How well do paramedics predict admission to the hospital? A prospective study

J Emerg Med. 2006 Jul;31(1):1-5. doi: 10.1016/j.jemermed.2005.08.007.


A study was designed to determine whether paramedics accurately predict which patients will require admission to the hospital, and in those requiring admission, whether they will need a ward bed or intensive care unit (ICU) monitoring. This prospective, cross-sectional study of consecutive Emergency Medical Service (EMS) transport patients was conducted at an urban city hospital. Paramedics were asked to predict if the patient they were transporting would require admission to the hospital, and if so, whether that patient would be admitted to a ward bed or require an ICU bed. Predictions were compared to actual patient disposition. During the study period, 1349 patients were transported to our hospital. Questionnaires were submitted in 985 cases (73%) and complete data were available for 952 (97%) of these patients. Paramedics predicted 202 (22%) patients would be admitted to the hospital, of whom 124 (61%) would go the ward and 78 (39%) would require intensive care. The actual overall admission rate was 21%, although the sensitivity of predicting any admission was 62% with a positive prediction value (PPV) of 59%. Further, the paramedics were able to predict admission to intensive care with a sensitivity of 68% and PPV of 50%. It is concluded that paramedics have very limited ability to predict whether transported patients require admission and the level of required care. In our EMS system, the prehospital diversion policies should not be based solely on paramedic determination.

MeSH terms

  • Allied Health Personnel / standards*
  • Cross-Sectional Studies
  • Decision Making*
  • Emergency Medical Services / standards*
  • Emergency Service, Hospital / statistics & numerical data*
  • Female
  • Health Services Research
  • Humans
  • Male
  • Predictive Value of Tests
  • Professional Competence
  • Prospective Studies
  • Sensitivity and Specificity
  • Surveys and Questionnaires
  • Transportation of Patients / statistics & numerical data*
  • Triage