Abstract
BACKGROUND: During the last year of life, persons with cancer should probably have similar care needs and costs, but studies suggest otherwise. METHODS: A study of direct medical costs (excluding costs for expensive prescription drugs) was performed based on registry data in Stockholm County, which covers 2.4 million inhabitants, for all deceased persons with cancer during 2015-2021. The data were mainly analyzed with the aid of multiple regression models, including Generalized Linear Models (GLMs). RESULTS: In a population of 20,431 deceased persons with cancer, the costs increased month by month (p < 0.0001). Higher costs were mainly associated with lower age (p < 0.0001), higher risk of frailty, as measured by the Hospital Frailty Risk Scale (p < 0.0001), and having a hematological malignancy. In a separate model, where those 5% with the highest costs were identified, these variables were strengthened. Sex and socio-economic groups on an area level had little or no significance. Systemic cancer treatments during the last month of life and acute hospitals as place of death had only a moderate impact on costs in adjusted models. CONCLUSIONS: Higher costs are mainly related to lower age, higher frailty risk and having a hematological malignancy, and the effects are both statistically and clinically significant despite the fact that expensive drugs were not included. On the other hand, the costs were mainly comparable in regard to sex or socio-economic factors, indicating equal care.