This study aims to describe concrete results, in a hospital, of changing a traceability process to an electronic data interchange (EDI) based model.This study aims to describe concrete results, in a hospital, of changing a traceability process to an electronic data interchange (EDI) based model.In November 2007, EDI was implemented between the blood bank (French Blood Establishment) and the University Hospital of Reims, concerning distribution data of blood products. We report the effects of this change on electronic traceability, after an 18-month follow-up.This study aims to describe concrete results, in a hospital, of changing a traceability process to an electronic data interchange (EDI) based model.In November 2007, EDI was implemented between the blood bank (French Blood Establishment) and the University Hospital of Reims, concerning distribution data of blood products. We report the effects of this change on electronic traceability, after an 18-month follow-up.Final traceability, after the haemovigilance service interventions, remains satisfying at 99.95 % after change. However, spontaneous traceability by clinical services fell brutally, and remains low at 86.1 % (versus 90.6 % before change). Although EDI concerns the informatic reception stage of traceability, both reception stage and final recording stage (transfusion, return to blood bank or destruction of products) were decreased. The process change itself lead to an increase of bad use dysfunctions, but human causes of omission type also increased at each stage. However, the majority of reception stage dysfunctions are technical causes, supported by EDI itself.This study aims to describe concrete results, in a hospital, of changing a traceability process to an electronic data interchange (EDI) based model.In November 2007, EDI was implemented between the blood bank (French Blood Establishment) and the University Hospital of Reims, concerning distribution data of blood products. We report the effects of this change on electronic traceability, after an 18-month follow-up.Final traceability, after the haemovigilance service interventions, remains satisfying at 99.95 % after change. However, spontaneous traceability by clinical services fell brutally, and remains low at 86.1 % (versus 90.6 % before change). Although EDI concerns the informatic reception stage of traceability, both reception stage and final recording stage (transfusion, return to blood bank or destruction of products) were decreased. The process change itself lead to an increase of bad use dysfunctions, but human causes of omission type also increased at each stage. However, the majority of reception stage dysfunctions are technical causes, supported by EDI itself.Although the new process was supposed to be more efficient and easier for the user, the global result on traceability by users is negative after 18 months.