BackgroundTo investigate variations in emergency medical service (EMS) pre‐arrival cardiopulmonary resuscitation (CPR), including both bystander CPR without dispatch assistance and dispatch‐assisted CPR (DACPR).MethodsWe carried out an observational study by implementing EMS pre‐arrival CPR reports in three fire agencies. We included adult, non‐traumatic, and non‐EMS witnessed out‐of‐hospital cardiac arrests. This reporting system comprised the dispatch instruction process and bystander CPR quality based on evaluations by EMS crews who arrived on the scene. Bystander CPR was categorized as “ongoing CPR” if the bystander was performing CPR when the EMS reached the patient's side and “good‐quality CPR” if the CPR was performed proficiently. We compared the frequencies of ongoing and good‐quality CPR in the bystander CPR already started without dispatch assistance (CPR in progress) group and DACPR group.ResultsOf 688 out‐of‐hospital cardiac arrests, CPR was already started in 150 cases (CPR in progress group). Dispatcher CPR instruction was provided in 368 cases. Among these, callers started chest compressions in 162 cases (DACPR group). Ongoing CPR was performed in 220 cases and was more frequent in the DACPR group (128/162 [79.0%] versus 92/150 [61.3%], P < 0.001). Good‐quality CPR was more frequent in the CPR in progress group, but the difference was not statistically significant (36/92 [39.1%] versus 42/128 [29.0%], P = 0.888).ConclusionsOngoing CPR and good‐quality CPR were not frequent in EMS pre‐arrival CPR. Detailed analysis of dispatch instructions and bystander CPR can contribute to improvement in EMS pre‐arrival CPR.