Skip to content

2012 Conference Presentation

DemandEvaluation United States

6 September 2012

Effects of obesity and smoking on the risk for nursing home use, 2002–2008

Wayne L. Anderson, RTI, United States
Joshua M. Wiener, RTI, United States
Galina Khatutsky, RTI, United States
M.S. Brian S. Armour, RTI, United States


Objective. To assess the effects of obesity and smoking on the risk of nursing home use in the United States. The goal is to provide information for developing policies to address unmet needs for health promotion for people with disabilities to promote community living.

Data and Methods: We used four waves of data (2002, 2004, 2006, and 2008) from the U. S. National Institute on Aging’s Health and Retirement Study (HRS) data set. The HRS is a nationally representative survey of older Americans living in the community providing data on health transitions that individuals undergo toward the end of their work lives and in the years that follow. Data on nursing home use was derived from the HRS linked to government health insurance claims containing data on short-stay nursing home use and the presence of chronic conditions from the US Centers for Medicare & Medicaid Services. Remaining data on obesity, smoking history, chronic conditions, exercise participation, functional limitations, cognitive impairment, supportive care, family structure, demographic, and socioeconomic characteristics were drawn from the HRS. Data on body mass index (BMI) was used to construct an overall measure and three separate levels of obesity. Smoking measures were created for whether respondents currently smoked and had ever smoked if not smoking at the time of the interview. The sample included approximately 17,000 HRS respondents including both younger and older people with disabilities having Medicare, Medicaid, both Medicare and Medicaid, private insurance, or uninsured. We estimated the likelihood (using Logistic regression) of overall nursing home use, and separately for short-stay, rehabilitation-oriented nursing home use vs. long-stay nursing home use. We also estimated a Cox proportional hazard model of nursing home use. The marginal effects of obesity and smoking were also derived from model estimates.

Results: Nursing home prevalence was approximately 5%, and obesity prevalence was 27%. Approximately half of the sample had ever smoked with 14% smoking at time of interview. In descriptive statistics, nursing home residents were less often obese than non-nursing home residents, with no differences in smoking rates across the two groups. In multivariate analyses, we found relatively small effects for being obese but larger effects for having smoked across most models tested. Policy Implications. Although the effects of obesity were of modest size, given the increasing rate of obesity in the United States, developing initiatives to improve health promotion efforts to address the risk of nursing home use resulting from obesity will be paramount to prevent an increase in institutionalization rates in coming years. Continued public health efforts to reduce smoking are also important.


Skip to toolbar