LIMITATIONS
As with any retrospective database analysis, it was not possible to have input into variables included in the tool nor to ensure accuracy of data recording. Among variables not reported in the database were socioeconomic status (such as income category, educational level, and insurance status) and a measure of co-morbidities. All of these variables could have conceivably affected D2B times of patients in the database. Although over 1000 patients were included in the database, only 818 had D2B times; further, the number of patients in the six hospitals differed greatly (varying from 122 to 322). Given the large number of variables, it is possible that the study was under-powered to detect important differences among the hospitals.