PP16 Prehospital recognition and antibiotics for 999 patients with sepsis: feasibility study results

Background Sepsis is a common condition which kills approximately 44,000 people annually in the UK. Early recognition and management of sepsis has been shown to reduce mortality and improve outcomes. Paramedics frequently attend patients with sepsis, and are well placed to provide early diagnosis and treatment. We aimed to assess whether a multi-centre randomised trial to evaluate pre-hospital antibiotics was feasible. Methods Volunteer paramedics used scratchcards to allocate patients with ‘Red Flag’ Sepsis at random between experimental and control arms. The primary outcome was mortality at six months. We also measured: adverse events, costs, final diagnosis, length of stay in hospital, and quality of care; and collected qualitative data about acceptability to patients in interviews, and paramedics in focus groups. We pre-specified criteria for deciding whether to progress to a fully powered trial based on: recruitment of paramedics and patients; retrieval of outcome data; safety; acceptability; and diagnostic accuracy. Results Fifty-four paramedics completed their training and were issued scratchcards to randomly allocate patients to trial arms. Patients were recruited from 1.12.17 to 31.5.18. In total, 118 patients were randomly allocated to trial arms; four patients dissented to be included in the trial, leaving 114 patients to follow-up. Sixty-two patients (54%) were allocated to the intervention arm. The mean age of the control arm was 71.2 years (range 28–97); 33 (65%) control participants were female. In the intervention arm the mean age was 75.6 years (range 30–99) and 38 patients (61%) were female. Nine patients in the control group (18%) and 17 in the intervention group (28%) were already taking antibiotics at the time of their 999 call. Twenty-three questionnaires were received. No serious adverse events were reported. Conclusions Complete results will be presented at the time of the conference, when routine linked anonymised outcomes are available.