Vi använder kakor för att förbättra webbplatsen

Den här webbplatsen använder kakor (cookies) för att underlätta ditt besök och för att göra webbplatsen bättre. Läs mer om kakor.


The terror attacks in Madrid, Spain, 2004 – KAMEDO-report 90

The Kamedo-reports are published by the Swedish Disaster Medicine Study Organisation (Kamedo), at the National Board of Health and Welfare. Observers study the medical, psychological, organisational and social aspects of disasters. The results, with a focus on experiences gained, are presented in the reports.


Course of events

A total of ten bombs placed on four different trains were detonated during the terrorist attack in Madrid on March 11, 2004. Fourteen explosive devices had been prepared and put in rucksacks and sports bags which had then been placed on trains at Acalá de Henares station east of Madrid. Each bomb contained explosive material and a detonator which was connected to the alarm function of a mobile phone.

The first bomb exploded at 7.39 am, after which a further nine bombs were detonated within a period of a few minutes. Four bombs never exploded. The locations that were attacked were the central station Atocha and the stations at Santa Eugenia and El Pozo plus the area around the station Téllez. The trains which were attacked at Téllez and El Pozo were each estimated to have between 1,000 and 1,800 passengers onboard.

This incident is the most serious that has occurred in a European country in peacetime. A total of 191 people were killed and more than 1,500 injured. The magnitude of the attack called for the mobilisation of resources from several municipalities in the region. This resulted in not only the regional but also the national command organisation being activated – something which has previously never happened.


During the first 16 hours more than 22,000 incoming calls were registered at the emergency service centre Madrid 112 in connection with the attacks. There was at no time a shortage of personnel at the centre since the incidents occurred at the same time as a shift change and the shift that was going off duty was ordered to remain. About 2/3 of the available switchboard capacity at the centre was engaged.

The first call concerning the attack at Atocha central station came in to the emergency service centre Madrid (112) at 7.39 am on March 11. At the same time as the person placing the call was being questioned the alarm was being forwarded to the emergency service centres for the two ambulance organisations in Madrid – Summa 112 and Samur. The alarm was then also forwarded to the police and the emergency and rescue service as well as to other concerned authorities. After some time, as calls came in concerning the other attack locations, this information was also forwarded. At about 8.30 am an emergency response regional command centre was set up in Madrid. A little later, at about 10.00 am, the Spanish government established co-ordination at a national level.

Madrid 112 was used as an information centre during the day, meaning that authorities and relatives could call there for information about the hospitals which victims had been taken to.

Prehospital medical care

The first ambulance arrived at Atocha seven minutes after the alarm was raised, and the first critically injured patient was transported from there 23 minutes later, following the arrival of the Samur ambulance organisation. Prior to this a large number of victims had taken themselves to hospitals in taxis or private vehicles.

Within 30 minutes of the alarm being raised collection points in hospital tents had been set up at all four incident sites. In addition, a sports hall at Téllez was used for this purpose. At one of the locations, at least, a tent was set up in direct proximity to a demolished railway carriage, thus placing it within the risk zone for possible further explosions.

There was a rapid flow-through at the collection points. Despite this, advanced treatment was applied at all of them, using intravenous access, intubation and pleural drainage. Life threatening haemorrhaging was also stopped by applying pressure bandages and cut-off bandages.

Medical transport and distribution

All the injured had been moved away from the incident sites by 10.17 am, i.e. 2 hours and 38 minutes after the first explosion. Medical transport officers at the sites attempted to distribute “their” patients evenly throughout Madrid’s emergency hospitals. There was, however, no distribution system or an updating of the situations at the various hospitals available at the incident sites.

In total 927 injured persons, of whom 165 were judged to be seriously wounded, were transported by ambulance to at least 15 different hospitals and clinics in Madrid. It did not come to light until later that the distribution among the hospitals had been rather uneven.

Prioritisation and command

No form of triage system by colour marking or similar for prioritisation was used, either at the incident sites or the collection points, despite the fact that the equipment for this was readily available. It was felt that it was “obvious who had received light, serious or critical injury to the extent that an indication system was not necessary”.

