Canadian donations and transfusion database (<scp>CANDAT</scp>): From blood donors to transfusion recipients

One in seven patients in hospital will require a blood transfusion.1 Transfusions are essential supportive treatments for all aspects of medical care, including surgery, cancer, critical care, obstetrics, and chronic diseases.2 A safe and sustainable blood supply is a prerequisite of modern health care delivery; however, this tenet is being challenged by an aging donor pool, devolving donor behavior, and a growing demand for blood due to increasingly complex medical and surgical therapies.3 In Canada, blood donation records are currently maintained by two blood operators (Héma-Québec serving Québec, Canadian Blood Services serving the rest of Canada); blood utilization and patient data are collected and stored at healthcare institutions, and health delivery data (e.g. physician billings) is held by publicly funded, province-specific health services databases. Canada lacks a comprehensive national database for blood product utilization which links blood donors, products, and recipients, and these data primarily exist in siloed repositories. This gap in comprehensive transfusion data coverage undermines the safety and sustainability of blood because, without it, trends in blood use, impending shortages, and emerging threats to the blood system cannot be effectively studied or addressed.4 The blood supply faces three main threats: infectious complications, non-infectious complications, and supply shortages. In the 1980s, thousands of blood transfusion recipients became infected with HIV and hepatitis in Canada and across the world.5 Since then, strategies to avoid infectious complications have been implemented, including donor screening, donor testing, pathogen reduction, and hemovigilance techniques. There has also been a heightened awareness of non-infectious risks of blood transfusion, including transfusion-associated circulatory overload, hemolytic transfusion reactions, transfusion-related acute lung injury, and transfusion-related immune modulation; these complications are more difficult to track. Blood shortages continue to occur in Canada with increasing frequency, and the fragility of the blood system has been evident during the COVID-19 pandemic and recent environmental disasters.6, 7 A comprehensive vein-to-vein database that can track the movement of blood products can enable monitoring of transfusion usage, identification of complications, and inventory audits during shortages, allowing data-driven policy and decision-making to address emerging threats to the blood supply. To address the need for a comprehensive blood transfusion database in Canada, we held a one-day in-person meeting with Transfusion Medicine experts, clinical and data scientists, database curators and custodians, partners and consumers of blood transfusion data, where we outlined a roadmap for the creation of CANDAT to link blood donor, blood product, and blood transfusion recipient data to enable research, uphold safety, and ensure sustainability of the blood supply in Canada. This report summarizes the output from that meeting. In Canada, contingency plans for blood shortages are the responsibility of the National Emergency Blood Management Committee (NEBMC) with cooperation from hospital transfusion medicine services, provincial Ministries of Health, and the blood suppliers.3 The NEBMC developed a plan to coordinate equitable access to blood, which informs inventory management and hospital-level transfusion policies during times of shortage.6 To support implementation and adjudicate access, all transfusion medicine services across jurisdictions are required to report inventory daily. However, the principal limitation of this approach is the lack of complete data, evident during the intermittent shortages of platelets and group O red blood cells (RBC) in 2023. The current system relies on manual reporting by hospital blood bank staff, making inventory reporting difficult. A transparent, high-velocity linked dataset will improve the timeliness and accuracy of provincial and national blood inventories to inform short-and long-term demand forecasting. Linked datasets with optimized data streams will also improve planning and inform measures to prevent blood shortages. Several existing linked databases in Canada and other countries can provide important lessons for building CANDAT. Canadian databases can be directly leveraged to accelerate the development of CANDAT linkages. International examples of vein-to-vein databases provide valuable roadmaps and serve as potential collaborators for multi-national big data studies in transfusion medicine. GEMINI, which was originally an initiative focused on studying the quality of care for general medicine inpatients,8 is a growing collaborative pooling data for hospitalized patients including all medical and intensive care admissions at over 35 large hospitals in Ontario, covering 60% of the province's patients with over 2.5 million hospital admissions. Hospitals provide data from their internal information systems to the GEMINI database, where data quality is assessed through computational and manual validation,9 and data variables are standardized to common data models. GEMINI data have been used to measure the quality of care at the physician and hospital level,10 and support annual quality reports