BACKGROUND: Survival after in-hospital cardiac arrest (IHCA) differs considerably between hospitals. This study tries to determine whether this difference is due to patient selection because of the hospital level of care or to effective resuscitation management. METHODS: Prospectively collected data on management of in-hospital cardiac arrests from Sahlgrenska Hospital, a tertiary hospital in Gothenburg, Sweden (cohort one) and from five Finnish secondary hospitals (cohort two). A multiple logistic regression model was created for predicting survival to hospital discharge. RESULTS: A total of 954 cases from Sahlgrenska Hospital and 624 patients from the hospitals in Finland were included. The delay to defibrillation was longer at Sahlgrenska than at the five Finnish secondary hospitals (p=0.045). Significant predictors of survival were: (1) age below median (odds ratio [OR] 2.0, 95% confidence interval [CI] 1.5-2.8); (2) no diabetes (OR 1.9, CI 1.2-2.9); (3) arrests occurring during office hours (OR 1.5, CI 1.1-2.2); (4) witnessed cardiac arrest (OR 6.3, CI 2.6-15.3); (5) ventricular fibrillation or ventricular tachycardia as the initial rhythm (OR 4.9, CI 3.5-6.7); (6) location of the arrest (compared to arrests in general wards, GW): thoracic surgery and heart transplantation ward (OR 2.9, CI 1.5-5.9), interventional radiology (OR 4.8, CI 1.9-12.0) and other in-hospital locations (3.0, CI 1.6-5.7) and (7) hospital (compared to arrests at Sahlgrenska Hospital); arrests at Etelä-Karjala Central Hospital [CH] (OR 0.3, CI 0.1-0.7), Päijät-Hame CH (OR 0.3, CI 0.1-0.8) and Seinäjoki CH (OR 0.4, CI 0.3-0.7). CONCLUSION: The comparison of survival following IHCA between different hospitals is difficult, there seems to be undefined factors greatly associated with outcome. A great variability in survival within different hospital areas probably because of differences in patient selection, patient surveillance and resuscitation management was also noted. A locally implemented strong in-hospital chain of survival is probably the only way to improve outcome following IHCA.