The prehospital care operation was organised through co-operation between Samur and Summa 112. Both organisations are among the best qualified ambulance organisations in the world, in terms of both availability of advanced technology and qualified personnel. The command on site was executed by Samur, the police, and the fire and rescue service. Both Summa 112 and Samur felt that they alone had led and co-ordinated the medical operations in the affected areas. According to the regulations the command of medical care operations lay with Samur.

Hospital care

Directly after the alarm the hospital management at Gregorio Marañón – one of the largest hospitals in Madrid and also one of the hospitals that received the most patients – took the decision to postpone all the operations planned for that day. This immediately made 22 operating theatres available. Following this, beds were made available by beginning the process of releasing patients. One hundred and sixty-one beds were made available within two hours, and within six hours 438 beds were available. Most of the intensive care patients could be moved to a lower care intensity level, intermediate care or post-operation care.

The emergency ward was organised in such a way that patients were categorised at the ambulance entrance. The most serious cases were taken to the, so called, trauma room. From there they were taken in for operation, to an ICU or to a ward that was opened especially in connection with this alert.

Relatives and confidentiality

All the hospitals were placed under extreme pressure from relatives. Just a few hours after the attacks more than 600 relatives had gathered at the accident & emergency ward at Gregorio Marañón hospital. They were shown to a large assembly hall where a list of injured patients was read out every 30 minutes.

Spanish law concerning confidentiality with regard to medication corresponds to Swedish law, but on this occasion a deliberate decision was made to override this regulation. This decision was initially taken by the affected hospitals separately, but was later sanctioned at regional and then national level.

A problem that soon affected the hospitals was that both the land line and mobile phone systems became overloaded and unserviceable. This caused difficulties in terms of information distribution both internally and externally.

Psychosocial care

The psychosocial care of the injured and their relatives at Gregorio Marañón hospital had not been prepared for in the disaster plans. The hospital management appointed a senior, experienced psychologist to take responsibility for this aspect. The point of departure was that ”the first priority was to save lives, but then to apply all possible resources to protect and restore mental health”. The next priority was to, firstly, care for the injured, and then the relatives and finally for the team that had cared for the victims.

According to the guidelines given in crisis management literature, one should gather families and separate them from other relatives. This was, however, not possible in this situation, since personnel had altogether too many relatives to attend to.

The significance of accurate information

Those responsible for psychosocial care soon saw the significance of giving accurate information. For this reason the names of the injured and where they were being cared for was read out. This information was updated every half hour. In order for the information to be as accurate as possible, it was first checked by a control group before being issued.

After two days a special website was set up showing a list of injured persons. The Ministry of the Interior took responsibility for the web site and the decision to publicise the list of patients. A list of those who had been died was not made public, however.

The hospitals also strove to give accurate information to the media, in order to avoid the spreading of rumours.

Identification of the dead

The bodies and remains of the dead were taken to Ifema, which is Madrid’s exhibition area. Here personnel from the ”Office for mental health” took on the task of caring for the relatives of the dead. Between March 11 and 13 a team of psychiatrists, psychologists, social workers and secretaries worked around the clock to support waiting relatives. Parallel to this Summa 112 organised voluntary psychologists who worked from March 11 to 22.

Identification of the dead went very quickly; after just 24 hours 155 of the total 191 bodies had been identified and the relatives notified.

On March 13 the identification work was moved from the exhibition area to a cemetery with a cold room. Also here psychosocial help was available to relatives. By March 15 all the dead had been identified and the work then finalised. In 37 cases DNA analysis was required in order to certify identification.

Psychosocial support in the aftermath

Children and young people who were involved appeared not to be in great need of psychological support initially, but when a mobile unit actively started visiting schools the number of visits rose significantly. These visits were initially discouraged by headmasters and teachers, who felt that there was no need for such support in their particular school.

In addition to this, two afternoon units for involved children were set up in two of the most affected areas; these were still open a year after the event.

Relief and debriefing of operative personnel was arranged partly through the medical service’s own channels, and partly through a strengthening of psychiatric care resources at healthcare centres.

The “Office for mental health” offered to provide assistance for police and fire and rescue service personnel. Both of these authorities, however, declined the offer for the reason that they would provide psychosocial relief via their own channels. Representatives of the Office felt that the explanation given by the management of the police and fire and rescue services reflected a concern that external help with crisis management would undermine morale in the units and lead an increase in sick-leave among staff.

