UC Irvine

Background/objectives

Studies show that in nursing homes (NHs), the prevalence of moderate-to-severe obesity has doubled in the last decade and continues to increase. Obese residents are often complex and costly, and this increase in prevalence has come at a time when NHs struggle to decrease hospitalizations, particularly those that are potentially avoidable. This study examined the association between obesity and hospitalizations.

Design

We linked 2011-2014 national data using Medicare NH assessments, hospital claims, and the NH Compare.

Setting and participants

Individuals aged ≥65 years, newly admitted to NHs, who became long-term residents between July 1, 2011 and March 26, 2014. The analytical sample included 490,086 residents.

Methods

NH-originating hospitalization was the outcome; a categorical variable defined as no hospitalization, potentially avoidable hospitalization (PAH), and other hospitalization (non-PAH). The main independent variable was body mass index (BMI) defined as normal weight (30 >BMI ≥18.5 kg/m²), mildly obese (35 >BMI ≥30 kg/m²), or moderately-to-severely obese (BMI ≥35 kg/m²). Covariates included individual and NH characteristics. Multinomial models with NH random effects and state dummies were estimated.

Results

After adjusting for individual level covariates, the risk of non-PAH for the mildly and moderate/severely obese was not different from normal weight residents. But the risk of PAH remained significantly higher for the moderate/severely obese (relative risk ratio = 1.055; 95% confidence interval 1.018, 1.094). Several NH-level factors also influenced hospitalization risk.

Conclusions and implications

Obese residents are more likely to experience PAH but not non-PAH. Efforts to improve care for these residents may need to broadly consider the ability of NHs to commit additional resources to fully integrate care for this growing segment of the population.