We examined the interrater reliability of the interRAI AC. An observational study was conducted on 3 geriatric wards. Two trained raters completed independently the interRAI AC between 24 and 48 h after admission. A sample of 100 patients (age 84.5 ± 5.6 years; 45% female) was analyzed. Interrater reliability was tested using observed agreement, kappa coefficients, and intraclass correlation coefficients. The overall kappa median value for nominal items was 0.82 (almost perfect). For items on ordinal level, both the overall weighted kappa median and the intraclass correlation coefficient median were 0.86 (almost perfect). According to conventional cut-offs for interpreting kappa statistics, reliability was almost perfect (K ≥ 0.81) for 60% of all items, substantial (0.60 < K ≤ 0.80) for 26%, moderate (0.41 < K ≤ 0.60) for 10%, and poor (K ≤ 0.40) for 4% of the items. The median observed agreement was 0.89. For 83% of the items, the observed agreement was greater than 0.80. This study showed substantial or almost perfect interrater reliability for 86% of the items. In addition to previous evidence, the current results suggest that the estimates of the interrater reliability of the interRAI AC are acceptable and provide preliminary evidence that the current version is appropriate for clinical application.
Copyright © 2011 Elsevier Ireland Ltd. All rights reserved.