corridor (that sustained no damage) connecting the dormitory with a cafeteria. If the earthquake had occurred later in the day, fewer residents would likely have been in the two collapsed buildings and more would have been distributed in other buildings on the General Hospital grounds.

Location of Survivors Within the Building

The eighteen surviving residents interviewed, located in the dormitory at the time of the quake, were residing on the fourth to the eighth floors. One person was on the fourth floor, four on the fifth, three on the sixth, six on the seventh, and four on the eight. The four interviewed residents in the Obstetrics and Gynecology Hospital were making rounds on the fifth and sixth floor(s).

We asked the survivors to indicate others in the same physical setting at the beginning of the earthquake. We found that at least eight of the deceased residents were in the dormitory, and three were in the Obstetrics and Gynecology Hospital building.

The earthquake occurred at 7:17 a m., a time when most medical residents were preparing for the day. Consequently, seventeen residents were in their rooms, one was on the same floor as his room, two were in a corridor connected to the cafeteria on their way to breakfast, and four were already at work, making rounds at the nearby six-story Obstetrics and Gynecology Hospital that also collapsed.

Actions of Survivors

We know from our previous injury investigations that occupants of buildings that don't collapse frequently are able to engage in several actions prior to their injury. However, occupant capability in a collapsing structure (rather than a collapsed structure) has not previously been documented. We assume that if people are able to act, then training can possibly reduce their chance of injury. On the other hand, if building occupants cannot act, then training is useless.

We asked respondents to list their actions from the start of the shaking until they were safely outside the building. We asked them to show us when, in this activity sequence, the building collapsed, what actions they engaged in, and what was the sequence of actions. We found that numerous occupants had both the time and ability to take several actions before the building collapsed. The actions of occupants in each physical setting are summarized below.

The fact that most respondents indicated that their arms or legs were trapped in the collapse reinforces the conclusion that their indicated activities did actually occur prior to collapse.

Sequence of Actions of Survivors in the Dormitory

All eighteen residents in the dormitory at the start of the shaking reported being trapped at some point after the temblor began. Eight of these residents said that their first action before being trapped was to "stay where they were" because they thought it was the right thing to do. Five of these eight were subsequently trapped without engaging in any other further action; the remaining three later engaged in several actions before being trapped; one getting to a doorway, and two others leaving their dormitory rooms and going into the corridor.

Three others of the eighteen trapped residents said that their first action was to "stay where they were" because they were unable to move (probably because of the shaking intensity in those seconds). Two of these residents were then trapped immediately. The third was able seconds later to leave the space he was in, go into the corridor, and start up a fire stair before being trapped.

The other seven trapped residents were able to leave their rooms and either move to the corridor or stand in a doorway.

Sequence of Actions of Survivors in the Underground Corridor

The two residents in the underground corridor were able to leave the building prior to its collapse. One of these residents subsequently administered medical services to his comrades.

Sequence of Actions of Survivors in the Obstetrics and Gynecology Hospital

Two of the four residents, in the Obstetrics and Gynecology Hospital building when the shaking started, stayed where they were because they thought it was the right thing to do. One of these eventually moved into a doorway. The remaining two immediately moved into a doorway where they were trapped.

Injuries

Of the twenty-four surviving medical residents surveyed, twenty-two suffered an earthquake related injury or medical problem. Only two respondents had neither, and these had been located in a corridor outside of the building. Eighteen residents suffered multiple injuries. Table 28 contains a line list of the specific injuries suffered by each respondent. Sixteen of the twenty-two required further treatment, with thirteen requiring hospitalization. In addition, eleven respondents indicated that particular earthquake injuries resulted in a continuing disability.

The Role of Structural and Nonstructural Elements and Building Contents in Life Preservation

When reinforced concrete buildings collapse, structural elements, nonstructural elements, and building contents often interact to form space pockets within which occupants might be able to survive (46). With this possibility in mind, we asked our twenty-four survivors if building elements, furniture, or equipment prevented a total collapse in the area that they were in.

