General earthquake injury statistics based only on statistics from hospital emergency departments tend to overestimate the number of people seeking hospital treatment for earthquake-related problems since they also include individuals who seek treatment for medical complaints that are not earthquake-related.  On the other hand, looking only at hospitals and the problems they treat will likely underestimate the total health impact of an earthquake since such information does not take into account other settings in which people seek and receive treatment.  These settings include (but are not limited to) community clinics, urgent care centers, and Red Cross and Salvation Army shelters.  Of course, it is also difficult to obtain documentation on those patients whose injuries are self-treated.  According to Durkin, injury statistics based solely on data collected from hospitals may account for only 40% of the number of injuries that actually occurred (e.g., the actual number of injuries in the 1989 Loma Prieta earthquake may have been as high as 9,500 instead of the officially reported 3,800) (31).