Repeatability of a Rasch model of the AQ20 over five assessments

TY - JOUR

T1 - Repeatability of a Rasch model of the AQ20 over five assessments

AU - Barley, Elizabeth A

AU - Jones, Paul W

PY - 2006/5

Y1 - 2006/5

N2 - Rasch analysis is now used widely to assess the measurement properties of health status questionnaires. This study tested the stability of the AQ20 - a dichotomous-response measure of health status in asthma, using parameters estimated by a Rasch model. One hundred forty-four asthma patients completed the AQ20 on five occasions over 3 months. At visit 1, two items showed significant misfit and were deleted. At each visit, the overall mean item-person and item-trait interaction statistics from the remaining 18 items (AQ18) were very similar. The repeatability of the item calibrations was excellent (intraclass correlation coefficient 0.95), despite the patients' health having changed (repeated-measures ANOVA: FEV(1) and AQ18 score p

AB - Rasch analysis is now used widely to assess the measurement properties of health status questionnaires. This study tested the stability of the AQ20 - a dichotomous-response measure of health status in asthma, using parameters estimated by a Rasch model. One hundred forty-four asthma patients completed the AQ20 on five occasions over 3 months. At visit 1, two items showed significant misfit and were deleted. At each visit, the overall mean item-person and item-trait interaction statistics from the remaining 18 items (AQ18) were very similar. The repeatability of the item calibrations was excellent (intraclass correlation coefficient 0.95), despite the patients' health having changed (repeated-measures ANOVA: FEV(1) and AQ18 score p

U2 - 10.1007/s11136-005-5466-z

DO - 10.1007/s11136-005-5466-z

M3 - Article

C2 - 16721640

SN - 0962-9343

VL - 15

SP - 801

EP - 809

JO - Quality of Life Research

JF - Quality of Life Research

IS - 5

ER -