OBJECTIVES: To compare the interrater and intrarater reliability of the Alzheimer's Disease Assessment Scale (ADAS) with the Standardized Alzheimer's Disease Assessment Scale (SADAS). DESIGN: A randomized, double blind trial. Sixteen university students were randomized to administer either version of the instrument. Subjects were randomized to three assessments, at 2-week intervals, using the ADAS or the SADAS. Each subject's first and third tests were administered by the same rater, the second by a different rater. SETTING:A geriatric outpatient clinic in a university teaching hospital. PARTICIPANTS: Fifty-four patients with possible or probable Alzheimer's disease living in the community or in a long-term care facility. MEASUREMENTS: The primary outcome was the interrater reliability of total ADAS and SADAS scores. Secondary outcomes were ADAS and SADAS cognitive scores, noncognitive scores, duration of testing, and sample size estimates. RESULTS: The interrater reliability of the SADAS total score was significantly better than that of the ADAS (interrater ICC 0.93 SADAS vs 0.83 ADAS), and the interrater standard deviation of the total SADAS score was lower than that of the ADAS (38%, P < .05). The SADAS cognitive subscale inter and intrarater reliability, although higher than the ADAS, was not significantly different when used by different raters (interrater ICC 0.91 SADAS vs 0.90 ADAS; intrarater ICC 0.88 SADAS vs 0.86 ADAS). The SADAS noncognitive subscale was significantly more reliable than the ADAS (interrater ICC 0.89 SADAS vs 0.42 ADAS; intrarater ICC 0.87 SADAS vs 0.70 ADAS; P < or = .05) and had a lower standard deviation between raters (59%; P < .01) and within raters (40%; P < .05) compared with the ADAS. CONCLUSION: The improved reliability of the SADAS total score means that investigators can now use this score as a primary outcome measure, and important behavioral symptomatology can be included as a marker for treatment efficacy in AD. The smaller standard deviation of the SADAS means that clinical trials using the SADAS as a primary outcome will demonstrate differences, if present, with smaller sample sizes than with the ADAS.
RCT Entities:
OBJECTIVES: To compare the interrater and intrarater reliability of the Alzheimer's Disease Assessment Scale (ADAS) with the Standardized Alzheimer's Disease Assessment Scale (SADAS). DESIGN: A randomized, double blind trial. Sixteen university students were randomized to administer either version of the instrument. Subjects were randomized to three assessments, at 2-week intervals, using the ADAS or the SADAS. Each subject's first and third tests were administered by the same rater, the second by a different rater. SETTING: A geriatric outpatient clinic in a university teaching hospital. PARTICIPANTS: Fifty-four patients with possible or probable Alzheimer's disease living in the community or in a long-term care facility. MEASUREMENTS: The primary outcome was the interrater reliability of total ADAS and SADAS scores. Secondary outcomes were ADAS and SADAS cognitive scores, noncognitive scores, duration of testing, and sample size estimates. RESULTS: The interrater reliability of the SADAS total score was significantly better than that of the ADAS (interrater ICC 0.93 SADAS vs 0.83 ADAS), and the interrater standard deviation of the total SADAS score was lower than that of the ADAS (38%, P < .05). The SADAS cognitive subscale inter and intrarater reliability, although higher than the ADAS, was not significantly different when used by different raters (interrater ICC 0.91 SADAS vs 0.90 ADAS; intrarater ICC 0.88 SADAS vs 0.86 ADAS). The SADAS noncognitive subscale was significantly more reliable than the ADAS (interrater ICC 0.89 SADAS vs 0.42 ADAS; intrarater ICC 0.87 SADAS vs 0.70 ADAS; P < or = .05) and had a lower standard deviation between raters (59%; P < .01) and within raters (40%; P < .05) compared with the ADAS. CONCLUSION: The improved reliability of the SADAS total score means that investigators can now use this score as a primary outcome measure, and important behavioral symptomatology can be included as a marker for treatment efficacy in AD. The smaller standard deviation of the SADAS means that clinical trials using the SADAS as a primary outcome will demonstrate differences, if present, with smaller sample sizes than with the ADAS.