The results showed that building elements and building contents clearly played a role in life safety. Ten of the twenty-two residents (excluding the two who were out of danger) reported such an effect. Eight of the eighteen responding residents located in the dormitory building and two of the four in the Obstetrics and Gynecology Hospital building reported that elements or contents performed a role in survival. Table 29 presents a further breakdown of responses.

Entrapment

Duration of Entrapment

All of the residents interviewed but the two in basement corridors (22 of the 24 respondents) were trapped. The survivors were mostly rescued quickly. The duration of entrapment ranged from twenty minutes to seventy-two hours, with eighteen of the twenty-two trapped for under five hours, and twenty-one of the twenty-two trapped for under twenty-four hours. A breakdown of entrapment time is presented in Figure 12.


Figure 12. Duration of Entrapment (first 24 hours)

Occupant Actions After Entrapment

If trapped occupants are capable of taking action following a collapse, they can presumably assist themselves. We therefore asked respondents if 1) they attempted to free themselves, and 2) if their attempts were successful.

Table 28. Specific Injuries and Severity by Case

Case

Age

Sex

Injury

Locus

Severity

1

28

M

a) lacerations

back and shoulders

light




b) contusions

entire body

light

2

29

M

a) contusion

lumber region of back

mod/severe

3

28

M

a) multiple abrasions

entire body

moderate

4

29

M

a) laceration

forehead

moderate




b) fracture

left femur

severe




c) open contusions

arm and forearm

moderate




d) compression

thorax

moderate

5

32

M

a) fracture

right tibia

severe

6

32

M

a) crush

left ankle

moderate




b) hemorrhage

left elbow region

severe




c) contusion

right hip

severe

7

30

M

a) fracture

severe





b) fracture

left femur

severe

8

28

M

a) fracture

left femur

severe

9

27

M

a) scratched cornea

left eye

light




b) rupture

left calf

moderate




c) ruptured tendons

right foot

light

10

27

M

a) fracture

right rib

moderate




b) superficial crush


moderate




c) cuts and scratches


light

11

29

M

a) laceration

scalp

light




b) laceration

hand and forearm

severe




c) fracture

hand and forearm

severe

12

31

M

a) fracture

both ankles

severe

13

29

M

a) fracture

pelvis

severe




b) laceration

right side

light

14

28

M

a) fracture

right femur

severe




b) fracture

hip

severe




c) fracture

right leg

severe

15

28

M

a) sprain

right ankle

moderate




b) cuts and bruises

entire body

light/mod.

16

27

M

a) fracture


moderate




b) contusion

spine

moderate




c) fracture

lower back

severe

17

26

F

a) fracture

right knee

moderate




b) wound and compression

anterior tibia nerve

severe




c) bruise

right knee

severe

18

26

F

a) compound, exposed fracture

left foot or leg

severe




b) dislocation

left foot or leg

severe

19

28

M

a) sprain

torso

moderate

20

28

M

a) lacerations

front and back

light

21

30

M

a) compression

pelvis

moderate

22

27

M

a) fracture

left femur

severe




b) fracture

left humerus

severe




c) scratches and bruises

entire body

light

Table 29. Performance of Building Elements and Contents


Dormitory N = 18


Obstetrics Hospital N = 4


Yes

No

Don't know


Yes

No

Don't know

a) structural

4

14

0


0

4

0

b) nonstructural

1

17

0


1

3

0

c) contents

2

16

0


1

3

0

d) don't know

1

0

1


0

0

0

Only seven of the residents trapped in the dormitory attempted to free themselves, and only one was successful. The resident who was successful had his right foot bent and trapped. He was able to free the foot, move debris out of the way, and get to the exit.