Police investigation

The police investigation after the attack led to seven of the suspected terrorists being localised in a flat in the Leganés district on April 3. Police units surrounded the property and the neighbouring inhabitants were evacuated. Shots were fired during the break-in, after which the terrorists blew up both themselves and the flat. Continued police work during the following months resulted in the arrest of more than 20 other suspects, the majority from Morocco. The police suspect that there is a strong link between the Madrid attacks and the terrorist network al-Qaida.

Observer’s experiences and conclusions

Prehospital care

Ambulance organisations

Emergency prehospital activities in Madrid are divided between two organisations – Summa 112 and Samur. Less demanding medical transport is also provided by other ambulance organisations, e.g. the Red Cross.

Summa 112 is operational in the whole of the Madrid region, whereas Samur only operates in the city of Madrid. Consequently there are two emergency ambulance organisations operating in central Madrid, of which Summa 112 is responsible for emergency calls within private housing areas and Samur is responsible for incidents occurring outdoors and in some official buildings.

Comments: Both Summa 112 and Samur are among the best qualified ambulance organisations in the world in terms of advanced technology and qualified personnel.

The allocation of emergency services depending on whether the incident has occurred in private housing or a public place appears rather strange from a Swedish perspective. The representatives for the two organisations, however, saw no problem in this. There is obviously a degree of “rivalry” between the organisations, which was apparent during the actual event, in that both organisations felt that they alone had led and co-ordinated the healthcare operations at the incident sites.

Regulations stipulate that responsibility for the healthcare operations lay with Samur. The doctors who were sent out by Summa 112 were therefore assigned no managerial roles by Samur as they were working within an organisation that was not integrated with Samur. Instead these doctors were allocated tasks concerning individual patients.

The importance of recognising and adhering to the framework that sets the limits between co-operating organisations is made very apparent. In Swedish terms this would apply more closely to co-operation between the fire and rescue services, the police and the medical service, but it could also apply to the Civil Aviation Administration, Maritime Administration and the mountain rescue service when incidents occur within their respective areas of responsibility.

Summa 112 has at its disposal 70 ambulances and emergency vehicles, as well as two helicopters, while Samur has about 130 ambulances, emergency vehicles and motorcycles. Both organisations also have their own staff vehicles and special disaster vehicles. Samur has 30 emergency vehicles manned with doctors. These go out along with an ambulance in order to follow and assess the work of the ambulance crew in accordance with a special model. Directly after the response the ambulance crew is graded for its performance, and feedback on the response is given.

Comments: Even taking into account the fact that Madrid is about three times more populated than Stockholm, the total resources there are significantly greater than Stockholm’s. It is also interesting to note the primary purpose of the emergency vehicles manned with doctors, namely to assess the work of the ambulance crew. This resource is an excellent way of increasing the quality of ambulance responses, and it is also an extra resource should it be required.

Response plans and alarms

The response plan Platercam is based on the law regulating civil protection and emergency response. This law defines the areas of responsibility of the different authorities including the police, emergency and rescue service, and medical service. All responses begin at a local level (level 1). If the incident is of a scale requiring the mobilization of resources from several neighbouring municipalities, a regional command is set up (level 2), based at the Madrid 112-centre. The highest level is level 3, which is national. This level involves the Ministry of the Interior and the establishment of a national crisis cabinet to support the affected region.

At about 8.30 am, just under an hour after the explosions, Madrid 112 took the initiative to set up a regional command, i.e. level 2 in Platercam. Sometime later level 3 was activated by the Spanish government.

Comments: The Platercam response plan enables extremely rapid intervention and take-over of command at a national level. Sweden should consider establishing a model for a similar plan, i.e. for a rapid increase in preparedness and a take-over of the command of emergency and medical responses at the optimal level, based on the characteristics and scope of the incident.

The first incoming call concerning the explosions was received by Madrid 112 at 7.39 am on March 11. At the same time as the person placing the call was being questioned, the alarm report was forwarded to the emergency centres at Summa 112 and Samur in accordance with established procedures. In addition, the alarm was forwarded to the police and the emergency rescue service and other affected authorities.

Comments: A direct transfer of the alarm made it easy to rapidly forward the alarm to where it was required. Also in Sweden emergency centre operators hold a key position in terms of the initial assessment and alarm in a serious situation. They should therefore be trained for and should simulate such situations.

