Comparative Cost and Utilization Analysis of CAR-T Therapy and Autologous Stem Cell Transplantation for Diffuse Large B-Cell Lymphoma in Germany: Insights from Statutory Health Insurance Billing and Nationwide Inpatient Data

Abstract Purpose Chimeric Antigen Receptor (CAR) T-cell therapy has transformed the treatment of diffuse large B-cell lymphoma (DLBCL) but carries higher drug acquisition costs than autologous stem cell transplantation (autoSCT), while administration and follow-up costs remain underrepresented in economic evaluations and reimbursement decision-making. This study compares administration-related and subsequent treatment costs of CAR-T and autoSCT for DLBCL from the German statutory health insurance (SHI) perspective. Methods A two-step analysis based on two German databases was performed. In step 1, SHI billing data (representing 5 million insured individuals, ≙ 8% of SHI) from 2020–2022 were analyzed longitudinally. Costs were divided into initial regime (IR) (all services except CAR-T drug acquisition) and a following regime (FR). In step 2, a cross-sectional analysis of all German inpatient cases from 2020–2023 (17 million cases/year) assessed inpatient costs. Results Step 1 included 12 CAR-T and 59 autoSCT patients. Median IR costs per patient were €61,703 (CAR-T) versus €55,802 (autoSCT), while FR costs were lower for CAR-T (€44,832 versus €69,453). Most frequently reported adverse events (any grade) for CAR-T versus autoSCT were: B-cell aplasia/neutropenia (77% versus 64%), thrombocytopenia (38% versus 64%). Step 2 included 1,229 CAR-T and 2,239 autoSCT inpatient cases (2020–2023). Average LOS was comparable (27.3 versus 27.8 days), and average inpatient costs were lower for CAR-T (€13,991 versus €29,432). Conclusion Although associated with higher IR costs, CAR-T therapy in DLBCL demonstrated lower inpatient and FR costs compared to autoSCT. This analysis focused on administration and follow-up costs, excluding drug acquisition costs, characterizing the economic burden beyond drug pricing.