At least three of the remaining six who tried unsuccessfully to free themselves had a trapped appendage. One resident couldn't free his legs. Another couldn't get completely free, although he was ultimately able to free his right hand. A third resident had almost escaped from the building but the cot/apse caught his foot, leaving his body hanging outside the collapsed structure. Three others reported that they were unable to move by themselves. These data suggest that in certain types of collapses, injuries or building obstacles make self-help nearly impossible.

Nature of Entrapment

We asked a series of open-ended questions about entrapment. One question helped characterize the physical setting surrounding the trapped person. We obtained the following results. In eight cases, respondents indicated that part of their body was trapped by falling debris, leaving them with little room in which to maneuver. This manner of entrapment (a function of collapse pattern) determined whether or not they attempted to free themselves, and how successful they were. A second important factor in entrapment was the type and severity of injury.

Who Were the Rescuers?

Knowing who actually is likely to rescue trapped earthquake victims is important to effective training and preparedness planning. Twenty-two of our twenty-four respondents were trapped. Of these, only one was able to free himself. Table 30 provides a breakdown of who actually performed the rescues. Thirteen residents were freed by fellow residents. Three residents reported being freed by a combination of fellow residents and other hospital workers. Only two trapped residents were freed by search-and-rescue teams.

Table 30. Rescuers by Building Type (N = 21)

Rescuers

Dormitory

OBGYN

Total

Fellow residents

12

1

13

Fellow residents and other hospital personnel

2

1

3

Fellow residents and search and rescue personnel




Fellow residents and taxi driver

0

1

1

Search and Rescue Personnel

0

1

1

Police

1

0

1

Rescue Techniques

We also asked people what methods were used in finding them, including how they knew that rescuers were near. In seventeen of twenty-one cases, the method used was human voices - calling out by both the rescuers and by those trapped. Other methods mentioned were knocking and making noise with objects such as an iron bar. No one mentioned sophisticated technology as a method of contact. Voices were likely the predominant method because colleagues performed most of the rescues prior to organized search and rescue and use of sophisticated equipment at these collapse sites.

Key Fears Of Those Trapped

We also asked victims if, while they were trapped, any of a list of factors constituted a serious threat to their survival. The results appear in Table 31. Most of the trapped residents indicated that the major threat was the removal of rubble. Those trapped heard the rescue operations occurring and were afraid that they would be crushed as rubble was shifted and removed.

Table 31. Fears of Those Trapped

Fears

Trapped persons

Removal of debris(afraid of being crushed)

15

Severity of injury

7

Lack of oxygen

6

Lack of liquids

3

Other (pressure on abdomen)

1

Gas

0

Fire

0

Discussion of Subsequent Results, Conclusions, and Implications

The following discussion integrates the previous results with our subsequent research on survival in buildings collapse both in Mexico City (case series of 120 survivors) and San Salvador (cohort study; five-story reinforced concrete office building) (49) (50), completed since publication of the Spanish version of this paper in 1988.

These cumulative research results have important implications for future earthquake casualty estimations and their application to preparedness, response, and recovery planning. The results suggest mitigation approaches for reducing the demand for health services and for increasing the supply of these resources.

First, complete building collapse does not result in death for all building occupants. Even collapsed structures can have substantial numbers of survivors. For example, three reinforced concrete buildings investigated so far had gross survival rates of 24%, 45% and 73% respectively suggesting that collapse patterns affect survival rates (53).

Second, we found that severe injury in structural collapse is not the exclusive misfortune of the trapped. Thirty-eight percent (38%) of the nontrapped occupants surveyed suffered a severe injury. Conversely, a substantial proportion (37%) of those trapped suffered relatively minor injuries (52). Although General Hospital enabled the victims to be rescued by medical personnel, we found that the majority of trapped, injured in other settings, were either rescued by lay persons (co-workers, family, and neighbors with no rescue or medical skills) or managed to free themselves. In addition, even trained rescue personnel did not have medical expertise (47).

