RESUMO
BACKGROUND: Deficiencies in communication in health care are a common source of medical error. Preferred communication patterns are a component of resuscitation teaching. We audio-recorded resuscitations in a mixed paediatric medical and surgical ICU to describe communication. METHODS: In the intensive care unit, resuscitation events were prospectively audio-recorded by two trained observers (using handheld recorders). Recordings were transcribed and anonymised within 24h. We grouped utterances regarding the same subject matter from beginning (irrespective of response) as a communication epoch. For each epoch, we describe the initiator, audience and content of message. Teamwork behaviours were described using Anesthesia Nontechnical Skills framework (ANTS), a behavioural marker system for crisis-resource management. RESULTS: Consent rates from staff were 139/140 (99%) and parents were 67/92 (73%). We analysed 36min 57s of audio dialogue from 4 cardiac arrest events in 363h of prospective screening. There were 180 communication epochs (1 every 12s): 100 (56%) from the team-leader and 80 (44%) from non-team-leader(s). Team-leader epochs were to give or confirm orders or assert authority (61%), clarify patient history (14%) and provide clinical updates (25%). Non-team-leader epochs were more often directed to the team (65%) than the team-leader (35%). Audio-recordings provided information for 80% of the ANTS component elements with scores of 2-4. CONCLUSION: Communication epochs were frequent, most from the team-leader. We identified an 'outer loop' of communication between team members not including the team-leader, responsible for 44% of all communication events. We discuss difficulties in this research methodology. Future work includes exploring the process of the 'outer loop' by resuscitation team members to evaluate the optimal balance between single leader and team suggestions, the content of the outer loop discussions and in-event communication strategies to improve outcomes.
