Prevalence of Nutritional Risk at Admission in Internal Medicine Wards in Portugal: The Multicentre Cross-Sectional ANUMEDI Study

Introduction: Disease-related undernutrition is highly prevalent and requires timely intervention. However, identifying undernutrition often relies on physician judgment. As Internal Medicine wards are the backbone of the hospital setting, insight into the prevalence of nutritional risk in this population is essential. We aimed to determine the prevalence of nutritional risk in Internal Medicine wards, to identify its correlates, and to assess the agreement between the physicians' impression of nutritional risk and evaluation by Nutritional Risk Screening 2002. Material and Methods: A cross-sectional multicentre study was performed in Internal Medicine wards of 24 Portuguese hospitals during 2017. Data on demographics, previous hospital admissions, primary diagnosis, and Charlson comorbidity index score were collected. Nutritional risk at admission was assessed using Nutritional Risk Screening 2002. Agreement between physicians' impression of nutritional risk and Nutritional Risk Screening 2002 was tested by Cohen's kappa. Results: The study included 729 participants (mean age 74 +/- 14.6 years, 51% male). The main reason for admission was respiratory disease. Mean Charlson comorbidity index score was 5.8 +/- 2.8. Prevalence of nutritional risk was 51%. Nutritional risk was associated with admission during the previous year (odds ratio = 1.65, 95% confidence interval: 1.22 - 2.24), solid tumour with metastasis (odds ratio = 4.73, 95% confidence interval: 2.06 - 10.87), any tumour without metastasis (odds ratio = 2.04, 95% confidence interval:1.24 - 3.34), kidney disease (odds ratio = 1.83, 95% confidence interval: 1.21 - 2.75), peptic ulcer (odds ratio = 2.17, 95% confidence interval: 1.10 - 4.25), heart failure (odds ratio = 1.51, 95% confidence interval: 1.11 - 2.04), dementia (odds ratio = 3.02, 95% confidence interval: 1.96 - 4.64), and cerebrovascular disease (odds ratio = 1.62, 95% confidence interval: 1.12 - 2.35). Agreement between physicians' evaluation of nutritional status and Nutritional Risk Screening 2002 was weak (Cohen's kappa = 0.415, p < 0.001). Discussion: Prevalence of nutritional risk in the Internal Medicine population is very high. Admission during the previous year and multiple comorbidities increase the odds of being at-risk. Subjective physician evaluation is not appropriate for nutritional screening. Conclusion: The high prevalence of at-risk patients and poor subjective physician evaluation suggest the need to implement mandatory nutritional screening.