Collection points and evacuation

Collection points and emergency care tents were set up extremely quickly, within 30 minutes, at three of the incident sites - Atocha, Santa Eugenia and El Pozo. A sports hall was used for the purpose at Téllez. All the injured that were cared for by Samur passed through the collection points and were then transported using different types of ambulances.

The processing at the collection points went very quickly – from 1 hour 10 minutes to, at the most, 2 hours 28 minutes. Despite this, advanced treatment was applied at all of them, using intravenous access, intubation and pleural drainage. Life threatening haemorrhaging was also stopped by applying pressure bandages and cut-off bandages. A large number (about 30 percent) of those less seriously injured made their own way to a hospital, without passing through a collection point.

Comments: The spontaneous/wild evacuation of the injured using various forms of transport results in a swift influx of wounded to the hospitals, with the least injured often arriving first. This is a well known phenomenon when incidents occur close to hospitals, but it is something that is often disregarded when planning or training for major disasters. At the same time it is acknowledged that efforts should be made to avoid this as it creates disorder. The phenomenon is, however, in practice difficult if not impossible to influence.

It would perhaps, instead, be better to see it as an asset rather than an impediment in situations where the medical transport capacity is limited in relation to the number of injuries. The important thing is that the personnel at the hospital are aware of the phenomenon and plan their routines accordingly. In practice it means that the hospital emergency reception quickly receives a large number of wounded that have not been diagnosed, and consequently not prioritised either. Because of this the accident & emergency wards become in practical terms a part of the incident site.

According to Swedish disaster medicine the first priority in city environments is to move the wounded to hospital as quickly as possible, in line with the scoop and run principle. In Madrid, however, with a larger number of ambulances and close proximity of many large hospitals, the decision was made to set up collection points and to stabilise patients.

Both ambulance organisations consider that success lies in stabilising critically injured patients prior to transportation. In this instance they considered it to have been directly life-saving in many cases, in addition to relieving the pressure on the emergency wards. Such a strategy is based on the assumption that the stabilisation process, with the obvious increase in time taken to reach the hospital, can take place without increasing the risk of death or disablement. This assumption has never been substantiated.

From a Swedish perspective the above should be something to reflect upon. It is the opinion of the observers that collection points with protection and warmth should always be used for major disasters in remote areas with few ambulances and long transportation distances. There is perhaps also good reason to reconsider the scoop and run doctrine for highly populated areas in cases involving many and seriously injured. The time factor is, of course, critical in both situations, i.e. how long it takes to set up a tent or gain access to a building. Each separate case involving injury requires an assessment of the response tactics at an early stage, based on the prevailing conditions.

Prioritisation and distribution

Triage marking was not used in Madrid, despite the fact that the equipment was readily available. It was judged to be sufficiently obvious how seriously injured people were.

Comments: Within Swedish disaster medicine, training is carried out on a prioritisation basis; triage marking is included as a natural part of this. In Madrid there were no special disaster journals or injury cards, instead ambulance journals made of thin paper with a carbon function were used. The weather was favourable but had it been raining or snowing, this would not have worked. In Swedish conditions triage and triage marking such as injury cards should continue to be used as support for fast and appropriate care and as a part of the quality system during disaster medication training.

No form of distribution key or updating of the capacity situations at the hospitals was available at the incident sites. There are at least eight hospitals in Madrid with large emergency care capacity, and thus each medical transport officer tried to distribute “their” patients evenly to these.

In total 927 injured persons, of whom 165 were judged to be seriously wounded, were taken to hospital by ambulance. Not all the injuries were registered, however. There are also reports of people with slight injuries being referred directly to healthcare centres by the hospital emergency receptions, and some patients with minor injuries were treated without being registered before being sent home. Consequently the picture of the number of injuries is not completely accurate.

Comments: A common system for the distribution of the injured to the hospitals was never implemented. Despite the fact that a regional command was established quickly (8.30 am) three of the four incident sites were more or less evacuated before the command function was up and running (10.00 am). Shortcomings in the distribution of the injured to the various hospitals were compensated for to a large extent by the huge hospital capacity in Madrid and the fact that the hospitals were warned at such an early stage. The distribution of patients between the different hospitals was not optimal, however.

Spontaneous evacuation of patients, in combination with an intensive flow of patients to the closest, large hospitals, meant that the load placed on the key functions at the hospitals varied by the minute.

