The 1980 earthquake in southern Italy: rescue of trapped victims and mortality*

Bulletin ofthe World Health Organization, 61(6): 1021-1025 (1983) © World Health Organization 1983 The 1980 earthquake in southern Italy: rescue of trapped victims and mortality* MARC DE BRUYCKER,1 DONATO GRECO,2 ISIDORO ANNINO,3 MARIA A. STAZI,2 NICOLA DE RUGGIERO,3 MARIA TRIASSI,3 YVES P. DE KETTENIS,' & MICHEL F. LECHATI A retrospective survey was undertaken on the health effects of the 1980 earthquake in southern Italy. The study population included 3619 people living in 7 villages situated near the epicentre ofthe disaster. The overall casualty rate (dead and injured) was 19. 7%. Nearly all the deaths (192/202) occurred among trapped people who died before they could be rescued. Eighty per cent of all the trapped people were extricated within 2 days, mostly without the use ofsophisticated means. The probability ofsurvival decreased sharply, the longer the time before extrication. The crude mortality during the 18 monthsfollowing the earthquake was 19.0 per thousand among the injured people who received treatment, and 14.1 per thousand among non-injured people. After age standardization, there was no significant difference between these two figures and the expected mortality figuresfor the Italian population in normal times (14.4 per thousand). These results stress the importance ofproviding rescue activities in the first 48 hours after the impact. Strengthening the self- reliance of the community in disaster preparedness is suggested as the best way to improve the effectiveness of relief operations. In disaster-prone areas, training and education in methods of rescue should be an integral part of any primary health care programme. On 23 November 1980, at 19 h 34, an earthquake struck southern Italy, the epicentre of which was in the region of Campania. A study of the health effects of this disaster was undertaken with special attention to the casualties. The object of the study was to identify the risk factors, the most appropriate type of rescue operation and medical aid in terms of per- sonnel and material, and the optimum time for their introduction. SUBJECTS AND METHODS The people who were exposed to the earthquake of 23 November 1980 and who were living in the most affected area constituted the reference population. Seven villages covering a surface area of 261 km2 were * Requests for reprints should be addressed to Professor M. F. Lechat, School of Public Health, UCL, Centre for Research on the Epidemiology of Disasters, Clos Chapelle-aux-Champs 30, B-1200 Brussels, Belgium. ' WHO Collaborating Centre for Research on the Epidemiology of Disasters, School of Public Health, Catholic University of Louvain, Brussels, Belgium. 2 Istituto Superiore di Sanita, Laboratorio di Epidemiologia e Biostatistica, Rome, Italy. 3 Istituto di Igiene della 11 Facolta di Medicina e Chirurgia, Naples, Italy. selected according to three criteria: (1) a crude injury rate of over 50 per thousand, (2) a mortality rate of over 20 per thousand, and (3) a seismic activity of grade 8 or more on the Mercalli scale. The population registers were used for sampling, at random, one out of every three households totalling 3619 persons who were present on the day of the disaster. A house-to-house retrospective survey was performed by local interviewers using a standard questionnaire. Two matters were important for operational pur- poses: (1) the mortality immediately after the impact or, in the case of trapped people, at the time of extri- cation (these were defined as deaths from inmpact); (2) the subsequent mortality rate, up to 18 months following the earthquake. The efficacy of any inter- vention could then be assessed in terms of mortality, thus quantitatively relating the types of rescue action at different times to the outcome. People were said to be "trapped" when they had been buried underneath debris or been locked into an enclosed space and needed help from outside to be freed. They were classified as "injured" when indi- viduals reported with injuries and could indicate the part of the body that was affected. People who died before they could be rescued were not considered in the calculation of the injury rates. A "casualty" was defined as a person either dead or injured. 