Premium
Validation of the Triage Risk Stratification Tool to Identify Older Persons at Risk for Hospital Admission and Returning to the Emergency Department
Author(s) -
Lee Jacques S.,
Schwindt Graeme,
Langevin Mara,
Moghabghab Rola,
Alibhai Shabbir M. H.,
Kiss Alex,
Naglie Gary
Publication year - 2008
Publication title -
journal of the american geriatrics society
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.992
H-Index - 232
eISSN - 1532-5415
pISSN - 0002-8614
DOI - 10.1111/j.1532-5415.2008.01959.x
Subject(s) - medicine , triage , emergency department , confidence interval , observational study , emergency medicine , prospective cohort study , clinical prediction rule , receiver operating characteristic , risk stratification , predictive value of tests , psychiatry
OBJECTIVES: To assess the predictive validity of the Triage Risk Stratification Tool (TRST) to identify return to the emergency department (ED) or hospitalization in a multicenter patient sample. DESIGN: Prospective, observational study with 1‐year follow‐up. SETTING: EDs of three hospitals in Toronto, Canada. PARTICIPANTS: Seven hundred eighty‐eight subjects aged 65 to 101 (mean age 76.6, 58.5% female) who presented to the ED and were discharged home from the ED. MEASUREMENTS: Trained clinical assessors completed the TRST on patients aged 65 and older during a 4‐week study period. Patients who subsequently returned to the ED or were admitted to the hospital were identified using hospital information systems and classified as experiencing the composite endpoint at 30, 120, and 365 days. RESULTS: The mean TRST score was 1.55 (range 0–5), and 147 (18.7%) patients experienced the composite endpoint of return to the ED or hospital admission by 30 days. The sensitivity of a TRST score of 2 or greater was 62%, (95% confidence interval (CI)=54–70%), specificity was 57% (95% CI=53–61%), and likelihood ratio was 1.44 (95% CI=1.23–1.66). The area under the curve was 0.61 using a cutoff score of 2. CONCLUSION: The TRST demonstrated only moderate predictive ability, and ideally, a better prediction rule should be sought. Future studies to develop better prediction rules should compare their performance with that of existing prediction rules, including the TRST and Identifying Seniors at Risk tool, and assess the effect of any new prediction rule on patient outcomes.