organized through the provincial health administration agency, Ontario Health.11 GEMINI collects blood transfusion data and is currently collaborating with researchers to examine the variability in transfusion practices across participating institutions. In addition to being an important source of data, the GEMINI experience can help inform the creation of CANDAT. GEMINI is an example of a clinical dataset that can be shared across healthcare institutions for research with robust oversight and privacy regulations. The GEMINI experience highlights the importance of incremental and rigorous approaches to data validation and quality assessment such as combining computation and manual assessments of data quality.9, 12 GEMINI has also demonstrated that clinical data can hold tremendous potential for quality measurement and reporting back to institutions and clinicians. Through collaboration with CANDAT, GEMINI can help enable large-scale linkages across blood donor and recipient databases, with the aim of understanding and improving the use of the blood supply in Canada. Founded in 1992, ICES is a publicly funded research institute with a repository of coded and linked longitudinal healthcare data sets, which encompass most publicly funded health services for the population of Ontario. ICES currently holds health services records for residents in the province of Ontario (population of 16.1 million) with linked datasets including physician billings, medical claims, long-term care, laboratory results, cancer therapy, stroke registry, demographic surveys, and other data on patients eligible for publicly funded healthcare in Ontario.13 The mission of ICES is to support research and analytics that translate data into trusted evidence that makes policy and healthcare better and people healthier. The longevity and comprehensiveness of ICES databases have allowed for large-scale studies and analyses in health services research. Other strengths include the availability of long-term outcomes data, diverse data-sharing agreements, strong privacy and security standards, project-specific data linkages with over 100 data holdings, large participating research programs, and established Indigenous data governance agreements that support Indigenous data sovereignty. Researchers from across Canada can access ICES data and analytic resources remotely. Lessons learned from the ICES experience can help CANDAT establish a national transfusion database. Firstly, CANDAT should leverage ICES and other provincial and national data repositories to avoid duplication. Second, CANDAT should identify specific data elements required for transfusion research and linkage so that these can be built or added to existing datasets. Third, CANDAT should aim to use existing data to conduct desired analyses and identify limitations, and document where gaps exist. The Health Data Research Network (HDRN) is a pan-Canadian network of organizations that either hold linkable health and health-related data for entire populations or have mandates and roles relating directly to the access or use of those data. ICES and other similar organizations across the country are members of HDRN, as are the Canadian Institute for Health Information and Statistics Canada. The goal of HDRN is to ensure that multi-regional data are used to improve health and health equity for Canadians. Challenges inherent in building HDRN Canada include variation in data collection and infrastructure across provinces, the need for strong privacy safeguards, the value of engaging smaller provinces for diversity, and the need for a strategy that acknowledges different data capabilities across data holders. Leveraging the HDRN experience, early opportunities for CANDAT include advocating for change in data management such as the use of common practices in data collection and coding; the use of existing resources and layering additional needs on those; building initiatives with a view toward sustainability; and focusing on the implementation of research findings to improve healthcare delivery. All blood products and components in Canada are supplied by either Héma-Québec (HQ, supplying the province of Québec) or Canadian Blood Services (CBS, supplying all other provinces and territories). Both CBS and HQ collect and analyze donation data for operational, research, and surveillance purposes. CBS data include donations, deferrals, donation reactions, test results (e.g. phenotyping and genotyping), and donor demographics with over 30 million donor records from 1987 onward. Past studies have evaluated donors (e.g. donation risk, donor health, donation frequency, genotypic prevalence) and recipient risks (e.g. transmissible disease risk).14-16 Infectious disease surveillance estimates have been used by physicians to counsel recipients about infection risks.17, 18 Data are also used for evaluating the impact of donor policies, such as the reduction in time deferral for gay, bisexual, and other men who have sex with men.19 CBS is also currently developing a program to systematically collect donor data for health systems research and surveillance. Similar data are collected by HQ from nearly 2 million donors and over 10 million donations in the province of Quebec. In addition, HQ also maintains specialized data including a plasma biobank since the COVID-19 pandemic, a genotype database for all donors (including ABO, Rh, and