Real‐world data on home end‐of‐life care for older adults with cancer: A retrospective claims data analysis

Abstract Background Cancer incidence is expected to increase with population aging, making the availability of places for treating patients with terminal cancer a pressing issue. However, little is known about the actual state of home end‐of‐life care (HEC) in Japan. Objective The objective of this study was to examine the real‐world state of HEC for older adults with cancer. Methods The Yokohama Original Medical Database was used to identify the cohort. Data of target patients was extracted based on three criteria: age ≥65 years, malignant neoplasm diagnosis, and having a specific billing code of HEC. Multivariable linear and logistic regression models were used to evaluate the association between age groups and HEC services or outcome indexes. Results Overall, 1323 people (554 and 769 aged < 80 and ≥ 80 years, respectively; men, 59.2%) had planned to receive HEC. The < 80 years group had more frequent emergent home visits than the ≥ 80‐year group (P < 0.001), but the number of monthly home visits was similar between the two groups (P = 0.267). The rate of emergent admission was 5.9% in the ≥ 80‐year group, which was higher than that in the < 80‐year group (3.1%; P = 0.018). Conversely, the rates of central venous nutrition and opioid use were higher in the < 80‐year group than those in the ≥ 80‐year group. Conclusions This study reported patterns of use of HEC among older adults with cancer in the terminal stage. Our findings may provide the basis for providing HEC for older adults with cancer.

percentage of older individuals in Yokohama City will exceed 25% in 2025 and 30% in 2035. 2 The proportions of patients aged > 65 years who died from cancer (from 2014 to 2015) were 83.1% in Yokohama City 3 and 84.2% nationally according to data from the National Cancer Center. 2 With the growth in the aging population, the proportion of older adults expected to die from cancer is estimated to increase to 88.2% by 2025. 3 Thus, it is important to focus on cancer care for older adults.
Most patients with cancer in Japan have so far spent the last moments of their lives in hospitals. However, due to the abovementioned factors, it is anticipated that there will be bed shortages for end-oflife care of patients with terminal stage cancer, and meeting this increased demand for older adults with cancer is a vital social concern.
Therefore, a community-based integrated care system is being pro- In contrast to the Japanese situation, in Canada, which is known as a well-organized country in the implementation of palliative care, the location of death at home exceeds 60% in older adults. 4 In addition, in many studies, the effectiveness of end-of-life care has been proven through population-based studies. 4,5 The community-based care is recommended to achieve better palliative care. 6 Awareness of the current situation of home end-of-life care (HEC) is indispensable to understand future supply and demand balance.
Thus, we aimed to evaluate the current situation of HEC for older adults with cancer in Yokohama City using an original administrative database to achieve further dissemination of HEC. Unlike existing medical databases, such as registry records of specific diseases, this database was developed for policymaking at the local government level. It is expected to be the first step of evidence-based policymaking (EBPM) through the gathering of information on medical policy problems, which was recommended by The Center for Government Excellence at Johns Hopkins University. 7 2 | ME THODS

| Database and setting
Yokohama City is the most populous city in Japan and is governed by the local government for the Greater Tokyo Area, which includes the Tokyo metropolis. Yokohama City has a population of approximately 3.75 million (January 2021), and the population distribution is as fol- Hence, it has a strong representation of the older adults and is an especially reliable database for those older than 75 years.

| Statistical analyses and definition of variables
Patients' demographics, including age, sex, insurance type, cancer type, and institution type, were collected from the database. In this study, we divided the patients into the < 80 years and ≥ 80 years groups according to their age. The age of 80 years is commonly used as one of the criteria when doctors make a decision for the options of cancer treatment in actual clinical practice. 9 Cancer type was clas-  Emergent admission could not be directly confirmed due to the payment system. Cases with data on planned/emergent home medical care fee and hospitalization fee in the following month were referred to as cases of emergent hospitalization. Survival time at home was defined as the period from the first home medical care event to the month in which the additional fee for end-of-life care was charged.
Comparisons were made between the groups of patients aged < 80 years and ≥ 80 years. Multivariable linear and logistic regression model were used to evaluate associations between age groups and HEC services or outcome indexes. Age group, sex, and location of institution were included in these models. For the survival analysis, a log-rank test stratified by age group was used. In addition, we used a Cox proportional hazards model to estimate a hazard ratio (HR) with 95% confidence interval (CI) and evaluate factors associated with time to death after HEC introduction. Age group, sex, and location of institution were included in the Cox hazard model.
Insurance type was not included in the model because this type was strongly connected to the recipient's age. Chi-square test was used to compare baseline characteristics between age groups. The  Table 1.  Table 1). Clinics were largely responsible for patients receiving HEC (94.5%).

