REDS Banner Logo

The U.S. Linked Donor Component Recipient Databases (Vein-to-Vein Databases)

Background

The REDS-III domestic program has established for the first time a detailed research database infrastructure that links data from blood donors and their donations, the components made from these donations, and the recipients of these components; i.e., a particular donation can be traced through component production and, if transfused at a participating hospital, to a data extract from the electronic medical record of the transfusion recipient. These cumulative databases and the ability to link donor and recipient data enable investigators to address important issues in blood banking and transfusion medicine and to inform blood policy decisions.




Establishing a multicenter transfusion recipient database: Feasibility and potential use
Annual Meeting of the American Association of Blood Banks 2015

Triggers for Red Blood Cell Transfusion in the NHLBI Recipient Epidemiology and Donor Evaluation Study III (REDS-III) Transfusion Recipient Database
Annual Meeting of the American Association of Blood Banks 2015

Demographic and epidemiologic characterization of transfusion recipients from four U.S. regions: Evidence from the REDS-III Recipient Database
Karafin, M, Bruhn R, Westlake, M, Sullivan M, Bialkowski W, Gustaf E, Roubinian N, Hauser R, Kor D, Fleischmann D, Gottschall J, Murphy, E, Truilzi, D, and for the National Heart, Lung, and Blood Institute Recipient Epidemiology and Donor Evaluation Study-III (REDS-III) Transfusion. 2017; doi: 10.1111/trf.14370



The REDS-III Donor/Donation and Recipient databases are anticipated to be available as public use databases through the NHLBI Biologic Specimen and Data Repository Information Coordinating Center (BioLINCC) program (at https://biolincc.nhlbi.nih.gov/home/)  around mid-2018. The REDS-III Component public use database and the ability to link information from the donors and their donations, to data on the components made from these donations, and information on the transfusion recipients who receive these donations/components are anticipated to be available through BioLINCC around the spring of 2019.

Methodology

Transferring Data Files to the DCC


There are multiple options for delivering REDS-III data to the DCC. The REDS-III website supports the secure uploading of files. This is the preferred option for frequently uploaded data files and/or files that are not exceedingly large. For large data file submissions, the DCC can provide a secure account for transfer of files across secure file transfer protocol, or SFTP. The DCC can coordinate with users as needed to provide the access and the connection details. If the data file size is exceedingly large or the other upload methods are not practical, the data can be shipped on physical media (CD, DVD, etc.) to the DCC. In this case the data should be encrypted on the physical media. The decryption information should be sent separately, for example by email. Once the study data are received, the DCC will store and process the data as needed.

Management and Compilation of Data Files


Data files received by the DCC and the detailed information associated with the transfer action are registered within our tracking system. The data files then progress through a process of review prior to being integrated with additional data sets. Throughout the review process, the data file status is updated within our tracking system, recoding relevant comments associated with the review. Data files are imported into a review database that allow data management and statistical staff the ability to inspect and report back to the submitter the acceptability of the submission. Only after review by the DCC and report acknowledgement by the submitter have been completed, and any identified problems have been resolved, will the submission move forward and be included in the REDS-III Linked Donor Component Recipient Databases.

Quality Control of Data Files


As submitted data files are imported into a review database, initial structural checks are performed. The next phase of quality procedures includes running defined validation checks for data integrity and consistency with defined data definitions. The rigorous data integrity checks include confirming required fields and detecting duplicate records, invalid data types, and out of range values. Once a data submission has passed all integrity validations, any necessary standardization procedures, such as controlled vocabulary mapping, will be applied to harmonize the data before integration into the REDS-III Linked Donor Component Recipient Databases. If errors are encountered during this process, detailed output reports are sent to the submitter for investigation. Once corrections are submitted, the quality review process is re-executed. Once a data file has passed the integrity validation and been integrated into the REDS-III Linked Donor Component Recipient Databases, additional qualitative processes will be applied by the DCC analysts on a regular basis. Cross-field, cross-record and cross-table validations will be applied to ensure accuracy and completeness and to detect data anomalies. Periodic data quality reports summarizing data inconsistencies and completeness, and data monitoring reports summarizing overall status by various criteria, will be distributed, as well as any ad-hoc reporting required by stakeholders.

Lists of the variables included in the databases can be viewed here. The following are detailed codebooks for each database:

REDS-III_DonorDonationDatabase_2012-2016.xlsx
REDS-III_ComponentsDatabase_2013-2016.xlsx
REDS-III_RecipientDatabase_2012-2016.xlsx


Current Database Analyses

Examples of on-going analyses of the domestic databases include:

Linked Analyses


REDS-III is in the process of using the linked donor, component, and recipient databases to perform recipient outcome analyses. A major analysis that is currently underway is to determine whether a blood donor’s sex and/or age affects a transfusion recipient’s in-hospital mortality and/or length of hospital stay. As part of this analysis, the effect of sex mismatch (female donor blood transfused to a male recipient, or male donor blood transfused to a female recipient) and the effect of a female donor’s pregnancy history will also be examined. Additional analyses using the linked databases are under consideration: these include the effect of a donor’s blood donation frequency on recipient outcomes, and the effect of various donor factors on red blood cell alloantibody formation. Also under consideration is a more focused analysis that will examine outcome parameters for recipients of units from donors enrolled in the RBC-Omics protocol for whom measures of in-vitro osmotic hemolysis of their stored red blood cell units have been obtained.

Anticoagulants


The use of non-vitamin k oral anticoagulants (NOAC) is common among patients with atrial fibrillation, but little is known regarding how the use of these newer medications impacts clinical outcomes in patients with major bleeding. The Recipient database is being analyzed to describe inpatient hospitalizations among NOAC and warfarin users and to test the hypothesis that outcomes following major hemorrhage differ based on which medication the patient was taking prior to bleeding.

Alloimmunization


The Red Blood Cell Alloimmunization Working Group is investigating this relatively common and often clinically significant complication of transfusion and pregnancy, in multiple sites across the US. The group is using the Recipient Database to investigate patient/recipient variables associated with red blood cell alloimmunization, and is using the Donor Database to investigate red blood cell alloimmunization patterns in healthy blood donors.

Platelets


This analysis will define the epidemiology of platelet use at the REDS-III hospitals. The aim is to expand our current knowledge of clinical practice patterns and provide information to identify populations that would be most fruitful for future clinical studies of platelet transfusion practice.

Thrombosis


Previous studies have suggested that red blood cell transfusion is associated with risk of thrombosis in select patient populations. The Thrombosis Working Group is testing the hypothesis that red blood cell transfusion is associated with thrombotic risk in any hospitalized patient and will determine if the risk of thrombosis differs based on surgical procedures, medical conditions, or underlying genetic thrombophilias.


REDS-III HOME | ACCESSIBILITY | READERS & PLAYERS | SITE INDEX | OTHER SITES | PRIVACY STATEMENT | FOIA | CONTACT US | OIG

National Institutes of Health    Department of Health and Human Services    USA.gov