4370 -1021 M. DE BRUYCKER ET AL. A follow-up of those individuals who had reported with injuries and had been accepted for medical care was made through a study of their hospital records. However, details on the management of these injured people are not included in this paper. Software designed specifically for this study was developed for the use of a microcomputer in the field. This reduced the time for collection, validation and processing of the data. RESULTS The total number of deaths from the moment of impact until 18 months later was 256 (7%) in the sampled population of 3619 persons who had been exposed to the earthquake. Immediate effects The death rate, including those found dead on ex- trication from the debris, was 5.6% (202/3619). Of the exposed population, 15.1 07o (548) had been trapped. Nearly all the deaths (95%0) were among trapped people who died before they could be rescued. The mortality in this group of trapped people was 35.0% (192), compared with 0.3% (10) among the non-trapped persons. The overall injury rate was 14.2% and the ratio of injured persons to dead persons was 2.5:1. Among the trapped people this ratio was 1.3:1. The overall casualty rate (dead and injured) was 19.8%. The casualty rate among the trapped persons was 80.0% compared with 9.0% among those who were not trapped. Follow-up All the people who had escaped death at impact or been extricated alive were followed up for 18 months (Table 1). The subsequent mortality among the non- trapped people was 13.4 per thousand (41/3061), compared with 36.4 per thousand (13/356) among those who had been trapped. The mortality rate, over the same period, differed widely between people who escaped injury and those who were injured. Among those who escaped injury, it was 12.9 per thousand for the non-trapped victims and 46.3 per thousand for the trapped. For the injured people, the mortality rate was calculated separately according to whether treat- ment had or had not been provided; 92 persons (17.8%) did not receive treatment, and 422 (82.2%) were given medical care on the spot or were evacuated to a medical centre either directly or after first aid on the field. Among the treated people, the mortality rate was 19.0 per thousand (8/422) during the 18 months after the earthquake, the rates for those who had and had not been trapped being 22.2 and 15.3 per thousand, respectively. Of the 92 injured people who were not treated, 5 died (3 of these had been trapped victims who died in the immediate post-impact period, i.e., the 24 hours after the earthquake, and before any treatment could be applied). Several individual and environmental factors (places, circumstances) have been correlated with the casualties. The fact of being trapped is in itself the most significant risk factor for injury and death. Table 2 shows that for those who were trapped, the probability of being alive at extrication from the debris declined sharply over time. From the fourth day onwards, no one was found alive. Fig. 1 shows the time distribution of people extricated after the Table 1. Number of deaths and mortality rates among the people who escaped death at impact, or who were extri- cated alive, during the 18 months following the earthquake Mortality No. of exposed people who escaped death at impact or who were extricated alive No. rates (per 1000) not injured 36 12.9 Not trapped (2795)notrae(3061) nottrete injured treated 2 28.6(266) ((19 6)3 5. treated 19.0 25.3 injured (226) 522.2J I (248) Jnot treated Trapped )(22) 3 136.4 (356) not injured 5 46.3 (108) 1022 THE 1980 EARTHQUAKE IN SOUTHERN ITALY Table 2. Probability of being alive, in relation to the number of days trapped, based on 548 casualties rescued after the earthquake No. of days Probability of after the No. of persons No. of persons being alive at earthquake extricated alive extrication (%) 0 379 333 87.9 1 57 20 35.3 2-3 35 3 8.5 >, 4 77 0 0.0 lOOr 0 - x -a. 0 - 0 461 471 436 3.6 - 379/ *-~353 355 356Ao _A ___ --- -4 O ALI I06A 1 2 3 4 Days 100 T 93.5 g (D 0 x 0 a. m _ v Time ( LibSo) =8h -Cumulative number of trapped people extricated ----Cumulative number of people extricated alive Fig. 1. Numbers and percentages of trapped people extricated, and the numbers alive during the first 4 days. impact. During the first day after the earthquake, 134 people (24.40o) were extricated within half an hour, 241 (44.0%) within 3 hours, and 305 (55.7%o) within 12 hours. The time within which 50%o of the trapped victims had been extricated (Time Libso) was 8 hours. Within two days, 79.6%o (436) of all the trapped people were extricated. Of all the people extricated alive, 93.6%o (333) were rescued within the first 24 hours. An analysis was made of the place of origin of the persons who helped in extricating the victims. As regards injured victims who were also evacuated, it was shown (Table 3) that 90.0%o of them had been ex- tricated by the inhabitants of the same village, mainly by persons residing in the same building (76.207o). Table 3. Place of residence of the people who helped in extricating the casualties and the number of victims rescued and evacuated to a medical centre Injured victims evacuated Place of residence No. % (1) Same house as victim 121 76.2 (2) Another house in the village 22 13.8 Subtotal: (1) + (2) 143 90.0 (3) Outside the village 13 8.2 (4) Unknown 3 1.8 Total 159 100.0 The means used for the extrication of injured victims who