In this observational study, we used anonymized national health insurance claims data (inpatient, outpatient, and dispensing pharmacy) from May 2013 to May 2021 in Hachioji City, located in the west of Tokyo. As of 2020, Hachioji City had approximately 580,000 residents covered by 35 hospitals, including two designated cancer hospitals. Our working group was commissioned to analyze medical care costs in Hachioji City as part of the city’s cancer screening project [12].

This study included female patients with breast cancer or C50 according to the International Classification of Diseases, Tenth Revision, between May 2013 and June 2016. We also included patients with codes for specific breast cancer treatment between August 2013 and June 2016. We created a study group comprising two breast surgery specialists accredited by the Japanese Breast Cancer Society, two clinical oncologists certified by the Japanese Society of Medical Oncology, two general clinical oncologists from the Japan Board of Cancer Therapy, and two oncologists with substantial experience in analyzing national health insurance claims data. This group extracted the procedure codes associated with certain breast cancer procedures from the Ministry of Health, Labour and Welfare and the codes for the anticancer agents insurance covered for treating breast cancer, as shown in Table 1. Our previous study identified patients with breast cancer from national health insurance claims data [13]. We used an algorithm that defined treatment using the aforementioned codes for specific breast cancer treatment: surgery, radiation, chemotherapy, antibody therapy, and hormone therapy (Table 1).

As for disease name, we omitted suspected cases and only used confirmed cases. Nevertheless, the disease name in national health insurance claims data does not accurately match the diagnosed disease name. For example, in some cases, the disease name can continue to be displayed even when the patient is not undergoing any breast cancer treatment or it can be simply entered for medical fee claims. Therefore, during the study period, the claims data could have three groups of patients with breast cancer as the disease name: the group with breast cancer that received treatment for it, the group with breast cancer that received no treatment for it (e.g., older patients and patients with multiple comorbidities), and the group without breast cancer. Hence, patients with breast cancer as the disease name and a code for the specific breast cancer treatment were determined to have breast cancer. However, we excluded patients with a code for specific breast cancer treatment between May and July 2013 in their claim because their treatment could have started before the study period. Given that the confirmation or modification of the disease name may take some time, we included patients with breast cancer as the disease name up to June 2016. As for the age limit, national health insurance does not cover people aged 75 years or older. Therefore, we included only those patients who had breast cancer and were aged less than 70 at the start of treatment, which allowed them to be followed up after 5 years.

We categorized patients with breast cancer into three groups based on the specific breast cancer treatment conducted during the study period: surgical treatment without axillary lymph node dissection (ALND; surgery without ALND group), other radical treatment with or without postoperative adjuvant chemotherapy (other radical treatment group), and palliative therapy (palliative group). The treatment in these groups corresponded to the respective treatment recommended for breast cancer classified as Stage 0/I, Stage II/III, and Stage IV according to the Classification of Breast Carcinoma by the Japanese Breast Cancer Society [14]. Radical treatment was a treatment that included the surgical resection of the primary lesion. Patients who underwent radical chemoradiotherapy were included in the other radical treatment group. Palliative treatment included chemotherapy for a more than a year, molecular targeted therapy, and treatment for distant metastases, including the radiation of metastatic lesions.

Patients with breast cancer who underwent tumorectomy, mastectomy, and other radical treatment (i.e., the surgery without ALND and other radical treatment groups) were defined as the curable group, while those in the palliative group were defined as the non-curable group. Patients with breast cancer were further classified according to their HER2 and hormone receptor status based on the specific breast cancer treatment they received during the study period and by referring to the diagnosis and pharmaceutical codes [15]. At least two experts from the working group reviewed individual claims data. Any discrepancies were resolved through expert discussions and we confirmed the accuracy of the classification based on the patient’s breast cancer treatment.

The primary outcome was the total medical care costs incurred by the curable and non-curable groups over the 5 years following the initial specific breast cancer treatment. Owing to the recording of multiple disease names within a claim, we could not accurately determine the costs associated with individual diseases [16]. Consequently, we aggregated the total medical care costs for managing breast cancer, which comprised the costs related to diagnostic procedures, perioperative management, handling of surgery-related complications, management of chemotherapy-induced side effects, and treatment for symptoms associated with cancer progression, particularly in the terminal phase. The median survival time reported for metastatic breast cancer is approximately 30 months [17]. Although breast cancer holds potential for long-term recurrence [18, 19], one of the indications for the treatment of hormone-positive and follow-up for HER2-positive breast cancer is five years [20, 21]. As such, the study period was set to five years. Total medical care costs were calculated by summing the claimed amounts each month, rounded to the nearest thousand.

Four secondary outcomes were identified: (1) the medical care costs of specific breast cancer treatment for the curable and non-curable groups over the five years; (2) medical care costs (at six-month intervals for the first year and then yearly for the subsequent four years) for the three groups; (3) medical care costs for different age groups (20–29, 30–39, 40–49, 50–59, and 60–69 years) for the three groups; and (4) medical care costs based on hormone and anti-HER2 therapies. The medical care costs of specific breast cancer treatment were determined using the national fee schedule for 2021 in Japan and by referring to the codes in Table 1. The incidence rate of breast cancer was calculated based on the number of patients identified with breast cancer by our algorithm or those identified in the claims data.

Continuous variables were described using means and standard deviations (SDs). Categorical variables were described using numbers. Medical care costs were calculated and displayed as mean (SD) and median (interquartile range, IQR). Patients with missing data were excluded from the analysis. These analyses were conducted using JMP Pro® 16.1.0 (SAS Institute Inc., Cary, NC, USA).