Post-earthquake engineering and epidemiologic assessments are important for the development of injury prevention strategies. This paper describes mortality and its relationship to building collapse patterns and initial medical responses following the 1992 earthquake in Erzincan, Turkey.

The study consisted of: 1) background data collection and review; 2) design and implementation of a field survey; and 3) site inspection of building collapse patterns. The survey included: 1) national (n = 11) and local (n = 17) officials; 2) medical and search and rescue (SAR) workers (n = 38); and 3) a geographically stratified random sample of lay survivors (n = 105). The survey instruments were designed to gather information regarding location, injuries, initial actions and prior training of survivors and responders, and the location, injuries, and management of dead and dying victims. A case-control design was constructed to assess the relationship between mortality, location, and building collapse pattern.

There was extensive structural damage throughout the region, especially in the city where mid-rise, unreinforced masonry buildings (MUMBs) incorporating a “soft” first floor design (large store windows for commercial use) and one story adobe structures were most vulnerable to collapse. Of 526 people who died in the city, 87% (n = 456) were indoors at the time of the earthquake. Of these, 92% (n = 418) died in MUMBs. Of 54 witnessed deaths, 55% (n = 28) of victims died slowly, the majority of whom (n = 26) were pinned or trapped (p <0.05). Of 42 MUMB occupants identified through the survey, those who died (n = 25) were more likely to have been occupying the ground floor when compared with survivors (n = 28) (p <0.01). Official medical and search and rescue responders arrived after most deaths had occurred. Prior first-aid or rescue training of lay, uninjured survivors was associated with a higher likelihood of rescuing and resuscitating others (p <0.001).

During an earthquake, MUMBs with soft ground floor construction are highly lethal, especially for occupants on the ground floor, suggesting that this building type is inappropriate for areas of seismic risk. The vulnerability of MUMBs appears due to a lack of lateral force resistance as a result of the use of glass store fiont windows and the absence of shear walls. The prevalence of this building type in earthquake-prone regions needs to be investigated further. A large portion of victims dying in an earthquake die slowly at the scene of injury. Prior public first-aid and rescue training programs increase participation in rescue efforts in major earthquakes and may improve survival.