This was a retrospective, longitudinal, observational cohort study in patients with MM. Retrospective claims data from the Medical Data Vision (MDV) database were analyzed from 01 January 2003 to 31 May 2021. The MDV database, comprising standardized healthcare insurance claims data provided by hospitals in Japan, has been utilizing the Japanese Diagnosis and Procedure Combination (DPC) fixed-payment reimbursement system for over 36 million individuals since 2003; it covers about 30000 patients with MM. Under the DPC system, treatments such as basic hospital stays, tests, and diagnostic imaging are reimbursed with a comprehensive payment set for different diagnosis groups. Treatments such as surgery and anesthesia are reimbursed based on a fee-for-service system.

Adult patients aged ≥20 years with a diagnosis of MM between 2003 and 2020, as per data available in the MDV database, were considered for this study. MM diagnosis was defined as the presence of at least one record with a confirmed MM diagnosis code (International Classification of Diseases [ICD]-10 code C90.0 in combination with the Japanese disease codes (2030003, 8839397, 8840039, 8842090, 8847152, 8847175, 8847250, and 8847258). The index diagnosis date was defined as the date on which the patient had the first record of a confirmed MM diagnosis. The baseline period was 12 months before the index diagnosis date. The follow-up period was ≥60 days from the index diagnosis date; however, patients who died within these 60 days were followed for <60 days.

Patients must also have continuous care in the baseline and follow-up periods. Those with an SCT record were included in the analysis even if they had a baseline period of <12 months. Patients with evidence of MM, other primary cancer(s), or metastatic disease in the baseline period were excluded from the study.

Patients were categorized into two groups based on the time of their initial diagnosis: 1) between 2003 and 2015 and 2) between 2016 and 2020. These categorizations were also executed considering approval of several new therapeutic agents for treatment multiple myeloma in Japan in 2015. Patients were subsequently divided into subgroups based on SCT after the index date (conducted and not conducted) and the presence of renal impairment during the baseline period (Fig 1). This analysis didn’t have the review from the Institutional Review Board since it includes the anonymized claims data from secondary database.

The main objective of the study was to evaluate treatment patterns in patients with MM receiving routine clinical care in Japan using the MDV database. Data were reported for different sub-groups as described above. Baseline demographics and clinical characteristics (age, sex, year of MM diagnosis, comorbidities, etc.) were reported. The proportion of the treatment regimens for each line (1L to 6L) was assessed and reported for different subgroups.

Duration of therapy was defined as the time from the initiation to the last day of the line of treatment. A regimen was established with at least one date of administration of the first drug following its administration in the follow-up; the interval between the dates fell within the range of the grace periods set up for each drug, considering the days of prescription. Additional drugs for the regimen were identified if another drug was administered within 28 days of the drug’s first administration. The date of administration of the target drug regimen was prescribed at least once after the date of the first administration, and the interval was within the range of the grace period for the drug. Switching to a new regimen was identified if a drug that was not the target drug in the previous regimen was newly administered at least once after the first administration date, and the interval between the two dates fell within the grace period of the drug.

The line of therapy (LoT) was considered as ended in one of the following conditions: (1) initiation of a new regimen; (2) 14 days or longer for any drug considering the grace period; (3) death or the end of the data. Attrition rates, defined as the percentage of patients who dropped out of the previous line, were also assessed, and reported for each line of treatment. Time to inpatient death was assessed from the index date through the date of the death of any cause. Those patients whose data ended without the death record were treated as ’censored’ at the end of the follow-up. Considering death information was only available from hospitalized patients, time to inpatient death was calculated only for such patients. For other patients, the duration to the last observation date was calculated and was handled as the right-censored value.

For baseline demographics, clinical characteristics, and treatment pattern data, descriptive summaries were provided. For survival data, the summary statistics (minimum, 25^th percentile, median, 75^th percentile, maximum) for time to inpatient death were presented, along with the Kaplan-Meier plot.

