Article

OBJECTIVES: To construct a new vestibular screening tool to identify likely
vestibular disorders and guide referral of dizzy patients presenting to hospital
and to test the vestibular screening tool for construct and discriminative
validity and reliability of physiotherapy assessors. DESIGN: Methodologic study.
SETTING: Emergency and acute hospital wards of a metropolitan hospital.
PARTICIPANTS: Adults (N=114) presenting to hospital with dizziness (mean age,
67.36+/-14.88y; 57% women). INTERVENTION: Not applicable. MAIN OUTCOME MEASURES:
Three vestibular screening tools (3, 4, and 5 items) were investigated.
Physiotherapy vestibular diagnostic tests categorized patients as vestibular or
nonvestibular patients. Subsets of patients were assessed twice by 2
physiotherapists (n=20) and twice by the same physiotherapist (n=30). RESULTS:
Each of the vestibular screening tools had a good fit to the Rasch measurement
model. Factor analysis demonstrated individual items loaded across 1 factor,
confirming unidimensionality of the 3 vestibular screening tools, and Cronbach
alpha determined internal consistency. The 4-item vestibular screening tool had
the greatest area under the curve using receiver operator curve analysis (.894),
with highest sensitivity (83%) and specificity (84%) for identifying vestibular
disorders (cutoff value >/=4/8). Sensitivity of the 3- and 5-item versions was
lower than the 4-item vestibular screening tool (80%). The 4-item vestibular
screening tool scores showed high intrarater (kappa item scores, .831-1;
intraclass correlation coefficient [ICC] total, .988) and interrater (kappa item
scores, .578-.921; ICC total, .878) reliability. CONCLUSIONS: The 4-item
vestibular screening tool is a reliable, valid tool for screening dizzy patients
presenting to hospital, with unidimensional construct validity, high sensitivity,
and specificity for identifying likely vestibular disorders. The vestibular
screening tool could be used clinically to streamline referrals of dizzy patients
to vestibular clinics.
CI - Copyright (c) 2015 American Congress of Rehabilitation Medicine. Published by
Elsevier Inc. All rights reserved.

Langue : ANGLAIS