Is your unconscious patient in cardiac arrest? A New protocol for telephonic diagnosis by emergency medical call-takers: A national study

Background: Worldwide, call-taker recognition of out-of-hospital cardiac arrests (CA) suffers from poor accuracy, leading to missed opportunities for dispatcher-assisted cardiopulmonary resuscitation (DACPR) in CA patients and inappropriate DACPR in non-CA patients. Diagnostic protocols typically ask 2 questions in sequence: 'Is the patient conscious?' and 'Is the patient breathing normally?' As part of quality improvement efforts, our national emergency medical call centre changed the breathing question to an instruction for callers to place their hand onto the patient's abdomen to evaluate for the presence of breathing.

Methods: We performed a prospective before-and-after study of all unconscious cases from the national call centre database over a 31-day period in 2018. Cases were placed in 2 groups: 1) 'Before' group (standard protocol) where call-takers asked 'Is the patient breathing normally?' and 2) 'After' group (modified protocol) where callers were instructed to place their hand on the patient's abdomen. In an intention-to-treat analysis, the accuracy, sensitivity and specificity of both protocols for determining CA were compared.

Results: 1557 calls presented with unconsciousness, of which 513 cases were included. 231 cases were in the 'Before' group and 282 cases were in the 'After' group. The 'After' showed superior accuracy (84.4% vs 67.5%), sensitivity (75.0% vs 40.4%) and specificity (87.9% vs 75.4%) when compared to the standard protocol. Adherence in the 'Before' group to the standard protocol was 100%. However, adherence in the 'After' group to the modified protocol was 50.4%. Per protocol analysis comparing the modified protocol with the standard protocol showed vastly improved accuracy (96.5% vs 69.3%), sensitivity (94.1% vs 39.0%) and specificity (97.8% vs 77.2%) of the modified protocol. In patients with true cardiac arrest, the median time to 1st compression was 32.5 s longer in the modified protocol group when compared to the standard protocol group, approaching significance (199.5 s vs 167.0 s, p = 0.059). Median time to recognize CA was similar in both groups.

Conclusion: Dispatch assessment using the hand on abdomen method appeared feasible but uptake by dispatch staff was moderate. Diagnostic performance of this method should be verified in randomised trials.