Dementia Malnutrition Risk in Austrian Hospitals
In annual Austrian acute-care assessments, documented dementia was not associated with a different prevalence of malnutrition risk, whereas nutritional-care documentation and the clinical characteristics associated with co-occurring dementia and malnutrition risk differed across patient groups.
This single-study briefing examines malnutrition risk and recorded nutritional interventions among older hospitalised patients with and without documented dementia in Austrian acute care hospitals. The investigators used annual Nursing Quality Measurement 2.0 assessments from 2017 through 2025 to describe group patterns and to model factors associated with the co-occurrence of dementia and malnutrition risk. The study is relevant to nutrition because it distinguishes screening-defined risk, documented care processes, and observational associations rather than treating them as interchangeable measures of nutritional status or clinical benefit. pmid:42887093
Evidence
The reported design was a multicentre, cross-sectional secondary data analysis. The final analytic sample comprised 9,991 hospitalised patients aged 65 years or older who had complete data on malnutrition risk. Data came from Austrian acute care hospitals participating in the Nursing Quality Measurement 2.0 program; annual participation included hospitals and wards that could recur across measurement years. The study therefore represents repeated annual cross-sectional assessments, not longitudinal follow-up of one fixed patient cohort. pmid:42887093
Information was collected with a standardized questionnaire incorporating patient interviews, clinical assessments, and medical-record review. Malnutrition risk was assessed with the Malnutrition Universal Screening Tool, while dementia status was based on International Classification of Diseases, 10th Revision documentation. Investigators recorded 13 nutritional interventions as dichotomous variables. For the association analysis, they used univariable and multivariable binary logistic regression with co-occurring dementia and malnutrition risk as the dependent variable; patients with both conditions were compared with all other patients in the analytic sample. pmid:42887093
The primary prevalence comparison did not show a difference in malnutrition risk between patients with documented dementia and those without it: 23.4% versus 23.7%, respectively, with p = 0.840. This is a comparison of Malnutrition Universal Screening Tool-defined risk in the included hospital population, not a measure of diagnosed malnutrition, nutritional intake, or subsequent outcomes. The reported result describes the distribution of screening-defined risk by documented dementia status within these assessments. pmid:42887093
Among patients already identified as at risk of malnutrition, several documented interventions supporting oral intake were more frequent for those with dementia than for those without dementia. The reported categories included mealtime assistance, modification of the mealtime environment, texture-modified diets, enriched diets, and oral nutritional supplements. These comparisons describe whether interventions were documented in routine care; they do not establish the content, intensity, timing, or appropriateness of an intervention for an individual patient. pmid:42887093
In the multivariable model, co-occurring dementia and malnutrition risk was independently associated with female sex (odds ratio 1.56, confidence interval 1.07–2.29), higher age (odds ratio 1.07 per year, confidence interval 1.04–1.09), end-of-life condition (odds ratio 1.89, confidence interval 1.21–2.95), a higher number of medical diagnoses (odds ratio 1.37, confidence interval 1.24–1.50), and greater care dependency (odds ratio 0.95 per Care Dependency Scale point, confidence interval 0.94–0.96). These estimates concern the modeled co-occurrence outcome, rather than separate causal effects of these characteristics on dementia or malnutrition risk. pmid:42887093
The study also reported variation in the use of several nutritional interventions across the annual assessments. Because the design draws on participating hospitals and patient populations that could differ by year, the observed pattern is situated within a changing series of cross-sectional samples. The results thus provide a description of documentation and associations in the participating Austrian acute-care setting, using standardized collection procedures and the stated screening and record-based definitions. pmid:42887093
Analysis — Observed Care Patterns
The central pattern is not a higher observed prevalence of Malnutrition Universal Screening Tool-defined risk among patients with documented dementia, but a different pattern of documented nutritional care among at-risk patients. That distinction matters analytically: intervention frequency may reflect care dependency, diagnoses, end-of-life condition, ward context, recorded needs, or other factors represented imperfectly in the available data. The regression findings likewise identify an observed profile associated with the combined outcome, not a causal pathway from any listed characteristic to dementia or malnutrition risk. pmid:42887093
The comparison structure also shapes interpretation. Patients with both dementia and malnutrition risk were contrasted with every other included patient, a broad comparator that combines patients with neither condition and patients with either condition alone. Consequently, the multivariable estimates characterize the co-occurrence group relative to that heterogeneous remainder of the sample. Statistical significance in the reported group comparisons does not itself establish clinical significance, and documentation of a nutritional intervention is not evidence that the intervention changed a nutritional or patient outcome. pmid:42887093
Limitations
The retrospective cross-sectional design cannot establish causation. The analysis measured co-occurrence at annual assessment points and used documented dementia diagnoses, Malnutrition Universal Screening Tool assessments, and recorded intervention status. It cannot determine temporal ordering between the associated characteristics and the combined outcome, nor can it attribute intervention documentation to dementia status or to a nutritional effect. pmid:42887093
Selection is also material. The final sample was restricted to hospitalised people aged 65 years or older with complete malnutrition-risk data, while people who were not hospitalised, refused participation, were younger than 65 years, or had missing Malnutrition Universal Screening Tool data were excluded from the secondary analysis. Such exclusions and selective participation can limit how fully the analytic sample represents all potentially eligible hospital patients. pmid:42887093
Annual variation may partly reflect differences in the hospitals and patient populations participating in a given assessment year, as well as selective participation, rather than changes in nutritional care. Finally, the data indicate whether named interventions were documented, but do not permit conclusions about their appropriateness for individual patients. These boundaries mean the findings should remain scoped to observed records, screening-defined risk, and associations in the participating Austrian acute-care assessments. pmid:42887093