Brief Report
Identifying high-risk patients for triage and resource allocation in the ED

https://doi.org/10.1016/j.ajem.2007.01.014Get rights and content

Abstract

Five-point triage assessment scales currently used in many emergency departments (EDs) across the country have been shown to be accurate and reliable. We have found the system to be highly predictive of outcome (hospital admission, intensive care unit/operating room admission, or death) at either extreme of the scale but much less predictive in the middle triage group. This is problematic because the middle triage acuity group is the largest, in our experience comprising almost half of all patients. Patients triaged to the 2 highest acuity categories (A and B) have admission/ED death rates of 76% and 43%, respectively. In contrast, the 2 lowest acuity categories (D and E) have admission/ED death rates of 1% or less. The middle category (C), however, has an overall admission/ED death rate of 10%, too high to be comfortable with prolonged delays in the ED evaluation of these patients. We studied this group to determine if easily obtainable clinical factors could identify higher-risk patients in this heterogeneous category. Data were obtained from a retrospective, cross-sectional study of all patients seen in 2001 at an urban academic hospital ED. The main outcome measure for multivariate logistic regression models was hospital admission among patients triaged as acuity C. Acuity C patients who were 65 years or older, presenting with weakness or dizziness, shortness of breath, abdominal pain, or a final diagnosis related group diagnosis of psychosis, were more likely to be admitted than patients originally triaged in category B. These findings suggest that a few easily obtainable clinical factors may significantly improve the accuracy of triage and resource allocation among patients assigned with a middle-acuity score.

Introduction

As emergency departments (EDs) face rising visit rates, increasing acuity, and longer lengths of stay [1], [2], it is becoming more important to have an accurate and quick triage system to ensure appropriate time to physician evaluation for patients who require urgent evaluation and treatment. Triage is the act of prioritizing patients based on limited history and vital signs for both the urgency with which they will receive emergency services and the types of service (critical care, fast-track) they receive. With many patients waiting in crowded EDs several hours before a physician evaluation [3], [4], it is increasingly necessary to have a triage system that has excellent sensitivity to identify patients who may potentially deteriorate in that time frame while having high enough specificity (identifying patients who may safely wait during that period) to realistically allow allocation of resources to higher-risk patients. Waiting times in the ED can range from 30 minutes to more than an hour, depending on patient acuity level; and combining waiting with treatment time can increase the time in the ED to 4 hours or more [5].

There are a number of triage systems in place. Most hospitals use either a 3-level system (emergent, urgent, and nonurgent) [6] or a 5-level system (resuscitation, emergent, urgent, less urgent, and nonurgent) [6], [7]. Previous studies found the 5-level system to be more reliable than the 3-level system [6], [8], [9], [10], [11] and better at predicting resource consumption, admission rates, length of stay, and mortality [12], [13]. The Canadian Emergency Department Triage and Acuity Scale (CTAS) is a 5-scale system ranging from resuscitation (level I) to nonurgent (level V). We have maintained the exact definitions for each group but designated the acuity categories by letter (A-E) rather than roman numerals to avoid confusion with other numerical scales (such as the diagnosis related group [DRG] severity index) and also to decrease confusion when reporting results (Table 1) [7], [11]. This scale was developed and endorsed by the Canadian Association of Emergency Physicians, the National Emergency Nurses Affiliation of Canada, and the L'Association des Medecins d'Urgence du Quebec.

Our hospital began using the CTAS system in 1999. We performed a 1-year review of all patient visits to our ED. As part of that project, the accuracy of triage was reviewed by comparing it to patient outcomes (death, hospital admission, intensive care unit (ICU)/operating room (OR) admission, transfer or home discharge, medical diagnosis, and DRG severity assessment). This review demonstrated that the middle level (level C, urgent) is the largest triage category. It is also the most problematic, with an overall admission rate that may be too high (10%) to safely delay the evaluation of these patients yet too low to take limited resources away from other potentially sicker patients by assuming all acuity C patients should be evaluated differently.

The objective of this study was to explore whether the addition of a small number of demographic and/or presenting characteristics to the triage protocol may better identify a group of higher-risk patients from among all those included in this midlevel (acuity C) category. Improving accuracy of triage of higher-risk patients would help better distinguish higher-risk from lower-risk patients in this large middle-acuity category—a distinction that could potentially improve the quality and efficiency of ED care.

