Trajectories of end-of-life medical and long-term care expenditures for older adults in Japan: retrospective longitudinal study using a large-scale linked database of medical and long-term care claims

BACKGROUND: An accurate understanding of the current state of end-of-life care is important for healthcare planning. The objectives of this study were to examine the trajectories of end-of-life medical and long-term care expenditures and associated factors. METHODS: This was a retrospective longitudinal study using a large-scale linked database of medical and long-term care claims—National Health Insurance, Advanced Elderly Medical Insurance, and long-term care insurance—covering Prefecture A in Japan. Patients aged ≥70 years who died between April 1, 2016, and March 31, 2017, were included (N = 16,084 patients; mean age = 85.1 ± 7.5 years; 7804 men (48.5%) and 8280 women (51.5%)). The outcome measures were medical expenditures (inpatient, outpatient, and prescription), long-term care expenditures, and total healthcare expenditures (the sum of medical and long-term care expenditures) during the 60 months before the date of death. We calculated each patient’s monthly medical and long-term care expenditures for 60 months before the date of death and applied group-based trajectory modeling to identify distinct trajectories. Factors associated with spending trajectories were examined via multinomial logistic regression analyses. Explanatory variables included age, sex, diseases, and the medical services used. RESULTS: We identified six distinct spending trajectories for the total healthcare expenditures: high persistent (45.6%), medium-to-high persistent (26.1%), early rise then high persistent (9.8%), late rise (6.4%), low persistent then very late rise (i.e., when spending starts increasing later than “late rise”; 6.4%), and progressive increase (5.7%). Factors associated with the high-persistent trajectory were chronic illnesses, various organ failures, neurodegenerative diseases, fractures, and tube feeding. The trajectory pattern of medical expenditures was similar to that of total healthcare expenditures; however, a different pattern was seen for long-term care expenditures. CONCLUSIONS: Regarding combined medical and long-term care spending of the last 5 years, most patients belonged to a pattern in which the healthcare expenditures remained high, and a combination of multiple factors contributed to these patterns. This finding can offer healthcare providers a longer-term perspective on end-of-life care. SUPPLEMENTARY INFORMATION: The online version contains supplementary material available at 10.1186/s12877-021-02215-9.