A total of 38370 patients with MM were registered in the MDV database, and 6438 (16.78%) were considered eligible; 53.4% (n = 3441) of these were men. The median (minimum-maximum) age at the index diagnosis date was 72.0 (20.0–100.0) years; 43.36% of patients were ≥75 years of age at the index diagnosis date (Table 1). Of the eligible patients, 1472 (22.9%) had an index diagnosis year between 2003 and 2015, and 4966 (77.1%) had between 2016 and 2020 (Table 1 and Fig 1). A total of 1257 patients were included in the SCT-conducted group, and 5181 were in the SCT not-conducted group (Table 1 and Fig 1). In the SCT-conducted group, 632 (50.3%) patients were aged <60 years at the index date, 331 (26.3%) were aged 60–64 years, 261 (20.8%) were aged 65–69 years, and 33 (2.6%) were aged ≥70 years (Table 1). In the SCT not-conducted group, 3632 (70.1%) patients were aged ≥70 years at the index date. Overall, the mean (standard deviation [SD]) Charlson Comorbidity Index (CCI) score was 0.7 (1.35), with 0.1 (0.52) in the SCT-conducted group without renal impairment and 0.8 (1.45) in the SCT not-conducted group without renal impairment (Table 1).

In the SCT-conducted group, before SCT, VD (53.9% of patients) was the commonly selected induction regimen in the 2003–2015 cohort, and VRD (44.3% of patients) in the 2016–2020 cohort, (Fig 2A and 2B). The duration of VD treatment tended to be short, from a median of 82 days (2003–2015) to a median of 53 days (2016–2020), and in recent cohorts, some patients were changed to VCD, VRD, and Rd treatment. VRD induction increased in recent cohorts, but it was confirmed that bortezomib and lenalidomide were discontinued in about 3% of patients, and many cases continued. The median duration of treatment was 99 days (S1 Fig in S1 File).

For post-SCT therapy, Rd was the most common regimen in the patients treated by post-SCT therapy, utilized by 50.3% of patients (2003–2015) and 49.5% of patients (2016–2020) (Fig 2C and 2D).

In the SCT not-conducted group, Vd was the most selected regimen in 39.9% of patients from 2003–2015 and 36.3% from 2016–2020. Additionally, MP in the 2003–2015 cohort (21.5%) and Rd in the 2016–2020 cohort (30.9%) were the standard regimens used in the SCT not-conducted group (Fig 3).

Regarding treatment duration in the overall SCT not-conducted group, the treatment duration was shorter in the 2016–2020 cohort than in the 2003–2015 cohort (Fig 3 and S2A Fig in S1 File). Treatment duration by age (>70 years and ≤70 years, Fig 4) analysis in the SCT not-conducted group demonstrated a similar trend in both the age cohorts and in >80 years (S2B Fig in S1 File).

Kaplan-Meier time to inpatient death curves per different lines of treatment in the SCT-conducted and not-conducted groups are available in Fig 5A and 5B. For the SCT-conducted group, the median time to inpatient death from the treatment initiation of 1L to 6L was 4043, not estimated (NE), 2531, 2363, 2013, and 880 days, respectively, in the 2003–2015 cohort. The median time to inpatient death from each treatment line’s initiation was NE for 1L to 5L and shorter in 6L as 519 days for 2016–2020 cohort than the 2003–2015 cohort. For the SCT not-conducted group, the median time to inpatient death from the treatment initiation of 1L to 6L was 1684, 1882, 986, 643, 599, and 406 days, respectively, in the 2013–2015 cohort. The median time to inpatient death from the treatment initiation of 1L to 3L was NE; for 4L, 5L, and 6L, it was 932, 486, and 412 days in the 2016–2020 cohort.

The current study evaluated treatment patterns and clinical outcomes in patients with MM using the MDV claims database in Japan. The claims data provided real-world information on the presence/absence of transplantation, initiation time-to-treatment, regimen selection by age, and the duration of treatment. Further investigations are required to characterize treatment patterns and outcomes in individuals who underwent SCT. Although it was difficult to clarify the characteristics of treatment changes in the era of new drugs, our study suggested an improvement in survival time for both the SCT-conducted and the SCT not-conducted groups over time. This study showed the improvement of treatment outcomes along with the recent diversification of treatment options.