OBJECTIVE: This study aimed to describe the nutritional status of geriatric home residents according to their place of dwelling and to identify institutional factors associated with higher rates of undernutrition. METHODS: All institutions (514) in the Aquitaine region were interviewed for staff ratio, nutritional procedures, staff training, and other procedures in the area of nutrition. A stratified random sample of 601 residents in a subsample of 42 institutions underwent Mini-Nutritional Assessment. RESULTS: The estimated prevalence of undernutrition was 19.1 (95% confidence interval [CI] 14.0-24.2), with a higher rate in long-term care (48.0%, 95% CI 15.9-80.2) than in nursing homes (14.5%, 95% CI 10.6-18.4, P < 0.0001). In univariate analyses the risk of undernutrition was higher in long-term care (P < 0.0001), in settings with better weighing equipment (P < 0.0001), with a higher staff ratio (P = 0.0001), and a higher rate of subjects needing help for eating (P < 0.0001) and was lower in settings with a higher rate of training in nutritional screening (P = 0.0001) and management (P < 0.0001). In nursing homes, each item of the Mini-Nutritional Assessment Short Form was independently predictive of undernutrition. In multivariate analyses in nursing homes only, better weighing equipment (adjusted odds ratio 2.34, 95% CI 1.39-4.12, P = 0.0017) and higher staff ratio (adjusted odds ratio 1.03, 95% CI 1.00-1.05, P = 0.0230) were associated with higher rates of undernutrition. CONCLUSION: Undernutrition in institutions was linked to the resident health problems, with little evidence in favor of the influence of institutional policies.