| Results of home end-of-life care
The clinical services and outcome index of HEC are shown in Table 2.
The average number of composite unscheduled HVs was 1.9 times per person-month.
The < 80 years group received more unscheduled HVs than the ≥ 80 years group (1.5 vs. 1.2 times/person-month, P < 0.001), especially outside the outpatient clinic (P = 0.001). The rate of emergent admission was 5.9% in the ≥ 80 years group, which was higher than that in the < 80 years group (3.1%; P = 0.018). Conversely, the rates of central venous nutrition and opioid use were higher in the < 80 years group than those in the ≥ 80 years group. The need for more than 3 days of HVs and/or nursing visits per week, the rate of death in the patient's home, and oxygen use were not significantly different between the two groups.
The maximum numbers of p-HVs, u-HVs, and e-HVs per month within 6 months from death are shown in Figure 1. The average number of home visits for all patients was 5.1 ± 2.6. There was no difference in the maximum number of home visits per month between the two groups (< 80 years group 4.9 months vs. ≥ 80 years group 5.1 months, P = 0.150).

| Survival time after HEC introduction
The time to death after HEC introduction is shown in Figure 2.

| DISCUSS ION
To promote the concept of EBPM, we analyzed an original and large local government-based medical database consisting of medical invoice data. We found that patients received HEC regardless of their cancer type. Furthermore, through our analysis, the necessary home medical care resources and monthly frequency of HVs were identified. This comprehensive database analysis with a high coverage of individuals aged ≥ 65 years showed that patients with terminal cancer older than 80 years were less dependent on home medical care than the patients aged younger than 80 years. To the best of our knowledge, this is the first such analysis performed using a large, local government-based medical administrative database in Japan.
One of the significant contributions of this study was that it provided high-value real-world data on survival time after HEC introduction, which is difficult to analyze through medical invoice data in the Japanese system wherein survival time cannot be directly obtained.
The YoMDB is strongly representative of the population of older adults because the insurance system in Japan focuses on this population. 8 Moreover, the YoMDB is a highly comprehensive database for those older than 65 years. This universality is useful for EBPM. We targeted patients aged ≥ 65 years in this study because the proportion of patients aged ≥ 65 years who died of cancer was > 83%. 3 Therefore, such a focused analysis was important from the viewpoint of advancing medical care in the near future for an aging society.
In the Japan HOspice and Palliative care Evaluation study, which was the first large nationwide study focusing on end-of-life care   This study has several limitations. First, given the nature of the receipt database (it is based on a fixed code for the practice performed), we could not confirm the practice or outcome that we would like to have directly obtained. This may have resulted in some information bias. Second, because the reason for emergency hospitalization could not be correctly identified, it was difficult to analyze in detail whether palliation was acute, whether it was for respite purposes, or whether there was a different underlying requirement. Third, the analyses of medical invoice databases are limited because Japanese receipt data are originally used for the calculation of medical treatment fee, and analysis of the patient's condition and severity of underlying disease or comorbidity using these data was not possible. Last, it could solve selection bias if we could analyze a combined dataset which includes all the terminal stage patients treated at the hospital, hospice, and home.

| CON CLUS IONS
This study reported patterns of use of HEC among older adults with cancer in the terminal stage. In areas where the population will continue to age further, such as Japan, the need for HEC for patients with cancer will definitely increase, and the development of social resources to meet this demand is an urgent social issue.
Our results are valuable as they can be used as the basis for providing HEC through a community-based integrated care system F I G U R E 2 Kaplan-Meier estimated curves of the survival time after the introduction of home end-of-life care (HEC). CI, confidence interval.