Intended for healthcare professionals


Safety and feasibility of prehospital thrombolysis carried out by paramedics

BMJ 2003; 327 doi: (Published 03 July 2003) Cite this as: BMJ 2003;327:27
  1. Phil Keeling, consultant cardiologist (phil.keeling{at},
  2. Debbie Hughes, clinical nurse specialist1,
  3. Linnie Price, research coordinator2,
  4. Steve Shaw, senior lecturer in statistics3,
  5. Andy Barton, RDSU coordinator1
  1. 1South Devon Healthcare Trust, Lawes Bridge, Torquay TQ2 7AA
  2. 2Research and Development Support Unit, Peninsula Medical School, Plymouth PL6 8BX
  3. 3Department of Mathematics and Statistics, University of Plymouth, Plymouth PL4 8AA
  1. Correspondence to: P Keeling
  • Accepted 23 June 2003


The benefits of thrombolysis in patients with acute myocardial infarction are time dependent, with a potential 48% reduction in mortality if treatment is received within an hour of onset of symptoms.1 The benefits diminish thereafter. However the “call to needle” standard set by the national service framework2 has been achieved for only a third of patients nationally because of lengthy ambulance journeys and delays in hospital. Prehospital thrombolysis may be the most feasible means of achieving timely reperfusion, and all published studies have shown the accuracy of diagnosis, patient selection, and administration of thrombolysis.3 The Joint Royal Colleges Ambulance Liaison Committee favours autonomous prehospital thrombolysis by paramedics as a long term objective.4 We tested the feasibility of physician assisted and autonomous models of paramedic prehospital thrombolysis.M

Summary statistics for time from event to response (minutes)

View this table:

Methods and results

All 64 paramedics serving the study hospital took part in the study. Of these, 48 (75%) received formal training in recording 12 lead electrocardiograms, diagnosing ST elevation acute myocardial infarction (STEMI), and assessing suitability for thrombolysis with a checklist of nationally agreed exclusion and inclusion criteria. Those who were not formally trained acquired the necessary skills from trained paramedics via a cascade system. During one year electrocardiograms of patients with category A (emergency) chest pain were transmitted to hospital by telemetry and reviewed by the duty emergency medicine physician. Paramedics recorded data on the diagnoses of physicians and paramedics, the paramedics' hypothetical decisions to treat, and timings of all events, although they did not actually administer the drug. The actual time to thrombolysis and a consultant's diagnosis from the transmitted electrocardiogram was used as the standard. We analysed data for sensitivity and specificity of diagnoses and potential reduction in call to needle times. Where appropriate we calculated 95% confidence intervals.

Of 1192 emergency admissions for chest pain, we had complete data for 861 (72%) regarding the paramedics' diagnosis of STEMI. We consider this sample to be representative because failures in data collection were due to random incidence of transmission and equipment and protocol failure. The number of “true” STEMIs (according to the consultant's diagnosis) was 118 (14%). Among the 660 cases for which both the paramedic's and the physician's diagnosis was available, the paramedics' sensitivity was 71% (84/118) and specificity 97%. McNemar's test showed that paramedics' sensitivity was lower than that of doctors (71% v 90%, P=0.001) but the specificity was higher (97% v 94%, P=0.001).

The median call to potential needle time (call for help to paramedic's decision) was 28 minutes (range 7-112 minutes), and for STEMI the median time saved was 48 minutes. The goal of call to needle time in less than 60 minutes was met in 95% of cases, and 22% of patients would have received thrombolysis within 60 minutes of onset of symptoms. At present only 3% do so. The table shows the summary statistics of times measured. We have included under each category all cases for which we have appropriate data but the inconsistencies of some databases and transmission and protocol failures have led to differing denominators.


This feasibility study of autonomous prehospital thrombolysis shows that paramedics can record and interpret 12 lead electrocardiograms and safely administer thrombolysis in the community. The potential average time saved was 48 minutes from the call for help to medication being administered. The physician assisted model was found to be unreliable because of technological failure and areas in which communication was hampered because mobile phones could not work, a problem observed elsewhere.5 We did not routinely collect data for these unanticipated events so their precise distribution cannot be reported. Autonomous paramedic prehospital thrombolysis seemed feasible and safe and was associated with improved call to needle times. Sensitivity may improve with experience and confidence.


We thank the paramedics who took part in the study. The study was supported by a grant from NHSE South West Research and Development Directorate.


  • Contributors PK was the principal investigator and was responsible for project design, data collection, advice on analysis, delivery of training, and drafting the paper. DH was responsible for project design, data collection, liaison with paramedics, delivery of training, and drafting the paper. LP was responsible for data collection, liaison with accident and emergency departments and paramedics, and drafting the report and paper. SS was responsible for project design, monitoring and analysis of data, and drafting the paper. AB was responsible for project design and management, advice on analysis, and drafting the paper and is guarantor for the study. Helen Myers and Dougie Williams (Westcountry Ambulance Service) and Paul Feasby (South Devon Healthcare Trust) ensured protocol compliance; Terry Phillips (Westcountry Ambulance Service) was responsible for delivery of training; and Margaret Somerville (Teignbridge PCT) and Ken Wenman (Westcountry Ambulance Service) advised on protocol.

  • Funding NHS Executive Research and Development directorate South West. The guarantor accepts full responsibility for the conduct of the study, had access to the data, and controlled the decision to publish.

  • Competing interests None declared.