Identifying dysphagia in the intensive care unit: Validation of the Swedish version of the Gugging swallowing screen-Intensive care unit.

Abstract

Background: Dysphagia is independently associated with adverse outcomes in intensive care units (ICU). Early identification through dysphagia screening does not occur routinely, negatively impacting optimal patient management. This study aimed to validate the Swedish version of the Gugging Swallowing Screen-Intensive Care Unit (GUSS-IVA).

Methods: This is a prospective multicentre study of 56 adult ICU patients with endotracheal intubation exceeding 48 h at three hospitals in Sweden. The GUSS-ICU was translated into Swedish (GUSS-IVA) and used to screen all prolonged intubated patients (>48 h) once extubated. The GUSS-IVA screen was conducted by ICU nursing staff and then compared with a gold standard Flexible Endoscopic Evaluation of Swallowing (FEES) within 2 h of the GUSS-IVA screen. Fifty-one of 56 patients underwent FEES (where assessors were blinded to the GUSS-IVA screen results). Sensitivity and specificity were calculated, as was the area under the receiver operating characteristic curves (AUC) with 95% confidence intervals (CI). For inter-rater reliability, within 2 h of the initial screen, 29/56 patients were GUSS-IVA screened a second time by a nursing staff blinded to the first GUSS-IVA results.

Results: Among the 56 patients, 38 (67.9%) were identified as dysphagic using the GUSS-IVA screen. With FEES, 42 of 51 patients (82.4%) were diagnosed with dysphagia; of these, 16 (31.4%) were classified as aspirating. Compared to FEES, GUSS-IVA showed high sensitivity and specificity values (81% and 89%, respectively) with an AUC of 0.85 (95% CI: 0.71-0.95) and a positive predictive value of 97%. High convergent validity was obtained for GUSS-IVA compared with the Dysphagia Outcome Severity Scale (ɸ = 0.57, p < .001) and the Functional Oral Intake Scale (ɸ = 0.52, p < .001) and moderate validity with the Penetration-Aspiration Scale (ɸ = 0.30, p = .033). The inter-rater reliability showed moderate agreement (Cohen's kappa κ = 0.501, p = .006).

Conclusions: This study indicates that the Swedish GUSS-IVA is a valid and reliable screen to identify dysphagic ICU patients. Given the negative impact of dysphagia on short and long-term patient outcomes, the Swedish GUSS-IVA is recommended as an essential first step by nursing staff for early identification of dysphagia for further diagnostics and subsequent patient management.