Study designs of evaluations included in the review
Inclusion criteria relating to the study design were not specified. Insufficient information on study design was provided.
Specific interventions included in the review
Studies of the commonly used version of the BI compared with at least two ratings for the same group of patients were eligible for inclusion (see Other Publications of Related Interest nos.1-2). Studies of other versions of the BI were excluded. The included studies used different types of assessors and methods of assessment. The assessors were nurses, researchers, occupational therapists, trained lay persons, physicians, care assistants and informal carers. The methods of assessment were face-to-face interview, interview by post and by phone, testing and observation. The studies used variable amounts of assessor training (from none to training sessions plus scoring guidelines). Some studies used proxy responses where necessary.
Participants included in the review
Studies of older people were eligible for inclusion. Studies that only included stroke patients or patients aged less than 65 years were excluded. The included studies were conducted in different settings: acute hospital, geriatric unit, rehabilitation in-patients, rehabilitation care home, psychogeriatric day centre, community and nursing home. Where reported, the age of the participants ranged from 18 to 103 years.
Outcomes assessed in the review
Studies that assessed reliability were eligible for inclusion. The included studies assessed reliability using the kappa statistic for individual BI items, intra-class correlation coefficients (ICC), analysis of variance (ANOVA), a method recommended by Bland and Altman (see Other Publications of Related Interest no.3) and Pearson's correlation coefficient. The included studies assessed inter-rater reliability using simultaneous observed testing and sequential testing by different assessors, compared different methods of administration, and compared assessments at different times using different methods of administration.
How were decisions on the relevance of primary studies made?
Two reviewers independently selected studies and resolved any discrepancies by discussion.