Establishing a hospital transfusion management system promotes appropriate clinical use of human albumin in Japan: a nationwide retrospective study

Abstract

Abstract Background Despite international recommendations to establish hospital transfusion management systems to promote appropriate use of blood products, the general efficacy of establishing such systems has not been proven. This study aimed to validate the effect of establishing such systems for promoting the appropriate use of human albumin. Methods In this retrospective observational study, we used a Japanese Diagnosis Procedure Combination (DPC) database from fiscal year 2012 to 2016, which included inpatient records from approximately 1200 hospitals for payment processes in the national medical insurance system. From this existing database, containing approximately 8 million inpatient records per year, we selected patients with emergency due to “bleeding,” “sepsis,” and “burn injury,” by using the International Classification of Diseases and Injuries 10th revision (ICD-10) codes, and hospitals that had one or more patients for each disease group in each fiscal year. We conducted multivariable logistic regression analysis to estimate the relationship between human albumin administration and the state of the hospital transfusion management system. We evaluated temporal trends of mortality rate and length of stay as an indicator of care quality. Results Overall, 139,853 eligible patients admitted to 682 hospitals were selected. The results of the multivariable logistic regression analysis show that patients who were admitted to hospitals with an established hospital transfusion department introducing good practice criteria of blood products were less likely to be administered human albumin compared with those who were admitted to hospitals not introducing it, by approximately 30% for each of the three disease groups; adjusted odds ratios (95% confidential intervals) were 0.70 (0.59–0.83), 0.75 (0.69–0.81), and 0.71 (0.58–0.87) in the “bleeding,” “sepsis,” and “burn injury” groups, respectively. The temporal trends evaluation shows that there were no increasing trends of mortality rate and average length of stay against decreasing trends of human albumin administration in any disease groups. Conclusions Establishing a hospital transfusion department responsible for promoting appropriate clinical use of blood products could reduce human albumin administration for critically ill patients without loss of care quality. These findings provide support for policy makers and hospital managers to consider establishing such systems.