Introduction: In our region Advanced Emergency Medical Technicians (AEMTs) respond to acutely ill or injured patients in rural areas. The AEMTs have been authorized to administer fentanyl intravenously in doses up to 2 μg/kg to selected groups of patients in pain. Higher doses can be allowed by a physician after a teleconference. We examined the effect of intravenous (IV) fentanyl treatment, expressed as pain reduction on a 10-point Numeric Rating Scale (NRS). Moreover we examined the occurrence of negative coincident events to assess whether it was safe to let non-medical staff administer potent opioids intravenously. Methods: Retrospectively we collected the case sheets for all patients treated with IV fentanyl by the AEMTs in 2005 and 2006. We excluded all patients where a physician had been directly involved in the prehospital treatment. We recorded the IV fentanyl dose, NRS-score before and after fentanyl administration, the use of naloxone, as well as all recorded incidents of nausea or vomiting, hypoxia, hypotension, bradycardia and bradypnea following treatment with IV fentanyl. Results: A total of 564 patients were treated with IV fentanyl by the AEMTs. Thirtyseven patients were excluded. The remaining 527 patients (mean age: 55 years) were treated solely by the AEMTs and received a median dose of 100 μg IV fentanyl (quartiles: 70 μg; 120 μg), which lead to a decrease in median NRS-score on 5 (quartiles: 4; 6). Negative coincident events were recorded in 25 (4.7 %) patients. Two of those became unconscious immediately following fentanyl administration. One was treated successfully with IV naloxone and regained consciousness promptly. The other regained consciousness before naloxone treatment was commenced. In the remaining 23 patients negative coincident events were considered minor, and were not treated with naloxone. Conclusions: Our results suggest that non-medical personnel safely can administer IV fentanyl to selected groups of patients with a satisfactory result in terms of a considerable reduction in pain score and an acceptable rate of negative coincident events.