Assessing adherence to the evidence base in the management of poststroke dysphagia

Objective: To evaluate the reliability and responsiveness to change of an audit tool to assess adherence to evidence of effectiveness in the speech and language therapy (SLT) management of poststroke dysphagia. Design: The tool was used to review SLT practice as part of a randomized study of different education strategies. Medical records were audited before and after delivery of the trial intervention. Setting: Seventeen SLT departments in the north-west of England participated in the study. Subjects: The assessment tool was used to assess the medical records of 753 patients before and 717 patients after delivery of the trial intervention across the 17 departments. A target of 10 records per department per month was sought, using systematic sampling with a random start. Analysis: Inter- and intra-rater reliability were explored, together with the tool's internal consistency and responsiveness to change. Results: The assessment tool had high face validity, although internal consistency was low (r_a = 0.37). Composite scores on the tool were however responsive to differences between SLT departments. Both inter- and intra-rater reliability ranged from 'substantial' to 'near perfect' across all items. Conclusions: The audit tool has high face validity and measurement reliability. The use of a composite adherence score should, however, proceed with caution as internal consistency is low.