phenotypes C, c, E, e, and Kell), and systematic screening data on all donors for Hepatitis A and Parvovirus B19.20 CBS and HQ have also collaborated on nation-wide projects such as the impact of changes to sexual risk behavior donor screening and the national disease seroprevalence of emerging infectious diseases in Canada.21 Both CBS and HQ have previously provided data for large-scale data linkage studies. Project-specific linkages of CBS variables (e.g. donor sex) with transfusion recipient data have allowed the study of donor characteristics impact on patient outcomes.22 HQ donor data have also been linked with Québec's universal health insurance registry to study long-term donor health outcomes such as the risk of iron deficiency among donors and the general population.23 Using blood donor data and insurance registry numbers to link the provincial infection and vaccination registry, HQ estimated the seroprevalence of anti-SARS-CoV-2 in Québec over the first three years of the pandemic. CANDAT is aimed at further expanding the linkage and usage of blood operator datasets for vein-to-vein research.24 Models of comprehensive transfusion medicine databases in other countries can help inform the creation of CANDAT and provide a structure for standardized data collection. The Swedish Donations and Transfusions database (SCANDAT) has nearly 40 years of data, including over 1 million donors and several million recipients with complete blood product traceability for several hundred thousand donor-recipient chains.25, 26 These data have allowed SCANDAT to study donor health to improve donation and blood transfusion safety, donor recruitment through refinement of eligibility criteria, the efficacy of transfusions for patients, activity-based costing for transfusion, forecasts for blood utilization, and artificial intelligence analysis for quality improvement. In North America, the Recipient Epidemiology and Donor Evaluation database (REDS-IV-P), funded by the National Institutes of Health in the United States and Brazil, has over 30 years of comprehensive transfusion data from 22 hospitals and 7 blood centers representing 1.3 million patients.27, 28 Projects have examined donor genetics and recipient outcomes, HIV transmission among donors, pediatric and maternal transfusion, and blood utilization in patients with sickle cell disease. Other large transfusion datasets include the Danish Blood Donor Study (DBDS), the Dutch Transfusion Data warehouse (DTD), the National Transfusion Dataset (NTD) in Australia, and the emerging Pan-European Transfusion Research Infrastructure (PETRA) in the United Kingdom with efforts toward Europe-wide linked transfusion databases.29-32 Current limitations of these databases include a lack of granular transfusion-related detail (e.g. vital signs, volume of transfused units), missing treatments (e.g. tranexamic acid), incomplete product characteristics (e.g. complete serologic profile), and details of transfusion reactions. Additional ongoing challenges are inconsistencies in coding across institutions, the need for validation of evolving healthcare technologies, complex data-sharing agreements, variable medico-legal frameworks, sustainable funding sources, emergent cybersecurity threats, and a lack of technological infrastructure for expanding data storage. Similar hurdles will need to be considered as CANDAT is established. The vision, aims, enablers, and action steps for CANDAT were refined using methods to facilitate the transition from a shared vision to specific goals.33, 34 The aims of CANDAT are: (1) To create a comprehensive database that links data on blood donors, blood components, and blood transfusion recipients; (2) to improve the safety and sustainability of the blood system; and (3) to enable research in transfusion medicine and related areas of study. The vision of CANDAT is to create a Canadian national database that links blood donors and blood transfusion recipients in real time to uphold safe, sustainable, efficient, and equitable transfusion practices. Through multidisciplinary collaborations, CANDAT will improve the understanding of transfusion-related patient outcomes (e.g. transfusion-associated adverse events), inform blood transfusion policy (e.g. avert shortages), and advance transfusion medicine research. Data for CANDAT will include blood donor and product data from the blood suppliers, transfusion recipient features, epidemiological data, financial cost information, patient registries, environmental variables, and metadata. Essential quality elements of any dataset are accuracy and internal and external validity.35 Developing a new database requires a careful and dedicated approach to data quality requiring significant upfront investment, though approaches can be increasingly automated over time.15 The use of quality indicators, data monitoring boards, and benchmarks for ongoing assessment are essential for maintaining data integrity. Transparency in the methodology used for data collection, computation, and sharing with thorough cybersecurity measures will mitigate data breaches. While comprehensive data linkages can enable scientific advancements, they increase exposure to risks related to privacy breaches and bring forth unique ethical considerations. CANDAT will require a well-defined governance framework and operational structure to support a centralized research