| Literature DB >> 11967134 |
K Titlestad1, T Kristensen, J Jorgensen, J Georgsen.
Abstract
As the description of transfusion practice over time requires the handling of massive amounts of data, and because the majority of records have until now been paper-based, little is known about the epidemiology of blood transfusion. Furthermore, most previous studies have collected data manually from patient records or from computerized registers created solely for study purposes. We describe a procedure for monitoring transfusion practices based exclusively upon the collection of data from the existing computerized registers. During the calendar years 1997 and 1998, more than 750 000 data records were collected from the blood transfusion registers, diagnosis and procedure registers, and clinical biochemistry registers at two university hospitals. Linking of the data required a unique and consistent identification of all data records. This was made possible by the national Personal Identification Number. In this way more than 98% of all transfusions could be identified and approximately 95% (77 950) could be connected to a principal diagnosis. The described procedure creates new opportunities for studies of transfusion practices both on a large scale and in richness of detail, and we propose to establish a network where transfusion practices data are available and comparable for individual departments, hospitals, whole regions, and possibly nations.Entities:
Mesh:
Year: 2002 PMID: 11967134 DOI: 10.1046/j.1365-3148.2002.00352.x
Source DB: PubMed Journal: Transfus Med ISSN: 0958-7578 Impact factor: 2.019