Section snippets

Methods

The study site was a large urban academic hospital ED, a part of a 1200-bed hospital and level I trauma center. This was a retrospective, cross-sectional study of all patients seen from January 1 to December 31, 2001. Data were obtained from hospital clinical, admissions, and financial records and were merged using a unique identifier. Barnes-Jewish Hospital and the Human Studies Committee of Washington University in St Louis, Missouri approved the study.

Acuity C patients were the primary

Results

Acuity level data were available for 77 709 of 80 209 patient visits to the ED during 2001. Documented levels were (a) acuity A, 1446 (1.9%); (b) acuity B, 29 316 (37.7%); (c) acuity C, 36 727 (47.3%); (d) acuity D, 9019 (11.6%); (e) acuity E, 1201 (1.6%). Women comprised 60% of the sample. The overall mean patient age was 43 years but differed significantly by acuity level from a low of 32 (acuity E) to a high of 50 (acuity B) years (P < .001). The percentage of patients 65 years or older was

Discussion

This study used hospital admission and the combined end point of ICU/OR admission or death as markers of significant illness, requiring expedient emergency care. We recognize that not all patients who are admitted to the hospital require emergency intervention. Nevertheless, hospital admission is a useful marker for potentially serious disease. Our results confirm that the most accurate patient triage is at either extreme of the scale. Those triaged as acuity A had a combined admission/death

Acknowledgments

We thank the Washington University Center for Health Policy for support.

Cited by (43)

  • Clinical Decision Support Systems for Triage in the Emergency Department using Intelligent Systems: a Review

    2020, Artificial Intelligence in Medicine
    Citation Excerpt :

    Regarding the variables used to develop the models, in addition to vital signs parameters, the authors in [103] acknowledged that complaints of patients should be considered in order to understand general principles of triage abnormal diagnosis. In [40], the authors stressed that variables such as time to surgery for surgical conditions or time to antibiotics for infections could have been included in the study and yielded better performance. In [42], only routine ED data collected at the time of triage were used for developing the model.

  • CCMapper: An adaptive NLP-based free-text chief complaint mapping algorithm

    2019, Computers in Biology and Medicine
    Citation Excerpt :

    This free-text information is a crucial element for clinical decision-making, which affects several facets of a patient's ED visit such as nursing focus, initial medical assessment, prioritizing the treatment, determination of patient flow through the ED, and generation of different hypotheses for the final diagnosis [1–3]. From a health system standpoint, early information about the arriving ED patients can lead to improvements in ED resource allocation and prediction of downstream demand (e.g., admission rates) [4]. Natural Language Processing (NLP) approaches have been used to transform free-text healthcare information (e.g., clinical notes and chief complaints) into structured data [5–8].

  • Predicting hospital admissions to reduce emergency department boarding

    2016, International Journal of Production Economics
    Citation Excerpt :

    Study objectives included identifying important factors in admission (Considine et al., 2011), identifying high-risk patients for admission (Ruger et al., 2007), developing hospital admission prediction models (Leegon et al., 2006; Li et al., 2009), and estimating the total number of ED-to-inpatient-unit admissions (Peck et al., 2012). The most common methods used in these studies were LR (Sadeghi et al., 2006; Ruger et al., 2007; Li et al., 2009; Sun et al., 2011), and Bayesian Networks (Leegon et al., 2005; Sadeghi et al., 2006; Li et al., 2009; Peck et al. 2012). A brief review of these studies, including their settings, methods, and results, are as follows:

  • Reducing uncertainty in triaging mental health presentations: Examining triage decision-making

    2014, International Emergency Nursing
    Citation Excerpt :

    In support of this, a study of mental health presentations to a Canadian general hospital ED found that almost 50% of patients triaged as Level V (not urgent) required admission (Clarke et al., 2006) suggesting that these patients may have been under-triaged. Studies of inter-rater reliability have also found wide variation in responses often across three triage categories (Considine et al., 2000) with the most accurate patient triage occurring at the most urgent and non-urgent ends of the scale (Ruger et al., 2007). It should be noted however, that a recent systematic review of the scientific evidence on triage scales concluded that “most triage scales present insufficient scientific evidence for assessing inter-rater agreement” (Farrohknia et al., 2011).

  • Thoracic ultrasonography

    2014, Critical Care Clinics
View all citing articles on Scopus
View full text