Test-Retest Reliability and Minimal Detectable Change of the Test of Visual Perceptual Skills-Third Edition in Patients With Stroke.

OBJECTIVES: To examine the test-retest reliability, calculate minimal detectable change (MDC), and report internal consistency of the Test of Visual Perceptual Skills-Third Edition (TVPS-3) in patients with stroke. DESIGN: Repeated-measures design (at an interval of 2wk). SETTING: Medical center. PARTICIPANTS: Patients (N=50) with chronic stroke who completed the TVPS-3. INTERVENTIONS: Not applicable. MAIN OUTCOME MEASURES: TVPS-3 that contains 7 subscales, namely, visual discrimination, visual memory, spatial relations, form constancy, sequential memory, visual figure-ground, and visual closure. RESULTS: The intraclass correlation coefficient value of the overall scale was .92 and those of the 7 subscales were .53 to .82. The MDC values of the overall scale and the subscales were 18.1 and 5.4 to 7.1, respectively. The MDC% value of the overall scale was 16.2% (<30%), showing acceptable random measurement error. However, the MDC% values of the subscales were 33.7% to 44.1% (>30%), indicating substantial random measurement errors. The Cronbach α of the 7 subscales were .71 to .89, indicating good internal consistency. CONCLUSIONS: Our results showed that the overall scale of the TVPS-3 had satisfactory test-retest reliability. However, the subscales demonstrated insufficient test-retest reliability. Therefore, the subscales should be used cautiously to explain the test results over repeated assessments in patients with stroke.