Objective: Little attention has been given to assessing the importance of self-care and communication in the caregiving setting, especially caregiving for those who are terminally ill. The Caregiver Inventory (CGI), a measure of self-efficacy for caregiving that includes these two dimensions, was subjected to psychometric analyses. Method: One hundred and thirty-three primary caregivers completed the CGI; of those, 81 also completed the Perceived Stress Scale (PSS), Caregiver Burden Inventory (CBI), and a measure of caregiver tasks (ADLR-CG). Based on home visits, social workers also rated the caregiver tasks required (ADLR-SW). Exploratory Factor Analysis, as well as reliability and validity analyses were conducted. Results: Fit indices in M + I (V. 5.1) indicated a four factors solution: Managing Medical Information (alpha = 0.64), Caring for Care Recipient (alpha = 0.78), Caring for Oneself (alpha = 0.88), and Managing Difficult Interactions/Emotions (alpha = 0.76). The CGI was highly negatively related to stress (PSS, r = -0.54, p = 0.001) and burden (CBI, p = -0.37, p = 0.001); ADLR-CG was related to burden (r = 0.43, p = 0.001) but not stress. In regression and relative importance analyses, Care of Oneself and Managing Difficult Interactions/Emotions emerged as equal in terms of having the strongest and most robust negative relationships with stress and burden. Significance of Results: Results suggest that the CGI is a reliable and valid measure of self-efficacy for caregiving, and indicate the importance of self-efficacy for self-care and for managing difficult communication in successfully navigating the demands of caregiving for terminally ill persons.