Randomised feasibility study of prehospital recognition and antibiotics for emergency patients with sepsis (PhRASe)

Severe sepsis is a time critical condition which is known to have a high mortality rate. Evidence suggests that early diagnosis and early administration of antibiotics can reduce morbidity and mortality from sepsis. The prehospital phase of emergency medical care may provide the earliest opportunity for identification of sepsis and delivery of life-saving treatment for patients. We aimed to assess the feasibility of (1) paramedics recognising and screening patients for severe sepsis, collecting blood cultures and administering intravenous antibiotics; and (2) trial methods in order to decide whether a fully-powered trial should be undertaken to determine safety and effectiveness of this intervention. Paramedics were trained in using a sepsis screening tool, aseptic blood culture collection and administration of intravenous antibiotics. If sepsis was suspected, paramedics randomly allocated patients to intervention or usual care using scratchcards. Patients were followed up at 90 days using linked anonymised data to capture length of hospital admission and mortality. We collected self-reported health-related quality of life at 90 days. We pre-specified criteria for deciding whether to progress to a fully-powered trial based on: recruitment of paramedics and patients; delivery of the intervention; retrieval of outcome data; safety; acceptability; and success of anonymised follow-up. Seventy-four of the 104 (71.2%) eligible paramedics agreed to take part and 54 completed their training (51.9%). Of 159 eligible patients, 146 (92%) were recognised as eligible by study paramedics, and 118 were randomised (74% of eligible patients, or 81% of those recognised as eligible). Four patients subsequently dissented to be included in the trial (3%), leaving 114 patients recruited to follow-up. All recruited patients were matched to routine data outcomes in the Secure Anonymised Information Linkage Databank. Ninety of the 114 (79%) recruited patients had sepsis or a likely bacterial infection recorded in ED. There was no evidence of any difference between groups in patient satisfaction, and no adverse reactions reported. There were no statistically significant differences between intervention and control groups in Serious Adverse Events (ICU admissions; deaths). This feasibility study met its pre-determined progression criteria; an application will therefore be prepared and submitted for funding for a fully-powered multi-centre randomised trial. Trial registration: ISRCTN36856873 sought 16th May 2017; https://doi.org/10.1186/ISRCTN36856873