Aim Retrospective case record reviews (RCRR) have been widely used to assess quality of care but evidence for their use in prehospital ambulance settings is limited. We aimed to review case records of potentially avoidable deaths related to ambulance care.

Method We identified patients who were transported to hospital or died using linked ambulance-hospital-mortality data from one UK ambulance service over 6 months in 2013. Death rates (within 3 days) for patient groups (based on age, dispatch code and urgency) were determined; 3 patients calling in-hours and 3 out-of-hours were selected from categories with the lowest death rates. Five reviewers (GP, nurse, 2 paramedics and medical health service manager) assessed anonymised patient records for quality of care and avoidable mortality.

Results We selected 29 linked records from 1 50 003 focussing on patients not transported to distinguish pre-hospital from hospital causes. Overall 8 cases out of 29 (27.6%) scored between 2.4 and 2.8 (1=Definitely avoidable, 2=Strong evidence of avoidability), 8 cases (27.6%) scored between 3.0 and 4.6 (3=Probably avoidable, 4=Possibly avoidable), and the remaining 13 cases (44.8%) between 4.0 and 5.8 (5=Slightly avoidable or 6=Definitely not avoidable). Variation between raters was satisfactory with ICC 0.84 (95% CI: 0.73 to 0.92). Common themes among cases with strong evidence of avoidability were symptoms or physical findings indicating a potentially serious condition and refusal by patients or their carers to be transported to hospital. RCRRs require linked ambulance, hospital and mortality data to ensure accurate assessment in light of the diagnosis and cause of death.

Conclusion Retrospective case record reviews (RCRR) have been widely used to assess quality of care but evidence for their use in prehospital ambulance settings is limited. We aimed to review case records of potentially avoidable deaths related to ambulance care.

Conflict of interest None

Funding National Institute for Health Research, UK.