Background/Case Studies: Continuous, effective donor recruitment is critical for blood centers to avoid supply shortages and donation rates by younger individuals have declined in recent years. Moreover, beyond adequate overall supply, diverse donor recruitment is needed to serve specific groups, such as patients with sickle cell disease (SCD). Some legacy methods (e.g., voice phone calls) may have limited effectiveness in the digital age and innovations leveraging technology may continue to play a larger role. Incorporating curiosity-driven, smartphone-based blood typing could be an effective tool to bridge the gap between potential donors and blood centers. Such an approach could appeal to younger donors and facilitate diversified outreach. Building on previously reported technology for antibody detection, the aim of this study was to assess the feasibility of an instrument-free, artificial intelligence (AI)-enabled ABO-RhD blood typing kit. The kit is designed for self-testing and based on smartphone readout, thereby providing an avenue for engagement and connection with blood centers from the attendant application.
Study
Design/Methods: Monoclonal IgMs for anti-A, anti-B, and anti-D and a negative control were formulated and dried down on a paper card with QR codes for smartphone image capture. The kit procedure comprises reconstitution of the dried spot with tap water. After a fingerstick blood self-collection and capillary tube transfer to each spot, the card is mixed/tilted, and agglutination is captured by smartphone. Images are interpreted by AI which reports the blood group based on hemagglutination patterns and monitors background agglutination in the negative control. Feasibility studies conducted at present include accelerated stability testing of dried IgM formulations at 37 °C and evaluation of 10 deidentified patient samples which were typed by standard methods at a blood center.
Results/Findings: Monoclonal IgM formulations were identified that are stable for at least 14 weeks at 37 °C when stored dried down on a paper card sealed in a poly/foil bag. All 10 deidentified patient samples were correctly identified by the AI application as AB+. Conclusions: The preliminary feasibility of a smartphone-based ABO-RhD blood typing kit using AI to interpret hemagglutination patterns was assessed. Future feasibility studies include extension of stability, method comparison, robustness, usability, and evaluating donor engagement. Smartphone activities are coordinated through an application which, in the future, could provide seamless connection to existing blood center recruitment/scheduling applications. Use case scenarios include enhanced donor recruitment efforts for specific blood types or demographics and will be the subject of future studies.
Importance of research: Blood donor recruitment is a continuous burden on blood programs with fewer than 5% of eligible donors actively donating and about 30% of the blood supply comprising first time donations. A 2020 HHS Report to Congress (Adequacy of the National Blood Supply) and the 2022 NHLBI/OASH State of Science in Transfusion Medicine both cite donor recruitment and diversification of the blood supply as critical needs. New tools for donor recruitment and diversification are needed by blood programs.
