Geriatric/original research
Triage of Geriatric Patients in the Emergency Department: Validity and Survival With the Emergency Severity Index

Presented at the American College of Emergency Physicians Research Forum, September 2005, Washington, DC.
https://doi.org/10.1016/j.annemergmed.2006.04.011Get rights and content

Study objective

We evaluate the validity of the Emergency Severity Index (version 3) (ESI) triage algorithm in a geriatric emergency department (ED) population and determine the association between ESI categorization and survival.

Methods

Validity was assessed using a retrospective analysis of hospitalization, ED length of stay, and observed resource use compared with resource utilization estimated at triage by the ESI. Survival analysis was conducted for the same cohort. The study was exempted by the hospital institutional review board, and the requirement for informed consent was waived.

Results

During the investigation, 1,087 patients older than 65 years were registered in the ED. Six patients were identified as direct admissions, and 152 were found to be repeat visitors, leaving 929 subject visits for analysis. In this cohort, hospitalization was associated with ESI triage assignment (Kendall’s τ-b=−0.476; 95% confidence interval [CI] −0.524 to −0.425). The area under the receiver operating characteristic curve for the predictive ability of the ESI for hospitalization was 0.77 (95% CI 0.748 to 0.806). Length of stay was associated with ESI assignment (Kruskal-Wallis test, P=.000). The relationship between triage categorization and resource utilization was significant (Spearman’s correlation=−0.683; 95% CI −0.716 to −0.647). ESI categorization was associated with vital status at 1 year (Kaplan-Meier χ2 67.85; df=4; P=.0000).

Conclusion

When used to triage patients older than 65 years, the ESI algorithm demonstrates validity. Hospitalization, length of stay, resource utilization, and survival were all associated with ESI categorization in this cohort.

Introduction

The Emergency Severity Index (ESI) version 3 algorithm is a 5-level triage instrument that has been shown to demonstrate reliability and validity in a wide range of patient populations.1, 2, 3, 4, 5, 6, 7 The algorithm was designed to evaluate both patient acuity and anticipated resource utilization necessary to reach disposition.3 The ESI instrument stratifies emergency department (ED) patients into 5 mutually exclusive categories, with ESI category 1 assignment representing the most acute, high-resource-utilization patients (cardiac arrest) and ESI category 5 representing patients who are the least acute and resource intense (medication refill).7 EDs nationally have identified crowding as one of their top areas of concern. The number of ED visits continues to climb, with fewer departments available to handle the volume.8 Coupled with hospital inpatient capacity that has become more constrained, entire initiatives have developed focusing on crowding and its ED impact.9 A reliable and valid triage instrument to accurately characterize waiting patients becomes essential for patient safety when one considers the resultant increased length of stay of ED patients.

Section snippets

Importance

Despite the increasing size of the older population and evidence suggesting that elders more frequently utilize emergency services than their younger counterparts, little is known about the validity of many frequently used instruments in the geriatric population.10 Several instruments have been tested in the geriatric population to risk-stratify after an ED visit. The Triage Risk Screening Tool is an instrument that stratifies older ED patients who are discharged from the ED into risk groups

Study Design

This investigation used both a retrospective health records survey methodology and a survival analysis for a cohort of older patients presenting to the study ED between May 13, 2004, and June 12, 2004. The institutional review board at Maine Medical Center exempted the study and waived the requirement for written informed consent.

Setting and Selection of Participants

The Maine Medical Center ED is an academic referral center that houses an emergency medicine residency training program. The annual ED census at Maine Medical Center

Results

During the investigation, there were 1,088 patients aged 65 years or older registered in the ED. Six patients were found to be direct admissions to the hospital, 1 patient had an incomplete ED record, and 152 registrations were found to be repeated patient visits, leaving 929 visits for analysis. Table 1 displays demographic characteristics of the subjects. Interrater agreement between the data abstractors on 5 key data elements represented excellent agreement, with weighted κ scores ranging

Limitations

This study was conducted at a single tertiary care hospital in a small city serving an extensive suburban and rural geographic area. These results may be difficult to generalize to EDs serving dissimilar communities. In addition, the retrospective nature of the health records review methodology introduces the possibility of bias. We made diligent efforts to limit bias introduced by the retrospective methodology by reviewing multiple data sources (dictated attending physician notes, test

Discussion

In this cohort of ED patients aged 65 years or older, the third iteration of the ESI triage algorithm accurately predicted hospitalization, ED length of stay, and resource utilization. In addition, ESI triage categorization was strongly correlated to survival at 1 year after the index ED visit. The ESI provides an accurate snapshot of overall acuity in this vulnerable ED population. Coupled with other screening tools for the geriatric ED population, it could provide the basis for a complete

References (22)

  • D. Eitel et al.

    The Emergency Severity Index triage algorithm version 2 is reliable and valid

    Acad Emerg Med

    (2003)
  • Cited by (0)

    Author contributions: MRB and TDS conceived the study, designed the trial, supervised the conduct of the trial, participated in data collection, performed data analysis, and drafted the manuscript, and both authors contributed substantially to its revision. TDS undertook data management. MRB takes responsibility for the paper as a whole.

    Supervising editor: Knox H. Todd, MD, MPH

    Reprints not available from the authors.

    Funding and support: The authors report this study did not receive any outside funding or support.

    Publication dates: Available online June 9, 2006.

    View full text