Malnutrition is prevalent among older adults, significantly impacting quality of life, functional abilities, and healthcare costs. Inconsistencies in outcome measures across nutritional intervention trials hinder data comparison and resource optimization. To address this, a Core Outcome Set (COS) – a standardized minimum set of outcomes – is crucial for future research. This study, conducted within the European Geriatric Medicine Society (EuGMS), aims to establish such a COS for nutritional interventions in older adults with or at risk of malnutrition. The project comprises five phases: a scoping review (completed), two-round Delphi surveys, a final consensus meeting, selection of measurement methods, and dissemination. The current study details the Delphi surveys and consensus meeting, building upon a previous scoping review identifying commonly used outcomes in existing trials.

A scoping review (Phase 1) identified commonly used outcomes in previous nutritional intervention trials in older adults. This review informed the selection of outcomes for the Delphi surveys. The literature highlighted the significant negative consequences of malnutrition and the lack of standardized outcome measures in the field, which hampers the ability to compare the effectiveness of different interventions and leads to wasted resources. The need for a Core Outcome Set (COS) in this area was therefore deemed crucial to ensure consistency and comparability of research findings.

The study employed a two-round modified Delphi method. Phase 2 involved online surveys where researchers and healthcare professionals rated 38 outcomes (derived from the scoping review and additional literature searches) on a nine-point Likert scale ('not important' to 'critical'). Consensus for inclusion was set at ≥75% 'critical' (≥60% for patient-reported outcomes [PROs]) and <15% 'not important'. Round 1 involved 93 participants; Round 2 (77.4% response rate) included those from Round 1, incorporating feedback and re-rating outcomes without consensus. Patient and public involvement (PPI) representatives provided feedback before and after the Delphi rounds. A final online consensus meeting (Phase 3) with 15 participants (representing academia, healthcare, health policy, industry, and PPI) finalized the COS, using a 70% agreement threshold for inclusion.

The first Delphi round identified five non-PROs (malnutrition status, dietary intake, body weight/BMI, muscle strength, functional performance) and three PROs (functional limitations, quality of life, and intervention acceptability) that met the consensus criteria for inclusion. The second round further refined the list, incorporating feedback and addressing outcomes that did not reach consensus in round one. This round showed that almost all participants agreed on the inclusion of malnutrition status, body weight or BMI, functional performance, dietary intake, muscle strength, functional limitations, quality of life and acceptability of intervention. Muscle mass and frailty also reached consensus. Appetite met the criterion as a PRO. The final consensus meeting resulted in the selection of ten outcomes: malnutrition status, dietary intake, appetite, body weight or BMI, muscle strength, muscle mass, functional performance, functional limitations, quality of life, and acceptability of the intervention. There were minimal differences in outcome selection across settings (community, hospital, long-term care) and follow-up durations.

The established COS addresses the significant issue of inconsistent outcome reporting in nutritional intervention trials for older adults. By standardizing outcome measurement, the COS will facilitate the comparison of results across studies, allowing for more robust meta-analyses and evidence-based decision-making. The inclusion of both PROs and clinician-reported outcomes ensures a comprehensive assessment of the impact of interventions. While the study aimed for broader representation across settings, the relative underrepresentation of long-term care participants might limit the generalizability of the COS to this setting. However, the findings suggest considerable overlap in critical outcomes across settings. Future research should focus on developing setting-specific COSs, particularly for long-term care, and explore alternative methods for involving older adults directly in the Delphi process to overcome potential participation barriers.

This study successfully developed a COS comprising ten critical outcomes for nutritional intervention trials in older adults with or at risk of malnutrition. This standardized approach will enhance the quality and comparability of future research, optimizing resource utilization and informing clinical practice. Future work will focus on defining appropriate measurement methods for these outcomes.

The study had a limited representation from long-term care settings, potentially affecting the generalizability of the COS to this specific population. The overrepresentation of European participants might also limit generalizability to other geographical regions. The reliance on online surveys could have excluded older adults with cognitive or physical limitations or limited digital literacy. While PPI representatives were involved, direct participation from older adults with malnutrition in the Delphi rounds was limited.