Critical outcomes to be included in the Core Outcome Set for nutritional intervention studies in older adults with malnutrition or at risk of malnutrition: a modified Delphi Study

Malnutrition in older adults is common across community, hospital, and long-term care settings and is associated with poor outcomes including diminished quality of life, functional decline, readmissions, mortality, and increased costs. Trials of nutritional interventions have used heterogeneous outcomes and measures, hindering comparison, synthesis, and evidence-based decision-making. The study aimed to develop a Core Outcome Set (COS)—a minimum agreed set of outcomes—to be measured and reported in all future nutritional intervention trials in older adults with malnutrition or at risk of malnutrition, to reduce heterogeneity, minimize selective reporting, and optimize research utility.

Background data show malnutrition prevalence ranging from 3–15% in the community, 18–29% in long-term care, and 22% in hospitals, with even higher proportions at risk, and documented adverse consequences for patients and healthcare systems. No COS existed for nutritional interventions in this population. The project followed COMET Initiative guidance and was registered in the COMET registry. A preceding scoping review identified commonly used outcomes in prior trials and informed the list for Delphi rating, supplemented by additional literature on patient-reported outcomes.

Design: Part of a multi-phase project under EuGMS Nutrition SIG following COMET guidelines and registered in COMET. Phases included a scoping review (Phase 1), modified two-round online Delphi surveys with validation by patient and public involvement (PPI) representatives (Phase 2), and a final online consensus meeting (Phase 3); later phases will select measurement instruments and focus on dissemination. Participants: Researchers and healthcare professionals with substantial experience in malnutrition in older adults were invited (criteria included ≥3 peer-reviewed publications as first/last author on the topic or relevant clinical experience). PPI representatives (older adults and caregivers) were involved for feedback and validation. Delphi Round 1 (July–December 2022): 93 experts rated 38 outcomes (e.g., malnutrition status, dietary intake, body weight/BMI, muscle strength, muscle mass, functional performance, functional limitations, QoL, appetite, etc.) on a 9-point Likert scale (1–3 not important; 4–6 important but not critical; 7–9 critical). Participants also specified primary setting (community, hospital, long-term care) and typical follow-up duration (≤12 or >12 weeks) for context. Consensus for inclusion: ≥75% rating critical (non-PROs) or ≥60% (PROs) and <15% rating not important. Consensus for exclusion required ≥75% not important (<15% critical) for non-PROs or ≥60% not important (<15% critical) for PROs. The survey order was randomized across four versions. Nine outcomes pre-judged as non-critical were put to agreement for exclusion; participants could propose additions. Delphi Round 2 (March–April 2023): 72 of the original participants (77.4% response) re-rated 29 outcomes without consensus in Round 1; physical activity was added based on feedback. Outcomes reaching consensus in Round 1 were also put to agreement for inclusion. Adverse events were not re-rated (considered mandatory in any trial); adherence was removed from the acceptability outcome for conceptual clarity. PPI validation: Five PPI representatives from the Netherlands, Türkiye, and Portugal listed outcomes important to them and indicated agreement with Delphi-derived outcomes in their own languages. Final consensus meeting (27 July 2023): 15 participants (steering group, Delphi participants, PPI, and industry) reviewed Delphi and PPI results and voted anonymously. Inclusion/exclusion threshold: ≥70% agreement. Initially, exclusion of undecided outcomes did not reach consensus (64%); after discussion, 80% agreed to exclude all undecided outcomes. All outcomes with prior Delphi consensus were included except frailty, which failed to reach ≥70% due to overlap with other outcomes. Data and ethics: Data stored securely at Universidade Nova de Lisboa; ethical approval 86/2022/CEFCM. Analyses in R v4.1.2; visualization with ggplot2 v3.4.2. Chi-square tested differences by setting and follow-up duration.

- Participation: 93 experts completed Round 1; 72 completed Round 2 (77.4% response). Majority were from Europe (≈72–75%), predominantly women, and working mainly in hospital settings; settings and demographics were similar across rounds. - Round 1 outcomes meeting inclusion criteria: malnutrition status (88% critical), dietary intake (83%), body weight/BMI (75%), muscle strength (82%), functional performance (85%); PROs meeting inclusion: functional limitations (72%), quality of life (80%), acceptability/adherence (79%). - Round 2 additional inclusions: muscle mass (82%), frailty (79%), and PRO appetite (64%). Near-universal agreement (≈99%) that previously included outcomes should be in the COS. - PPI feedback: prioritized well-being, appearance, mobility, fatigue, weight regain, strength, physical capacity/sports, and memory; generally agreed with Delphi results, with some neutral/negative votes on selected items. - Final consensus meeting: After discussion, 80% voted to exclude all undecided outcomes. Frailty did not reach the ≥70% inclusion threshold due to overlap concerns. Final COS includes 10 outcomes: malnutrition status; dietary intake; appetite; body weight or BMI; muscle strength; muscle mass; functional performance; functional limitations; quality of life; acceptability of the intervention. - Setting differences: Ratings were largely consistent across community, hospital, and long-term care; some hospital-specific outcomes (e.g., mortality, complications) were rated more critical in hospital but not enough to warrant setting-specific COS.

The study achieved its goal of defining a consensus-based Core Outcome Set for nutritional intervention trials in older adults with or at risk of malnutrition. Standardizing outcome reporting will enhance comparability and synthesis across studies, reduce selective reporting and resource waste, and support clinical decision-making. Although some outcomes (e.g., mortality, complications, healthcare costs) were valued in specific contexts, the final COS focuses on universally relevant domains. The COS is intended as a minimum set and does not preclude inclusion of additional outcomes tailored to specific research questions or settings (e.g., including mortality in hospital-based trials). The group emphasized mandatory reporting of adverse events and adherence (outside the COS), attention to hydration, and consideration of healthcare costs in larger trials for cost-effectiveness analyses. The next phase will identify valid, reliable, and feasible measurement instruments for each COS outcome.

A Core Outcome Set comprising 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—was established for trials in older adults with malnutrition or at risk of malnutrition. This COS should guide outcome selection in future studies and can inform routine care and observational research. Future work will determine standardized measurement instruments for each outcome and promote dissemination and implementation.

- Limited representation from long-term care experts may reduce applicability to that setting; however, identified priorities were similar across settings. - Overrepresentation of European participants (>2/3) may limit global generalizability, though stratified analyses showed few differences. - No older adults with malnutrition or at risk participated in the Delphi surveys due to practical barriers; PPI input was incorporated through other means and PRO thresholds were more lenient. - Did not reach the targeted Delphi sample size (aimed for 200 to ensure 50 per setting); 23% attrition from Round 1 to Round 2. Nonetheless, characteristics were similar across rounds, and methodology was rigorous (COMET registration and guidelines, published protocol).