In practice, the hospitals lacked appropriate tools for describing and reporting their load, something that would have facilitated a more even distribution. In this situation of high loading, the staff abandoned the IT-system normally employed for reporting, among other things, the number of beds available. This indicates that the system was difficult to use and not suitable for a disaster situation.

Conditions are the same in Sweden, and it is probable that if a similar situation occurred in one of our large cities the outcome would be the same as in Madrid. Additionally, in Sweden the requirement for leadership and co-ordination is generally even higher both at incident sites and within the healthcare itself, since immediate hospital space is less readily available.

From a Swedish perspective one should learn from the events in Madrid and train and assess the leadership and reporting systems.

Hospital care

Leadership, prioritisation and communication

The hospitals were managed by those normally responsible, who received no information directly from the incident sites other than via the patients and ambulance personnel. During the most intensive period about 120 patients arrived per hour. Many of these were critically injured and some were intubated. Despite this, there was never the feeling of shortage of either places or personnel at either of the two hospitals that took in the majority of patients. Management of psychological care had not been planned in advance, but rather various solutions developed during the course of events. An information support system was not available but was developed throughout the day, which, among other things, involved some conscious departures from the laws concerning confidentiality.

Comments: The hospitals were managed by their normal management personnel. From a Swedish perspective this is an aspect of interest, as we would normally, in extreme situations, set up a special management group. It is also interesting that management of psychological care was not planned or trained for in advance, that no information system was in place and that conscious departures were made from the laws concerning confidentiality. This illustrates the importance of the hospital management being familiar with their roles, that psychological care is prepared for and that an information support system is available.

The two hospitals which we visited received many injured persons who had arrived via taxi, the police or private vehicles. These were not diagnosed by healthcare personnel at the incident sites and consequently triage functions were set up at the entrances. Those who could walk were shown to a waiting room, while those on stretchers were taken directly to wards. Some triage markers were never used, with the explanation that ”we didn’t have enough forms”.

Another difficulty at the hospital was the inability to transfer information. Communication was a problem since both the mobile and land line phone systems collapsed both internally and externally due to overloading.

Comments: Because both the triage marking of patients arriving at the hospital emergency reception, as well as the inability to transfer information within the hospital via normal and mobile phones, are probably just as sensitive to extreme loading here in Sweden, it is likely that the same difficulties would occur in the event of such extreme circumstances. It is therefore particularly important that these aspects are taken into account in our disaster plans.

Psychosocial care, confidentiality and debriefing

With regard to the psychosocial care of the injured and their relatives that was carried out at Madrid’s largest hospital, Gregorio Marañón, one could quickly draw the conclusion that the hospital had not included this aspect in its disaster plans. The hospital management quickly appointed an experienced psychologist to take responsibility for this. The decision was made to refer the relatives to Aula Magna – a large assembly hall in the hospital area. At the time of this decision being made some 600 relatives had gathered outside the hospital and several more joined later.

As information on the injuries reached the management for the psychosocial response, a list of the names of the injured and where they were being treated was read out. By doing this the laws concerning confidentiality were breached. The information was updated every half hour.

In order to make the information as accurate as possible, it was checked by a control group before being issued. A special web site containing lists of the injured was set up after 48 hours. Responsibility for the web site and the decision to publish lists of patients lay with the Ministry of the Interior. This decision also involved the laws concerning confidentiality not being adhered to. A list of the dead, on the other hand, was never made public.

The psychosocial support focused on three groups:

  • all those affected
  • children and young people affected
  • personnel who worked with the injured and the dead.

All those affected could call a special number if they required help or support. The call centre was manned by psychiatrists and psychologists and was open until the end of June 2004. Thirtysix call lines were available in addition to those normally functioning within Madrid’s mental health organisation.

Children and young people who were involved appeared initially not to be in great need of psychosocial support, but when a team actively started visiting schools with a mobile unit, the number of visits increased considerably and a clear internalisation of the anguish and suffering they were undergoing was brought to the surface. Afternoon centres for affected children were opened in two of the worst affected areas; these were still open a year after the event.

