OBJECTIVE: The Korean advance directive (K-AD) comprises a value statement, treatment directives, preferences for cardiopulmonary resuscitation (CPR), artificial ventilation, tube feeding, and hospice care, as well as a proxy appointment. The K-AD can facilitate a patient's decision making with respect to end-of-life (EoL) care. The present study aimed to examine the extent to which patient-caregiver dyads would use the K-AD and agree on EoL care decisions. METHODS: Using a descriptive study design, 81 cancer patients were invited to participate. The final sample consisted of 44 patient-caregiver dyads who completed survey questionnaires, including the K-AD. One patient did not complete all parts of the questionnaire, and 36 (44.4%) declined to participate. Content analysis was conducted to examine the K-AD value statements. Cohen's kappa coefficient was calculated to determine the degree of patient-caregiver dyadic agreement on K-AD treatment directives (Sudore & Fried, 2010). RESULTS: Our patient participants had the following cancer diagnoses: colorectal 29.5%, breast 29.5%, and liver/biliary tract cancers, 15.9%. Half of the sample had advanced-stage disease. Spouses (70.5%) or adult children (20.4%) were the primary caregivers, with perceived bonding rated as fair (31.8%) or good (65.9%). Rejection of the K-AD was mainly due to the difficulty involved in deciding on EoL care (50%). Comfort while dying was the most common theme expressed by patients (73.8%) and caregivers (66.7%). In terms of treatment directives, dyads advocated for hospice care (66.7%) and reduced support for aggressive treatments of CPR or artificial ventilation. The use of CPR (κ = 0.43, p = 0.004) and artificial ventilation (κ = 0.28, p = 0.046) showed significantly mild to moderate concordance among the dyads. Some 16 of the 21 dyads identified their spouses as a proxy, with others designating their adult children. SIGNIFICANCE OF RESULTS: The degree of patient-caregiver concordance on the K-AD seemed applicable, and achieved mild to moderate concordance. Our findings are exploratory but suggest the need for EoL discussions where patient-caregiver dyads are encouraged to participate in EoL care decision making.