Third, our analysis of the proportion of major to minor injuries rescued by each type of rescuer revealed that over one-third of those rescued by each group, for example by lay persons, had serious injuries. Untrained rescuers encountered serious injuries that were immediately life threatening or could lead to permanent disability if mishandled. They encountered those injuries immediately or very soon after the collapse (42) (47).

Fourth, because co-workers in the 50 building General Hospital medical complex performed most of the rescues (only two respondents were rescued exclusively by search-and-rescue teams), training should focus on co-workers for essential rescue activities. This training seems essential to the survival of coworkers, as most residents in the collapsed building were trapped if not killed. The ways that people are entrapped are important to the possible training needs of co-workers.

In addition, our overall data on entrapment and rescue support the view that search-and-rescue teams (as presently constituted) provide relatively little help in saving lives. Most of those trapped were freed before organized search-and-rescue efforts could be launched. By the time the teams arrive, especially foreign teams, and have decided on methods, most survivors have been rescued. Both the relatively short duration of entrapment and the frequency of rescue by fellow residents or other hospital personnel underscore the importance of local resources and training. The injury list (Table 28) provides an inventory of types of injuries that first responders might anticipate and be trained to cope with. It also suggests the need for careful evaluation of rescue techniques.

Fifth, the results showed that building elements and building contents clearly played a role in life safety. The data have clear implications for training. Building occupants can be taught to identify possible shelters as well as hazards in the buildings they occupy. The findings also support our contention that occupants can act. These results have important implications for training (51). However, they do not clarify what specific actions helped. We need further investigation to achieve results from which such information can be derived.

Sixth, the length of time entrapment is crucial to survivability, especially to those who are severely injured. It is commonly accepted that twenty-four hours is the cut-off time for the survival of large numbers of trapped victims. Because most medical residents were rescued quickly, and because 36% of our overall group of trapped respondents were freed within two hours, and 72% within eight hours (53), our data do not help answer, at present, a key question: When did the people who perished begin to die? We therefore cannot confirm or refute the twenty-four hour cut-off time for survival.

Seventh, the rapid rescues also call into question the potential role of foreign rescue teams. We know that foreign search-and-rescue specialists do in fact rescue people. Since eighteen of our twenty-four respondents were trapped under five hours, and all but one under twenty-four hours, a rescue team that arrives twenty-four hours after a collapse is not going to be effective. Consequently, foreign teams (depending on the country) may be more important as purveyors of equipment and technology rather than of manpower.

These findings underscore the importance of the immediate availability of stabilizing care at the collapse site and the close proximity of definitive care. They also underscore the need to integrate medical assistance into the search-and-rescue process, and to more effectively combine "external" and "internal" expertise and resources. The external component includes the deployment of medical personnel and technology for the most efficient and effective coverage of collapse sites. The internal component rests on improving the capabilities of lay persons for providing detection, screening, stabilizing care and to a certain extent transportation.

Our findings also imply increased attention on managing rescue operations, training rescuers, including lay persons, in appropriate medical care, and integrating emergency medical response and local health care system capabilities in light of expected injury patterns. Specific actions include developing localized search-and-rescue procedures that are sensitive to the physical setting circumstances of victims, improved light and heavy search-and-rescue training for corporations and individuals, and improved contingency planning for local hospitals.

Durkin Associates is further investigating these and other heavily damaged structures to shed more light on the previous questions. We are locating, contacting, and interviewing additional collapse survivors to develop a more complete basis for survival and injury rates and their attribution. We are examining the influence of such factors as occupant location, action, evacuation, damage patterns, and available resources on injury, survival, and search and rescue. Our ultimate goal is to develop a comprehensive earthquake casualty estimation process for integration in loss estimation methods.

To obtain copies of this and other Durkin Associates reports write or call Michael Durkin, Michael E. Durkin and Associates, 22955 Leonora Drive, Woodland Hills, CA, 91367, (818) 704-1493.

(Copyright 1987, Michael E. Durkin, not for publication, quotation, or citation without written permission).