Factors associated with low body mass index and weight loss in nursing home residents

BACKGROUND. Undernutrition in nursing home residents is a significant and possibly modifiable public health problem. We evaluated the hypothesis that some potentially modifiable factors are associated with resident undernutrition. METHODS. This study is a cross-sectional, secondary data analysis of 6,832 community nursing home residents sampled from 202 nursing homes in 7 states. Data were from the Minimum Data Set (MDS), and HCFA-mandated resident assessment instrument used in U.S. community nursing homes. Two dependent variables represented undernutrition: (a) low body mass index (BMI), defined as the lowest quartile BMI of the sample (19.42 kg/m2 and below); and (b) weight loss, an MDS measure defined as 5% decrease in weight in 30 days, or a 10% decrease in 180 days (9.9% of the sample). Independent variables included resident demographics, eating-related variables, variables measuring functional, cognitive, and affective statuses, and medical conditions. Separate logistic regression models were estimated for low BMI and weight loss to test multivariate associations. RESULTS. Poor oral intake, eating dependency, decubiti, and chewing problems increase the likelihood of both low BMI and weight loss. Female gender, age 85 or older, bedfast, and hip fracture increase the odds of low BMI only; depressed behaviors and two or more chronic diseases increase the odds of weight loss only. CONCLUSION. Undernutrition in nursing home residents is a multifactorial syndrome. Improved oral feeding methods and treatment of depression are potentially important ways to counteract undernutrition in nursing home residents by targeting reversible features.