HOMEFOOD randomized trial—beneficial effects of 6-month nutrition therapy on body weight and physical function in older adults at risk for malnutrition after hospital discharge

Malnutrition is prevalent among older adults and is associated with adverse outcomes including altered body composition and reduced physical and mental function. Short hospital stays often preclude reversing malnutrition during admission, shifting emphasis to post-discharge care at home. Available post-discharge options include Meals on Wheels (MOW), oral nutritional supplements (ONS), and nutrition therapy by a clinical nutritionist/dietitian. Prior evidence suggests MOW can improve energy intake and meal frequency, but few studies assess functional outcomes. Meta-analyses of ONS show modest benefits (~1.0 kg weight gain). In Iceland, standard care offers access to one hot meal/day via MOW, which may be insufficient for frail older adults at nutritional risk. The study aims to test whether a 6-month multi-component intervention—nutrition therapy plus home-delivered energy- and protein-dense foods and ONS—improves body weight and physical function (SPPB) in community-dwelling older adults after hospital discharge.

- Observational studies suggest aging is not inevitably linked to malnutrition; adequate diet associates with lower mobility limitations and better quality of life. - Systematic review of MOW interventions in older adults showed significant improvements in energy intake and meals/day, but only 3 of 12 studies reported functional outcomes. - In Iceland, standard care post-discharge includes the option to order one hot meal/day (MOW), which may be inadequate for frail, nutritionally at-risk older adults. - Meta-analyses of ONS interventions show modest benefits for weight gain (~1.0 kg) and nutritional status improvements; concerns exist that long-term ONS may displace usual food intake. - Prior RCTs using dietitian home visits (without delivered foods) over shorter durations (~12 weeks) showed smaller endpoint weight differences (~1.4–1.8 kg) than anticipated with combined approaches. - Evidence on nutrition interventions improving physical function and muscle strength in older adults is mixed, with some studies showing benefits and others not.

Design: 6-month randomized controlled, assessor-blinded trial (HOMEFOOD) conducted in community-dwelling older adults (66–95 years) in the Reykjavik capital area (Jan 2019–Jul 2020). CONSORT guidelines followed; ethics approval obtained; registered at ClinicalTrials.gov (NCT03995303). Participants: Screened N=1003 at Landspitalinn University Hospital; eligible if ≥65 years, discharging home to independent living, at risk for malnutrition (Icelandic Nutrition Screening Tool score ≥3), and provided informed consent. Exclusions: dietary allergies/special diet, severe CKD (eGFR <30 mL/min/1.73 m²), active cancer treatment, tube feeding, inability to communicate with study team, MMSE ≤20, no functional kitchen at home, not community-dwelling, outside capital area, <65 years. Randomized N=106 (53 intervention, 53 control). Two deaths occurred during follow-up (one per group); 52/53 in intervention received allocated intervention. Randomization: Computer-generated random numbers (SPSS v26). Intervention: Nutrition therapy by a clinical nutritionist following the Nutrition Care Process (assessment, diagnosis, intervention, monitoring, evaluation). Schedule: five home visits (1 day post-discharge; at weeks 1, 3, 6, and 12) plus three interim telephone calls. Education on adequate energy and protein intake; identification and management of nutrition-related problems; involvement of caregivers when possible. Free weekly delivery of supplemental energy- and protein-rich foods: one hot meal/day and two in-between-meals/day (including ONS); education on storage, opening, and reheating provided. Control: Standard care—nutrition booklet on healthy eating in aging; encouraged to order MOW; no further dietetic counseling. Outcomes and assessments: Conducted at baseline (hospital) and endpoint (home). Assessors blinded to group for anthropometrics, physical function, muscle strength, and nutritional status. - Primary outcomes: Body weight and physical function (Short Physical Performance Battery, SPPB). Due to home setting, gait speed not assessed; SPPB score range 0–8 (balance and chair rise components only). Also asked: “Do you have difficulties walking?” (Yes/No). - Anthropometrics: Body weight (Seca 708), height from hospital records; BMI (kg/m²) and BMI categories: <23, 23–30, ≥30 kg/m². Body composition via handheld BIA (Omron HBF-306C). Calf and mid-arm circumferences measured by standard protocols. - Muscle strength: Dominant-hand grip strength via hydraulic dynamometer (two trials; max recorded). - Dietary intake: 24-hour dietary recall (24-HR) analyzed in ICEFOOD using Icelandic food composition databases (ISGEM). Additional endpoint food frequency questions (hot meals, food groups, liquids). - Nutritional status: Icelandic Nutrition Screening Tool (13 items; score 0–30; higher worse). Food-related adverse events tracked (diarrhea, nausea, constipation, stomach pain). Sample size: Based on prior data, n=44 per group to detect 1.8±3.0 kg between-group weight difference; n=45 per group to detect 1-point SPPB difference (SD 1.7). Targeted >50 per group to allow ~10% dropout. Statistical analysis: SPSS v26. Normality assessed by Kolmogorov–Smirnov. Baseline comparisons by t-test or Mann–Whitney U; categorical by chi-square. Intention-to-treat approach stated; endpoint calculations represent per-protocol (dropouts included at baseline but not endpoint). Due to slight sex imbalance, all multivariate endpoint analyses adjusted for sex; unadjusted shown in supplement. Linear mixed models for continuous endpoints (parameter estimate B = adjusted group difference). Logistic regression for categorical physical tasks and difficulty walking, adjusted for baseline and sex. Subgroup analyses: sex, marital status, BMI categories, and age tertiles; independent t-test or ANOVA with LSD for subgroup comparisons; interaction tested via general linear model. Significance P<0.05.

