Increased risk of rehospitalisation and death in older hospital patients at risk of malnutrition: A cohort study.

Objectives: The aim of the study was to investigate associations between risk of malnutrition and risk of rehospitalisation and death in older hospital patients, and whether the possible associations were modified by age, gender, comorbidity or Ambulatory Care Sensitive Conditions (ACSCs).

Design: Prospective cohort study.

Setting: Somatic hospital in Western Norway.

Participants: 9,768 hospital admissions for patients aged ≥65 years.

Measurements: Information on the risk of malnutrition was based on nutritional risk screening data from 34 point prevalence surveys conducted between 2008 and 2018. Risk of malnutrition was assessed using Nutritional Risk Screening 2002 (NRS 2002) during the initial hospital admission. Outcomes and possible effect modifiers were obtained from the hospital's patient administration systems. The short-term outcome was the length of initial hospital stay. Long-term outcomes included total number of days in hospital, number of hospital stays and risk of death within one year following nutritional risk screening. Statistical analysis involved negative binomial and Cox regression models with adjustment for age, sex and number of diagnoses at time of nutritional risk screening.

Results: Overall, 34% of the patients were classified as being at risk of malnutrition. A higher proportion of the initial admissions were related to ACSCs for patients at risk of malnutrition than for those not at risk. Risk of malnutrition was associated with longer initial hospital stay (adjusted hazard ratio (95% confidence intervals) 1.31 (1.25, 1.37)), more days in hospital (adjusted risk ratio 1.25 (1.18, 1.32)) and a higher risk of having more than two hospital admissions the year following nutritional risk screening (adjusted risk ratio 1.16 (1.07, 1.26)). Patients at risk of malnutrition also had an increased risk of death within one year (adjusted hazard ratio 2.45 (2.25, 2.67)). All associations were more pronounced in the '65-69' and '70-79' age groups compared to the 80+ years age group, and in patients with fewer than four diagnoses compared to patients with four or more diagnoses. No significant interaction was detected between sex and risk of malnutrition with regard to patient outcomes.

Conclusion: Older patients at risk of malnutrition have a higher risk of rehospitalisation and death during the first year after nutritional risk screening compared with those not at risk. Among patients at risk of malnutrition, the initial hospital admissions were more often due to ACSCs. The impact of the risk of malnutrition on outcomes appears stronger in patients aged 65-79 years and in patients with less comorbidity. These findings underline the importance of nutritional risk screening and subsequent nutritional support in all groups of older patients.