were later evacuated to a medical centre are shown in Table 4. These means were divided into two groups: unsophisticated (hands, shovel, ladder) and sophisticated (tractor, crane). Most of the severely injured victims (96.8%) were extricated with- out any sophisticated equipment. On the first day, only simple measures were available, but by the second day more sophisticated means came into use. For each of the seven villages, the proportion of people extricated alive and the proportion of use of sophisticated means were calculated. There was a negative correlation between these two indices (r = - 0.76). When only the first day after the impact was considered, no correlation was found, indicating that an earlier use of sophisticated means would not have improved survival rates. DISCUSSION Few studies have attempted to determine the factors specifically responsible for deaths at impact; 1023 79 I 24 1024 M. DE BRUYCKER ET AL. Table 4. Means used for extricating trapped persons, by time' No. of No. of trapped persons injured people (dead included) extricated on: Means of evacuated to a extrication medical centre Day 1 Day 2 Day 3 and later Hands 119 (74.8) 297 (76.5) 10 (17.2) 13 (12.7) Shovel 33 (20.7) 58 (14.9) 28 (48.3) 33 (32.4) Ladder 2 (1.3) 14 (3.6) 1 (1.7) 1 (1.0 ) Subtotal 154 (96.8) 369 (95.0) 39 (67.2) 47 (46.1) Tractor, crane 5 (3.2) 10 (2.6) 12 (20.7) 36 (35.3) Unknown 9 (2.4) 7 (12.1) 19 (18.6) Total 159 (100) 388 (100) 58 (100) 102 (100) a Figures in parentheses give percentages. they have mainly concentrated on the use of different building materials and techniques in relation to the risk of death (J).' In the present study, the fact of being trapped was the single most significant risk factor. Indices have been proposed for assessing the health effects of natural disasters and related needs (2). One of these indices is the time-related death rate follow- ing the impact, taking into consideration the rescue measures used and the self-reliance of the com- munity. The mortality figures over a period of 18 months after the 1980 earthquake in southern Italy show that, once people were liberated alive from the debris, very few died. Indeed, of the 514 injured victims, only 13 died over this period; 3 of them died within 24 hours of the earthquake before any treat- ment could be applied. In this 18-month period, 8 people, irrespective of being trapped, died among the 422 victims who were treated. The expected number of deaths for the same age-standardized Italian popu- lation is 9.87. There is thus no evidence of any in- creased mortality among the earthquake survivors who received medical guidance. Medical interven- tion, after the earthquake, on those still requiring attention (e.g., the injured victims extricated alive) was very effective, with only a narrow margin for improvement, if any. Increasing the medical re- sources available for helping the trapped victims after their extrication would therefore not be expected to be very cost-effective. It was not possible to ascertain the cause of death in the 5 trapped victims who escaped injury at impact but died subsequently (Table 1). The high death rate a MITCHELL, W. A. ET AL. Predicting casualties and damages caused by the earthquake in Turkey: apreliminary report. Colorado, US Air Force Academy, Department of Economics, Geography and Management, 1978 (USAFA-TN-78-2). (46.3 7) observed in this group calls for further study. Trapped people with no recognizable injury after the impact could constitute a potentially vulnerable population. The great majority (95%) of deaths were among trapped people who died before they could be rescued. Although the proportion of persons who died instantly on impact is not known, the probability of survival in relation to time after the impact leads to the conclusion that if any significant reduction in earthquake mortality is to be achieved, attention should be given to rescue action within the first 2 days after the impact. The aim should be to decrease the number of victims found dead at rescue (35070 of the trapped). The death rate of people liberated from the debris was about the same for all age groups exposed; only the oldest age group experienced a slightly higher mortality (1.5 times the mean). Similar mortality pro- files were observed by de Ville et al. (3) after the earthquake in Guatemala in 1976, and by Sommer & Mosley (4) after the cyclonic storm of 1970 in Bangladesh. From Fig. 1 and Table 2, it is clear that the bulk of the rescue work was done by the local people within 2 days of the impact. After this period, the probability of rescuing people alive is very low. A similar relationship between time of extrication and survival has been reported in Tang Chan, China, in 1976 (Wen-Kui Mai, personal communication). Con- sidering this short time period within which any effective rescue work can