Latent transition analysis of end-of-life care trajectories among cancer patients in South Korea: a nationwide cohort study

BACKGROUND: Despite the prolonged and often fragmented course of care before death among patients with cancer, end-of-life care research has largely relied on cross-sectional snapshots rather than the longitudinal trajectories needed to inform quality improvement. This study aimed to identify distinct end-of-life care trajectories among Korean cancer decedents. METHODS: This retrospective cohort study used a nationwide dataset of all 75,784 individuals who died from cancer in South Korea in 2016 and examined health care utilization trajectories during the final year of life. Health care utilization during the final year of life was summarized using 5 binary indicators: chemotherapy or radiotherapy; high-intensity treatment (intensive care unit admission, mechanical ventilation, cardiopulmonary resuscitation, and vasopressor use); acute-care hospital admission; long-term care hospital (LTCH) admission; and hospice use. These indicators were analyzed using latent transition analysis to identify end-of-life care states, transition probabilities between states, and the most frequent trajectories. RESULTS: Latent transition analysis identified 5 end-of-life care states: high anticancer therapy use state, high acute-care hospital admission state, high LTCH admission state, minimal inpatient care state, and high hospice use state. Near the end of life, two primary trajectories emerged, collectively accounting for 36.0% of all patients: transitioning from either minimal inpatient care or high anticancer therapy use state into the high acute-care hospital admission state. High state persistence was observed across subsequent intervals; 90.2% of individuals remained in the high LTCH admission state, and 99.1% remained in the high hospice use state. CONCLUSIONS AND RELEVANCE: These findings suggest that end-of-life anticancer therapy, LTCH admission, and hospice use are concurrently or sequentially linked to acute-care hospital utilization. The marked state persistence, coupled with the late-stage initiation and limited duration of hospice care, underscores the critical need for earlier, proactive palliative care integration and timely hospice referrals before high-utilization patterns become entrenched.