- Participants: 106 randomized (53 intervention, 53 control); very low attrition (one death per group); 52/53 intervention participants completed planned visits/calls; no intervention-related adverse events. - Baseline: Groups similar on most characteristics; more females in intervention (72% vs 53%); similar baseline energy intake (~1500 kcal/day) and protein (~75 g/day). - Dietary intake changes (endpoint means ± SD): - Energy intake: Control 731 ± 320 kcal vs Intervention 2412 ± 403 kcal (P<0.001); change approx −815 vs +919 kcal/day, respectively. - Protein: Control 31.2 ± 15.5 g vs Intervention 118.2 ± 34.3 g (P<0.001); protein per kg BW 0.4 ± 0.2 vs 1.5 ± 0.4 g/kg (P<0.001). - Carbohydrates: 77.2 ± 34.4 g vs 203.5 ± 43.0 g (P<0.001); Fat: 31.1 ± 18.3 g vs 122.0 ± 29.8 g (P<0.001); Fiber: 6.4 ± 4.2 g vs 11.2 ± 3.8 g (P<0.001). - ONS contributed 24% of total energy and 29% of protein at endpoint (~1.75 ONS servings/day) in intervention. - Body weight: - Within-group changes: Intervention +1.7 ± 2.5 kg (~2% BW) vs Control −3.5 ± 3.9 kg (~5% BW); 42/53 controls lost >1 kg; only 1/53 in intervention lost >1 kg. - Sex-adjusted between-group difference at endpoint: B = −5.121 kg (control vs intervention), 95% CI −6.381 to −3.860; i.e., intervention had 5.1 kg higher weight (P<0.001). - Body composition and anthropometrics: - Lean body mass: B = −4.181 kg (control vs intervention), 95% CI −5.647 to −2.715; P<0.001 (intervention higher). - BMI: B = −1.693 kg/m² (control vs intervention), P<0.001; Waist: B = −2.624 cm, P=0.007; Mid-arm: B = −2.185 cm, P<0.001; Calf: B = −1.266 cm, P=0.020; Body fat % difference NS (P=0.176). - Physical function: - SPPB score: B = −0.906 (control vs intervention), 95% CI −1.787 to −0.293; P=0.024 (higher in intervention). - Task improvements (OR vs control, adjusted): Semi-tandem OR 3.77 (1.45–9.80), P=0.007; Side-by-side OR 3.15 (0.98–10.10), P=0.052; Tandem OR 2.21 (0.95–5.10), P=0.065; Chair test OR 2.02 (0.80–8.55), P=0.137; Difficulties walking (deterioration): OR 0.34 (0.14–0.83), P=0.018. - Nutritional status: ISNST score higher (worse) in controls at endpoint: B = 2.226 (1.381–3.071), P<0.001. - Acceptability and behaviors: >94% of intervention participants liked the provided foods; higher frequency of hot meals, meat consumption, and liquid intake vs controls. - Subgroups: Intervention effects consistent across sex, marital status, and age tertiles; interaction significant for BMI categories (P=0.027) with greater weight gain in low BMI vs middle ≈ high BMI groups.

The multi-component intervention—individualized nutrition therapy by a clinical nutritionist combined with provision of energy- and protein-dense foods and ONS—effectively countered the typical post-discharge decline seen under standard care. Compared with controls who experienced substantial weight and lean mass loss and worsened nutritional status, the intervention group achieved increased energy and protein intake, gained weight predominantly as lean mass, and improved lower-extremity function (SPPB) and perceived walking ability. Handgrip strength did not change, reflecting mixed evidence for strength responses to nutrition-alone interventions. The magnitude of benefit (≈5.1 kg higher body weight at endpoint, ≈7% difference) exceeds effects reported in studies using single-modality approaches (dietary advice, ONS, or MOW alone), which typically show smaller gains (≈0.6–1.5 kg) and modest intake increases. The combination of frequent, individualized counseling (NCP framework), long intervention duration (6 months), and high-acceptability, protein- and energy-rich foods likely drove better adherence and outcomes. Public health implications are notable given the high prevalence of malnutrition among hospitalized older adults and the observed continued decline under standard care. The intervention demonstrates a scalable model to improve post-discharge nutritional and functional outcomes.

After hospital discharge, older adults receiving standard care commonly experience weight and muscle mass loss, reduced dietary intake, and deterioration in nutritional status. A 6-month program of individualized nutrition therapy delivered by a clinical nutritionist, combined with freely delivered energy- and protein-dense foods (including ONS), led to significant improvements in body weight (primarily lean mass), dietary intake, physical function (SPPB), and nutritional status, with consistent effects across most subgroups. Future research could explore broader implementation, cost-effectiveness, optimization of components (e.g., tailoring by BMI or functional status), and integration with physical activity interventions to enhance strength outcomes.

- Imbalance in sex distribution between groups at baseline, though analyses adjusted for sex and conclusions were consistent in adjusted and unadjusted analyses. - Not double-blinded; however, outcome assessors were blinded. - Home-based assessments limited measurement scope: SPPB lacked gait speed component; body composition relied on handheld BIA and circumference measures rather than gold-standard methods. - Modest sample size (N=106) may allow some baseline imbalances despite valid randomization. - Endpoint analyses effectively per-protocol for dropouts (included at baseline only), which may introduce bias if attrition differs by group (although attrition was minimal and similar).