The purpose of this study is to evaluate the efficacy and feasibility of a multi-level self-management intervention to improve nutritional intake in a group of homebound older adults (HOAs) who are at especially high risk for undernutrition. The study will be guided by the theoretical approaches of the Ecological Model and Social Cognitive Theory and will use a prospective randomized controlled design to estimate whether individually tailored counseling focusing on social and behavioral aspects of eating results in increased caloric intake and improved nutrition-related health outcomes in a sample of 104 HOAs. The investigators hypothesize that intervention at these levels will improve caloric intake and indirectly improve health outcomes.

Inclusion Criteria:

1. at least 65 years of age
2. homebound (based upon Medicare's definition of homebound status)
3. able to communicate or have a caregiver who is able to communicate
4. living in a private residence
5. experiencing either an acute illness or chronic condition
6. not consuming enough calories to maintain body weight or weight loss of > 5% over past 6 months (if known)

Exclusion Criteria:

1. significantly cognitive impaired (MMSE < 24 if living alone or MMSE < 15 if living with caregiver) (if known)
2. terminally ill
3. any cancer diagnosis within the past five years (melanoma excluded)
4. end-stage renal disease
5. gastric or enteral tube-feedings
6. dependent on a ventilator