Relief and debriefing of operative personnel was arranged partly through the medical service’s own channels and partly through a strengthening of resources at healthcare centres and psychiatric clinics. The ”Office for mental health” offered help with debriefing for police and fire and rescue service personnel, but their participation was surprisingly low. Both organisations claimed that they managed these aspects via their own channels, but they have not made it easy for their personnel to apply for such help. Some people within the Madrid office for mental health said that management was afraid that psychosocial support could lead to absence due to illness and undermine morale, and that this could have been the reason behind their response.

Comments: The disaster plans were followed, the emergency ward resources were strengthened, operations were postponed and the triage function was activated at both hospitals. At Gregorio Marañón hospital the psychological aspect had not been considered in the disaster plan. Despite this, the decision to solve the care of relatives by gathering them in a large assembly hall was quickly reached – a solution which, incidentally, was also used at the 12 October hospital. This initiative was the result of one person’s decisiveness, and should not be taken as justification for not preparing for the problem of caring for the relatives of victims in extreme situations.

The confidentiality aspect with regard to issuing information proved to be problematic; laws concerning confidentiality were not followed, and most likely the problems surrounding both suitable locations for relatives and the need for information were underestimated.

From a Swedish perspective it is crucial that psychological disaster management (PDM) is trained for and that premises and a model for issuing information to relatives are taken into account in the disaster plans. It is also important to consider how to distribute accurate information should the phone systems become overloaded. Focussed support for the most seriously affected groups should be planned for, and special attention paid to children and young people, bearing in mind the experiences gained from Madrid The care of operative personnel is often neglected, and if some groups see the need for help as a weakness as opposed to a need, then this should also be borne in mind.


The process of identifying the dead went very quickly. By March 12, twenty-four hours after the event, 155 of the total 191 dead had been identified and their relatives notified. Hotel rooms were booked for the relatives that could not be informed that quickly, and a special reception providing psychosocial support was opened.

On March 13, the continued work of identification was moved to a cemetery with a cold room. Here too psychosocial support was available to relatives who received the information that a family member had been identified among the dead. The work was finalised on March 15 once the identification of all the bodies had been completed.

Comments: The identification process must be considered as having been extremely quick and effective. In disasters involving many dead, especially in hot climates, this work can become very complicated. This was clearly demonstrated during the tsunami disaster in south east Asia in 2004. This aspect should be taken into account in disaster plans.

The nature of terrorism

On March 11, ten bombs were detonated on four different trains. Four bombs never exploded. One reason for this was that the timer was set 12 hours too late. The police collected a bag containing one of the four unexploded bombs and took it to a police station without realising what the contents were – 10 kg of explosives and a detonator. In addition to the explosive device, the bag held a large amount of metal fragments, including nails, the purpose of which was to maximise injury to passengers close to the bombs.

Comments: The main purpose of a terrorist attack is to strike terror into the population and paralyse the society which has been targeted. For this reason terror bombings are often carried out at locations where many people are gathered. Experience has shown that suicide bombing, i.e. bombings carried out by people carrying a bomb about their person, are the most difficult to protect oneself against. Furthermore, the explosive devices are often constructed to cause maximum injury to people; so called anti-personnel bombs.

The emergency care of physical injuries does not depend on whether the injury has been caused by an accident or a terrorist attack. There is, however, every reason for management to pay extra attention to the security of personnel in the case of a terrorist attack. The work should be managed bearing in mind the risk of further explosions and the collapse of buildings. In Sweden there is a lack of recognition of the possibility of becoming the target of a terrorist attack, and that such could be planned to first detonate one bomb to draw in emergency service personnel and then set off further bombs to cause a maximum amount of damage to the various emergency service organisations.

It could be questioned whether the healthcare service in Stockholm could cope with a disaster similar in scale as that in Madrid, with almost 1,500 injured. In Stockholm there are seven hospitals with emergency wards, whereas in Madrid the injured were taken to 15 of a total of 24 different hospitals that were available.

One way of tackling this situation in Sweden would be to redistribute the injured – particularly those requiring intensive care – to other hospitals, either close to Stockholm (by road) or further a field (by air). The capacity to transport patients requiring intensive care is limited, however. Moreover the management and co-ordination of such a large scale transfer has never been trained for in Sweden.

Read the full Summary

Year: 2007
Article number: 2007-123-36
Format: PDF
Pages: 72
Language: Engelska
Price (VAT included): 0 kr


Susannah Sigurdsson
+46 (0)75 247 30 00