• J Turner, J Nicholl, L Webber, H Cox, S Dixon, and D Yates.
• Medical Care Research Unit, School of Health and Related Research, University of Sheffield, UK.
• Health Technol Assess. 2000 Jan 1;4(31):1-57.

ResultsIn total 1309 patients were entered in the study: 699 (53.4%) were treated by paramedics operating protocol A and 610 (46.6%) were treated by paramedics operating protocol B. The randomisation worked well and there were no significant differences between treatment groups in incident characteristics, ambulance performance times, or patient or injury characteristics, apart from slightly more moderate or severe head injuries in the protocol A group (25.3% versus 20.3%). Protocol compliance was poor, with only 31% of protocol A patients receiving prehospital fluids and only 80% of protocol B patients not given fluids. The estimated odds ratio for being given prehospital fluids when treated by protocol A compared to protocol B was 2.09 (95% confidence interval (CI), 1.53 to 2.81).MortalityThere were 73 deaths within 6 months in the 699 patients in the protocol A group (10.4%), and 60/610 (9.8%) in the protocol B group. Thus the crude odds ratio for deaths when managed by protocol A was 1.07 (95% CI, 0.73 to 1.54). Excluding 26 patients whose cause of death may not have been trauma related, the odds ratio was 1.04 (95% CI, 0.69 to 1.55). Excluding 17 patients who may have been dead on arrival of the ambulance at the scene the odds ratio was 1.04 (95% CI, 0.70 to 1.53). Adjustment for age, injury severity and whether the patient was unconscious at the scene did not significantly alter these odds ratios.ComplicationsA total of 106 patients were identified from hospital notes as having at least one of eight major complications (adult respiratory distress syndrome, sepsis, acute renal failure, coagulopathy, wound infection, pneumonia, fat embolism or pulmonary embolism). The proportions with recorded complications were similar in the two groups: 60/699 (8.5%) in the protocol A group versus 46/610 (7.5%) in the protocol B group. HEALTH STATUS: A total of 878 questionnaires were sent to patients, and 559 (64%) usable replies were received. The response rate was similar in the two groups (62.9% versus 64.6%). In all eight dimensions of the SF-36 health status measure patients who had been managed by paramedics operating protocol A reported better average health than did patients in the protocol B group. However, none of the differences were at a level considered clinically important and only for one of the eight dimensions was the difference statistically significant. COMPOSITE OUTCOMES: No significant differences in outcome were found between the two protocol groups in terms of patients who either died or had serious complications, nor for patients who either died or had known poor health. SUBGROUPS: Subgroups of patients were defined on eight characteristics (ambulance service area, whether a doctor was on scene, paramedic-patient contact time, injury severity, whether taken to theatre for emergency surgery, type of injuries, type of area, and whether the patient was treated before or after protocol cross-over). There was no evidence of any difference in mortality rates or composite outcomes between any subgroups, or between protocols within any subgroup. Time to A&E department The analysis suggests that patients given fluids spent 12-13 minutes longer at the accident scene than did patients not given fluids. However, because only one-quarter of patients were given fluids, and the specific protocol used made little difference to this, average on-scene times were largely unaffected by protocols.CostsIn the prehospital and immediate-care phase (including A&E treatment), the mean costs of the protocol A and protocol B groups were ¿419 and ¿416, respectively. This small difference reflects two small and offsetting effects of protocol B: reduced on-scene time (p = 0.08) and increased use of blood in the A&E department (p = 0.03). There were no other statistically significant differences in costs, with the mean total costs being ¿2706 and ¿2678 in the protocol A and protocol B groups, respectively (p = 0.52). (ABSTRACT TRUNCA

16 keywords...

hide…