INTRODUCTION: Oxygen for refractory dyspnea at the end of life is commonly prescribed, even when the criteria for long-term home oxygen therapy are not met. Is palliative oxygen less likely to be prescribed when a person lives alone at the end of life? METHODS: Three years of consecutively collected clinical data from a regional community palliative care program in Western Australia were used in univariate analyses utilizing chi(2) and analysis of variance, as appropriate. Multivariable logistic regression was used to identify characteristics of people for who oxygen has been prescribed. RESULTS: Of the study population (n = 5203), 9.8% (n = 507) did not have a caregiver. Oxygen was less likely to be prescribed for patients living alone (11.8% versus caregiver 20.6%; p < 0.001), those with cancer (18.8% versus noncancer 26.5%; p < 0.001), and older people (oxygen 68.1 years versus no oxygen 69.6 years; p = 0.005), and more likely to be prescribed for those with worse dyspnea (mean score 7/10, oxygen 7.6 versus no oxygen 4.2; p < 0.001). With severe dyspnea (= 7/10), people who lived alone had twice as many clinician visits before oxygen was prescribed (4.2 versus caregiver 1.7; p = 0.03). In multifactor analyses lung disease, dyspnea severity, and female gender predicted oxygen prescription, while not having a caregiver reduced the likelihood significantly (odds ration [OR] 0.51, 95% confidence interval [CI] 0.35-0.74; p < 0.001). DISCUSSION: Presence of a caregiver in the home is associated with palliative oxygen prescription having controlled for time in hospital. This study raises questions about the reasons for oxygen prescription, and the role caregivers may play in initiating requests for therapy.