ethics board, ensuring a streamlined yet rigorous ethical oversight process to facilitate both compliance and efficiency in research. Rather than establishing a new independent database, CANDAT can leverage existing databases by advocating for more detailed transfusion-related variables an direct linkages. These databases have been interlinked in the past on a per-project basis. For instance, public health data (e.g. collected by the Canadian Institute for Health Information), hospitalized patient electronic data (e.g. collated through GEMINI), and Canadian Blood Services donor data have all previously been linked with provincial administrative health services databases (e.g. ICES) to perform studies related to the COVID-19 pandemic. Such linkages have demonstrated feasibility through deterministic methods using shared universal health insurance numbers across healthcare databases (e.g. OHIP, or Ontario Health Insurance Plan number) within each province, and through probabilistic linkages using patient and donor details where this is not feasible.36, 37 As CANDAT expands, additional data can be linked incrementally through feedback, validation, and growth cycles driven by ongoing research needs and funding. Engaging with policymakers and funding agencies is essential to ensure fiscal sustainability of CANDAT. These efforts entail navigating and aligning CANDAT research objectives with the priorities of funders wherever possible. Given the importance of blood transfusion sustainability across the entire country, an approach that includes all provinces and territories is essential for an integrated national database. With this objective in mind, CANDAT will need to transcend institutional and regional boundaries, and leadership will need to build relationships ideally formalized through evergreen data agreements. Furthermore, advances in data management, including federated data platforms, represent an important technology to promote a culture of open science, information exchange, and national collaboration. Executive planning will engage stakeholders and emphasize key milestones. Strategic partnerships across local, provincial, and national institutions, including governmental agencies, will enable the creation of the database, and robust legal frameworks and cybersecurity standards will safeguard privacy and build trust. Research capacity will be fostered through training and mentorship. Barriers to the creation and implementation of CANDAT are the potential for data fragmentation, funding limitations, incomplete data collection, inconsistent coding practices, and ethical-legal risks in data linkages across provinces. Table 1 summarizes requirements for the successful creation and continuity of CANDAT in the short- and long-term. A demonstration project will show feasibility. We selected a project that will track immunoglobulin usage across provinces, as this is a priority area for Canada. Subsequent steps will be to identify other high-priority areas of research, secure funding, and create core governance teams. Table 2 provides strategies and specific actions toward the development and implementation of CANDAT linkages. Create a white paper outlining the objectives for CANDAT Develop a plan for a proof of concept Identify support staff Consider needed expertise Determine leadership Create terms of reference Perform a needs assessment and identify priorities for CANDAT Create a roadmap the of data stakeholders in the planning and of the demonstration project of is a blood product that is used for patients who either lack or require for immune The use of products to increase at an annual of In Canada, products for nearly a of the on blood components and products in to plasma collection during the COVID-19 pandemic in supply. In we are developing a linked database combining electronic health data and clinical information in Ontario, and from to This initiative will understanding of and to products are Using data we aim to compliance with provincial identify patients with utilization or and to inform product demand and supply The project will leverage two data inventory and recipient transfusion and which key information and patient not in at each will be linked using unique patient for comprehensive of clinical outcomes, and of data robust data and data this project will vein-to-vein data can improve system efficiency and help the Canadian blood system better for product demand and potential shortages. Data blood donors, blood products, and blood transfusion recipients is currently in Canada. CANDAT, a national database for blood transfusion, will help the safety and sustainability of the blood supply through comprehensive health data that blood donors, blood components, and blood transfusion and provided into contingency planning for and provided into the GEMINI database. provided into the ICES database. provided into the Health Data Research Network and provided into blood transfusion datasets and the that priorities for CANDAT. We and for their during the CANDAT meeting. This was funded by support from Canadian Blood Services and the Canadian Institute for Health The funding Canadian Blood Services provided support toward the and conduct of and CBS in the or of the have financial or relationships or that the and on this The study not patient data. All from data were to sharing data collected during and are on this data are