SHBC1396

C.Y.CHEONG¹, C. MALHOTRA², P.L.K. YAP¹

Khoo Teck Puat Hospital¹, Duke-NUS Graduate Medical School²

Measuring disease severity in persons with severe dementia (PWSDs) is essential for clinical care and research. Most instruments encounter issues, e.g. floor effects in characterizing PWSDs. We aimed to validate and evaluate the psychometric properties of the Bedford Alzheimer Nursing-Severity Scale (BANS) in community-dwelling PWSDs.

Panel study Investigating Status of Cognitively impaired Elderly in Singapore is a multi-centre prospective longitudinal study. 215 PWSDs (≥FAST staging 6c) were recruited (mean age 83.6±8.2). We evaluated BANS’ construct validity with exploratory factor analysis and correlation with other established measures, and predictive validity with multiple regression.

Factor analysis revealed a two-factor solution (variance 58.4%) with item-2 (sleep-wake cycle) not loading onto any factor. Dropping item-2 (BAN-6) revealed a single-factor solution (variance 40.2%) and  internal consistency reliability (Cronbach’s α) improved from 0.701 (BANS) to 0.782 (BANS-6). FAST did not correlate with the Cohen-Mansfield Agitation Inventory (CMAI) and Quality of Life in Late-Stage Dementia Scale (QUALID), but both BANS and BANS-6 did. After adjusted for age, sex and comorbidities, BANS-6 was significantly associated with pneumonia (β=2.09,95%CI=0.13-4.05), fever episode (β=1.24,95%CI=0.10-2.40) and oral antibiotics used (β=1.33,95%CI=0.11-2.55) within 4 months prior to the assessment, as well as  tube feeding (β=4.42,95%CI=2.62-6.22), pressure sore (β=2.51,95%CI=0.85-4.18), eating problem (β=3.27,95%CI=2.20-4.34),  and malnutrition (β=1.63,95%CI=0.24-3.02). BANS was not significantly associated with pneumonia, oral antibiotics, and pressure sore, and FAST was not significantly associated with any outcome.

The 6-item BANS(BANS-6) is psychometrically reliable and valid in assessing disease severity in PWSDs and outperformed FAST, a commonly used in dementia staging tool, in our population.