Study design and setting
We performed a retrospective cohort study of HEMS dispatch to older trauma in the south-east of England between 1 July 2013 to 30 June 2019.
Air Ambulance Kent Surrey Sussex (AAKSS) serves a population of approximately 4.3 million, with a transient population of up-to 11 million. AAKSS attends approximately 1600 patients each year. Two doctor-paramedic teams deploy by helicopter or response car, one of which operates a 24-h day and the other an 18-h day. The HEMS team brings advanced clinical procedures to complement the scope of practice provided by a land based Critical Care Paramedic (CCP), to include: pre-hospital emergency anaesthesia (PHEA), advanced analgesia and sedation, blood product transfusion and surgical intervention (thoracostomy and thoracotomy). The service works alongside the regional ambulance service of South East Coast Ambulance Service (SECAmb).
A CCP and HEMS dispatcher evaluate and task the critical care resources across the region from the Emergency Operations Centre (EOC). The tasking algorithm was devised internally and is previously published . Activations are categorised as: immediate, interrogated or crew request. Immediate dispatch is triggered by pre-determined criteria. Interrogated dispatch is triggered where subsequent clinical information is reviewed, and HEMS dispatch agreed. Both immediate and interrogate dispatches are based on mechanism of injury (MOI), clinical condition of the patient and geographical location. A crew request can be activated by crews on scene (figure 3, suppl. file).
All older trauma patients (≥65 years) attended by HEMS with suspected traumatic injuries during the study period were included. Exclusion criteria comprised: patients < 65 years and patients presenting with suspected medical aetiology. Inter-facility transfers were excluded due to the unknown interventions delivered prior to HEMS arrival, and mutual aid requests excluded due to a variable, and unaccounted passage of time prior to AAKSS receiving the tasking.
An electronic record system (HEMSbase Medic One Systems, Ltd. UK) is used at AAKSS. The following data were retrieved from the electronic patient record: patient identification number, timings (112/999-time, dispatch interval). Patient characteristics (age, gender), mechanism of injury (assault [blunt/penetrating], fall (< 2 m [m], > 2 m), intentional self-harm, road traffic collisions and other (for example, crush injury)), anatomical site of injury (head, neck, thorax, abdomen, upper leg, upper arm), GCS, advanced pre-hospital interventions provided by the HEMS team, drugs administered, patient disposition (pronounced life extinct on scene, or transport to local hospital, Trauma Unit or MTC), and conveyance (carry, ground escort, or ground assist) by transport modality were retrieved.
Advanced pre-hospital interventions comprise those not performed by ground ambulance teams where a CCP is not present: pre-hospital emergency anaesthetic (PHEA), open finger thoracostomy, resuscitative thoracotomy, ultrasound sonography (USS), administration of prothrombin complex concentrate (PCC, beriplex®), insertion of intercostal chest drain (ICD), administration of intravenous (IV) antibiotics and/or antiviral drugs, administration of hypertonic saline 5%, advanced analgesia (fentanyl and ketamine), and pre-hospital transfusion therapy. Transfusion therapy consisted of packed red blood cells (PRBC) and freeze-dried plasma (FDP). PRBCs were available throughout the study period and FDP (Lyoplas) available from 3 April 2015.
Data extraction and eligibility of patients was performed by one of the authors (JG) and any inaccuracies and discrepancies were resolved by a second author (JB). These included miscategorisation of calls with regard to mechanism, and coding of variables.
This project met National Institute for Healthcare Research (NIHR, UK) criteria for service evaluation and formal ethical approval was therefore not required. The project was approved by the Research & Development Committee at AAKSS.
Descriptive statistics are given as mean [95% CI] or median [IQR]. Patients were stratified into three groups according to age: 65–74, 75–84, 85 and over. Comparisons across groups were made using Chi-square or Kruskal-Wallis tests where appropriate. Where statistical significance was found (set at a p-value < 0.05) Dunn’s post-hoc testing with Bonferroni correction was performed. The study applied Strengthening the Reporting of Observational Studies in Epidemiology (STROBE) Guidelines , with missing values reported. All statistical analyses were conducted using SPSS 26.0 (IBM).