Aim: By evaluating tape recordings of true cardiac arrest calls, to judge the dispatchers ability to (a) identify cases as suspected cardiac arrest (CA), (b) give the case the right priority, (c) identify CA cases suitable for dispatcher-assisted, telephone-guided cardiopulmonary resuscitation (T-CPR) and (d) accomplish T-CPR. Methods: Evaluation of 99 tape recordings of consecutive cases that had been admitted to the two city hospitals in Göteborg after out-of-hospital CA. Results: In 70% of the interviews, the dispatcher demonstrated impeccable behaviour with short, distinct questions, quickly resulting in a decision on how to handle the case. In 30%, serious criticism could be voiced as the dispatcher displayed very stressful behaviour, or omitted to ask important questions such as whether the patient was conscious and breathing. In 21%, the interviews indicated a clear opportunity to perform T-CPR. In another 10%, there was a possibility of performing T-CPR. Only in 8% was T-CPR actually accomplished. Conclusions: (1) In the majority of the interviews, the quality was very high, while in one-third, serious criticism could be voiced. (2) In our study, only one-third (95% confidence interval, 22–41) of CA cases were suitable for T-CPR, and T-CPR was performed in only 8% of the 99 cases. (3) To optimise the dispatcher ability to identify suspected CA and initiate T-CPR, both medical knowledge and practical training are needed, preferably with protocols for pre-arrival instructions.