Inter-rater reliability of the Barthel ADL Index: how does a researcher compare to a nurse?

Objectives: To investigate whether a nonclinical research assistant, using standardized scoring criteria, can reliably administer the Barthel Activiti es of Daily Living (ADL) Index in a sample of elderly inpatients. Design: Paired comparison of nurse and nonclinical research assistant Barth el Index assessments. Setting: Acute hospital wards from two hospitals in a UK Healthcare Trust, with a catchment population of approximately 224 000 people. Methods: A consecutive sample of 94 elderly patients with a variety of medi cal problems. Main outcome measures: Barthel ADL Index, Folstein Mini-Mental Status Exami nation. Results: Whilst the inter-rarer reliability of the Barthel Index was within acceptable boundaries, two items out of ten had only fair agreement and lo w crude agreement (transfer and dressing) on Cohen's kappa scores. Conclusions: Depending on the differences observed in any particular contex t, the Barthel Index can be applied with reasonable reliability by nonclini cal staff applying the standardized scoring criteria. II should be noted, h owever, that the kappa coefficients between clinical and nonclinical assess ors tend to be lower than those found when comparing two clinically trained assessors in previous research.