Aim To quantify the difference between adjusted costs for home‐based palliative care and hospital‐based palliative care in terminally ill cancer patients. Methods We carried out a case–control study of home‐care patients (cases) who had died at home between January 2009 and December 2013, and hospital‐care patients (controls) who had died at a hospital between April 2008 and December 2013. Data on patient characteristics were obtained from insurance claims data and medical records. We identified the determinants of home care using a multivariate logistic regression analysis. Cox proportional hazards analysis was used to examine treatment duration in both types of care, and a generalized linear model was used to estimate the reduction in treatment costs associated with home care. Results The case and control groups comprised 48 and 99 patients, respectively. Home care was associated with one or more person(s) living with the patient (adjusted OR 6.54, 95% CI 1.18–36.05), required assistance for activities of daily living (adjusted OR 3.61, 95% CI 1.12–10.51), non‐use of oxygen inhalation therapy (adjusted OR 12.75, 95% CI 3.53–46.02), oral or suppository opioid use (adjusted OR 5.74, 95% CI 1.11–29.54) and transdermal patch opioid use (adjusted OR 8.30, 95% CI 1.97–34.93). The adjusted hazard ratio of home care for treatment duration was not significant (adjusted OR 0.95, 95% CI 0.59–1.53). However, home care was significantly associated with a reduction of $7523 (95% CI $7093–7991, P = 0.015) in treatment costs. Conclusions Despite similar treatment durations between the groups, treatment costs were substantially lower in the home‐care group. These findings might inform the policymaking process for improving the home‐care support system. Geriatr Gerontol Int 2017; ••: ••–••.