be done, the local people in the village itself are in fact the only available human resources. Moreover, the available means and skills for rescue (equipment, know-how) must also be con- sidered. There is no evidence that the use of sophisti- THE 1980 EARTHQUAKE IN SOUTHERN ITALY 1025 cated equipment, even very soon after the impact, can improve the survival rate among the rescued people. This could be because these methods were not appro- priate or because people were not sufficiently skilled in their use specifically for rescue operations. CONCLUSIONS A state of preparedness, with reliance on the com- munity's own resources and reinforcement of self- help, is the most important approach for the reduc- tion of mortality due to an earthquake. The dispatch of special rescue teams must be considered only if they are highly qualified and if they can be oper- ational within 24 hours. These considerations have clear implications for disaster preparedness. In disaster-prone areas, the training of community health workers and medical personnel, as well as lay persons, in methods of rescue, should be an integral part of any primary health care system. Locally appropriate rescue techniques should be introduced into the curriculum of such programmes. ACKNOWLEDGEMENTS This survey, carried out at the request of the Italian Ministry of Health and the World Health Organization, was conducted jointly by the Centre for Research on the Epidemiology of Disasters, University of Louvain in Brussels, Belgium, and by the Istituto Superiore di Sanita, Laboratorio di Epidemiologia e Biostatistica, in Rome, Italy, with the collaboration of the Istituto di Igiene della II Facolta di Medicina e Chirurgia in Naples (Professor B. Angelillo). The authors wish to thank the regional and local authorities for the administrative coordination of the field work, as well as the interviewers and their supervisors, especially Dr M. Moro and Dr T. Stroffolini. The processing of data by the micro- computer in the field was realized by C. Speeckaert, M.A. Stazi, E. Sansonetti and M. Pasquali, using the software prepared by the Department of Epidemiology, Catholic University of Louvain, Brussels. R. Scipioni, F. Timperi, G. Siepi and L. Fantozzi offered technical assistance during the survey. The support given by and valid suggestions from Professor A. Bouckaert and Professor A. Zampieri are gratefully acknowledged. Special thanks are due to M. A. Parent for help in the writing of the paper. RtSUME LA TREMBLEMENT DE TERRE DE 1980 EN ITALIE DU SUD: SAUVETAGE DES VICTIMES ENSEVELIES ET MORTALITE Une etude retrospective des effets sur la sante du tremble- ment de terre de 1980, dans le sud de l'Italie, a e r6alisee. La population etudiee consiste en un echantillon de 3619 personnes habitant 7 villages situes pres de l'epicentre du seisme. Le pourcentage total des blesses (morts et vivants) repre- sente 19,7%. Presque tous les morts (192 sur 202) sont des personnes emprisonn&es mortes avant d'avoir e secourues. Quatre-vingts pour cent des personnes ensevelies ont e degagees au cours des deux premiers jours, la plupart sans utilisation de moyens sophistiques. Au-dela de ce temps, la probabilite de survie au moment du degagement d6crolit rapidement. A partir du 4eme jour apres le seisme, on ne trouve plus aucune personne en vie. Parmi les personnes degagees vivantes, 93,6% l'ont e le premier jour. Parmi les blesses evacues ulterieurement, 90% ont e extraits par les habitants du village meme et, principalement, par les habi- tants du meme immeuble (76,2%). Le taux brut de mortalite en 18 mois est de 19,0 pour mille parmi les blesses qui recoivent un traitement et de 14,1% pour mille parmi la population non bless&e. Apres standar- disation des ages, on constate qu'il n'y a pas de difference significative entre ces deux taux et le taux de mortalite de la population italienne en temps normal (14,4 pour mille). Les resultats montrent l'importance primordiale des acti- vites de secours dans les 48 premieres heures apres le seisme. La meilleure voie pour ameliorer l'efficacite des secours se trouve dans le renforcement de l'autosuffisance de la communaute sinistree. Dans les regions predisposees aux desastres, formation et education au sauvetage doivent etre partie integrante de tout programme de soins de sante primaires. REFERENCES 1. GLASS, R. I. ET AL. Earthquake injuries related to housing in Guatemalan village. Science, 197: 4304 (1977). 2. LECHAT, M. F. Disasters and public health. Bulletin of the World Health Organization, 57: 11-17 (1979). 3. DE VILLE DE GOYET, C. ET AL. Earthquake in Guatemala: epidemiologic evaluation of the relief effort. Bulletin of the Pan American Health Organization, 10: 2 (1976). 4. SOMMER, A. & MOSLEY W. H. East Bengal cyclone of November 1970: epidemiological approach to disaster assessment. Lancet, 1: 1029-36 (1972).