Background: A relevant number of stroke patients have swallowing disabilities; assessment of swallowing ability is an important procedure on a stroke unit to reduce complications. There is still doubt whether or not a trained nurse has a similar efficacy in the assessment of swallowing difficulties as does a specialized speech therapist (ST).
Methods: Our ST has trained any stroke unit nurse (SN) in assessment of swallowing function. A test using first thickened flavored water (custard consistency) and 5 ml water is done.
For a pilot study, during 6 weeks all stroke unit patients with cerebral ischemia were included. Patients with TIA, non-stroke diagnoses or altered consciousness (score 2-3 for item 1a of the NIHSS) were excluded. Swallowing function was evaluated independently by the patient’s SN and latest on the next working day by the ST. A 5 step scale was used, measuring from no food, to 2 different texture-modified diets, to normal food.
Cohen’s kappa statistic was calculated to evaluate congruence of the assessments. Univariate statistics were done to analyze difference between cohorts with equivalent and different grading.
Results: Overall 90 patients were included (53% male, 93% ischemic stroke, mean age 71 years). In 72 (80%) of the SN and ST scored identical (10% no food allowed, 11% special food preparation, 79% no problem). In 11 (12%) SN were more careful than the ST, and in 7 (8%) the SN assessed less swallowing-problems than the ST. Cohens’s kappa was 0.6067. In four of those patients the ST found relevant swallowing problems, which could have led to feeding abstention (2 patients) or special food preparation (2 patients). Comparing those 18 patients with differences in the assessments and those 72 without, presence of dysarthria was the only different parameter and was more present in the patients with varying rating (Mann Whitney U, p=0.0019)
In this unselected cohort of ischemic stroke patients, swallowing problems were detected in 32%, when using the ST’s assessment as gold standard. Inter-rater agreement with a trained SN was good. In a small percentage (8%) nurses had overseen swallowing problems and in 4% this was relevant. Presence of dysarthria was the only factor, which was different between equal and unequal graded patients.