Using data envelopment analysis and the bootstrap method to evaluate organ transplantation efficiency in Brazil

• Published in: Health care management science 2021-09, Vol.24 (3), p.569-581
• Main Authors: Marinho, Alexandre, Araújo, Claudia Affonso Silva
• Format: Article
• Language: English
• Subjects: Bias; Blood & organ donations; Bootstrap method; Business and Management; Data envelopment analysis; Econometrics; Efficiency; Health Administration; Health Informatics; Management; Operations Research/Decision Theory; Transplants & implants
• ISSN: 1386-9620; 1572-9389
• DOI: 10.1007/s10729-021-09552-6

Brazil has the most extensive public program for organ transplantation in the world, and the Brazilian National Health System (SUS) provides full coverage of all costs involved in organ donation, transplants, and post-transplant. Despite the relevance of the subject and the shortage of organs for transplants, transplantation process efficiency assessments are still uncommon in Brazil and abroad. This study aims to evaluate the efficiency of the Brazilian states and the Federal District in transforming potential organ donors into actual donations. We applied data envelopment analysis (DEA) in conjunction with the bootstrap technique, using organ transplantation data from 2018. The bootstrap methods applied (bootstrap technique, the bootstrap-biased scores of efficiency, and the bootstrap bias–corrected scores of efficiency) allow to obtain a confidence interval for DEA scores and provide greater robustness to studies based on DEA methodology. The bootstrap bias–corrected model indicates that there is significant room for improvement in terms of converting potential donors into actual donors. The mean corrected score is 0.55, signalizing that altogether the Brazilian states could maximize in 45% the number of transplanted organs without necessarily increasing the pool of potential donors. The study provides insights into the Brazilian processes of organ donation and transplantation, helping to identify locations in need of resource allocation improvements. Given the scarcity of studies with a joint application of DEA and bootstrap techniques in this crucial health activity, we also intend to methodologically contribute to this type of benchmark analysis, emphasizing the importance of considering measurement errors, randomness, and bias at DEA models.