Premium
Old and New Geriatric Screening Tools in a Belgian Emergency Department: A Diagnostic Accuracy Study
Author(s) -
Heeren Pieter,
Devriendt Els,
Wellens Nathalie I.H.,
Deschodt Mieke,
Flamaing Johan,
Sabbe Marc,
Milisen Koen
Publication year - 2020
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/jgs.16503
Subject(s) - medicine , emergency department , triage , emergency medicine , receiver operating characteristic , confidence interval , prospective cohort study , diagnostic accuracy , pediatrics , psychiatry
OBJECTIVES To compare the diagnostic accuracy of the Identification of Seniors at Risk, the Flemish version of Triage Risk Screening Tool, and the interRAI Emergency Department Screener for predicting prolonged emergency department (ED) length of stay, hospitalization (following index ED stay), and unplanned ED readmission at 30 and 90 days among older (aged ≥70 years) community‐dwelling adults admitted to the ED. DESIGN Single‐center, prospective, observation study. SETTING ED with embedded observation unit in University Hospitals Leuven (Belgium). PARTICIPANTS A total of 794 patients (median age = 80 years; 55% female) were included. MEASUREMENTS Study nurses collected data using semistructured interviews and patient record review during ED admission. Outcome data were collected with patient record review. RESULTS Hospitalization (following index ED stay) and unplanned ED readmission at 30 and 90 days occurred in 67% (527/787) of patients and in 12.2% (93/761) and 22.1% (168/761) of patients, respectively. For all outcomes at cutoff 2, the three screening tools had moderate to high sensitivity (range = 0.71‐0.90) combined with (very) low specificity (range = 0.14‐0.32) and low accuracy (range = 0.21‐0.67). At all cutoffs, likelihood ratios and interval likelihood ratios had no or small impact (range = 0.46‐3.95; zero was not included) on the posttest probability of the outcomes. For all outcomes, area under the receiver operating characteristics curve varied in the range of 0.49 to 0.62. CONCLUSION Diagnostic characteristics of all screening tools were comparable. None of the tools accurately predicted the outcomes as a stand‐alone index. Future studies should explore the clinical effectiveness and implementation aspects of ED‐specific minimum geriatric assessment and intervention strategies. J Am Geriatr Soc 68:1454‐1461, 2020.