Background/objectives: A disease-specific nutrition screening tool (NST): the spinal nutrition screening tool (SNST) has been developed for use in patients with spinal cord injury (SCI) but its reliability and agreement with other published tools requires investigation. The aims of this study were to assess the prevalence of malnutrition risk in SCI patients and to confirm the diagnostic accuracy of the SNST.
Subjects/methods: Patients' baseline clinical data, anthropometric measurements and NST scores were assessed. The validity of the SNST was assessed by (i) comparing with a full dietetic assessment (criterion validity); (ii) comparison with a generic NST: malnutrition universal screening tool (MUST) (concurrent validity); and (iii) completion of an additional SNST to assess inter- and intra-rater reliability. Agreement was assessed using Cohen's κ-statistics.
Results: Using the SNST, the prevalence of malnutrition risk ranged from 22 to 64% on admission to four SCI centres. The SNST had substantial agreement with MUST (κ: 0.723, 95% confidence interval (CI): 0.607-0.839) and dietitian assessment (κ: 0.567, 95% CI: 0.434-0.699). The SNST had a moderate to substantial reliability (inter-rater reliability: κ: 0.5, 95% CI: 0.2-0.8; intra-rater reliability: κ: 0.64, 95% CI: 0.486-0.802). When compared with dietetic assessment, the SNST had a numerically lower specificity (76.1% vs 80.4%) and similar agreement to MUST (κ: 0.57 vs 0.58) but SNST showed a numerically higher sensitivity (85.7% vs 80.4%) and a numerically higher negative predictive value (92% vs 89.2%) than MUST.
Conclusions: This study shows that malnutrition is common in SCI patients. The SNST is an acceptable (valid and reliable) NST and may be a useful alternative to MUST in identifying SCI patients at risk of malnutrition.