Paramedics Cannot Determine Which Patients Require Transport

[Guest guest]

Abstract

Introduction. Reducing unnecessary ambulance transports may have operational

and economic benefits for emergency medical services (EMS) agencies and

receiving emergency departments. However, no consensus exists on the ability

of paramedics to accurately and safely identify patients who do not require

ambulance transport. Objective. This systematic review and meta-analysis

evaluated studies reporting U.S. paramedics' ability to determine medical

necessity of ambulance transport. Methods. PubMed, Cumulative Index to

Nursing and Allied Health Literature (CINAHL), and Cochrane Library

databases were searched using Cochrane Prehospital and Emergency Care Field

search terms combined with the Medical Subject Headings (MeSH) terms

³triage²; ³utilization review²; ³health services misuse²; ³severity of

illness index,² and ³trauma severity indices.² Two reviewers independently

evaluated each title to identify relevant studies; each abstract then

underwent independent review to identify studies requiring full appraisal.

Inclusion criteria were original research; emergency responses;

determinations of medical necessity by U.S. paramedics; and a reference

standard comparison. The primary outcome measure of interest was the

negative predictive value (NPV) of paramedic determinations. For studies

reporting sufficient data, agreement between paramedic and reference

standard determinations was measured using kappa; sensitivity, specificity,

and positive predictive value (PPV) were also calculated. Results. From

9,752 identified titles, 214 abstracts were evaluated, with 61 studies

selected for full review. Five studies met the inclusion criteria

(interrater reliability, kappa = 0.75). Reference standards included

physician opinion (n = 3), hospital admission (n = 1), and a composite of

physician opinion and patient clinical circumstances (n = 1). The NPV ranged

from 0.610 to 0.997. Results lacked homogeneity across studies;

meta-analysis using a random-effects model produced an aggregate NPV of

0.912 (95% confidence interval: 0.707-0.978). Only two studies reported

complete 2 2 data: kappa was 0.105 and 0.427; sensitivity was 0.992 and

0.841; specificity was 0.356 and 0.581; and PPV was 0.158 and 0.823.

Conclusion. The results of the few studies evaluating U.S. paramedic

determinations of medical necessity for ambulance transport vary

considerably, and only two studies report complete data. The aggregate NPV

of the paramedic determinations is 0.91, with a lower confidence limit of

0.71. These data do not support the practice of paramedics' determining

whether patients require ambulance transport. These findings have

implications for EMS systems, emergency departments, and third-party payers.

Prehospital Emergency Care 2009, Volume 13, Issue 4 pages 516 - 527