Introduction
In today’s rapidly changing world, artificial intelligence (AI) plays a critical role not only in science and communications, but also medicine. Transfusion medicine (TM) is one of the key areas that have a great potential to enhance people’s quality of life (Wang et al., 2020). In TM, AI can be used to strengthen the processes of procurement and manufacturing of blood transfusion, focusing on safety, justification for clinical application, and evidence-based decisions.
Examining the Present State of AI in TM
The key principle of AI in TM is emulating human intelligence in analyzing a patient’s need for blood transfusion and making a relevant decision. Examining patient blood management systems, AI contributes to a vein-to-vein digital footprint, which increases the safety and availability of the blood transfusion process (Levi et al., 2021). In particular, AI constructs artificial neural networks (ANN) that work on the collection of the daily mass of blood and corresponding data processing. Furthermore, raw data collection during the input layer is computed and analyzed in the output layer (Sibinga, 2021). In addition, the hidden layer refers to the so-called intermediate level, which characterizes the depth of the computerized learning (Sibinga, 2021). Accordingly, the safety of AI significantly depends on deep learning of the infrastructural environment.
The shortcomings of applying artificial intelligence in healthcare refer to limited data sets, overfitting due to using single-center data sets, and the difficulty of integrating clinical decisions in the process of blood transfusion prediction. According to Wang et al. (2020), the modern use of AI and machine learning is quite limited. For instance, one of the recent studies shows that they provide “limited information about preoperative blood ordering and intraoperative hemorrhage risk” in cardiothoracic surgery (Wang et al., 2020, p. 2). To predict the necessity of blood transfusion, artificial intelligence algorithms process preoperative variables, but the results are still not accurate enough to ensure coping with possible intraoperative hemorrhage (Wang et al., 2020). Nevertheless, the authors suggest that the future of AI is promising as a prediction method, while more studies are critical to validate current data.
Elaborating on the Perspectives of AI in TM
Speaking about the potential of AI and machine learning regarding blood transfusion, one should focus on the safety of implementing these technologies. Sibinga (2020) states that there is a need to develop consistent quality data bases through big data processes to integrate deep learning and algorithms. The mentioned author explains that it will improve vein-to-vein manufacture as well as the clinical application of blood transfusion (Sibinga, 2020). Accordingly, a more sophisticated approach to TM will allow the optimal use of procedures, equipment, resources, and human capital (Davenport & Kalakota, 2019). Credible data bases will be beneficial for advancing alert systems to identify any deviations in patients’ health indicators.
To ensure personalized blood transfusion and meet the needs of patients, healthcare providers should be educated to apply AI in their practice. To provide rationality to blood transfusion, they should be aware of how AI principles assess the needs of patients, which means that care providers cannot fully rely on the decisions made by technologies (Davenport & Kalakota, 2019). On the one hand, machine learning and AI are able to determine blood types, time required for transfusion, and the volume of the blood needed for a certain patient (Levi et al., 2021). On the other hand, research shows that individualized care tends to be a priority, which can be achieved only if care providers will make the final decisions (Sibinga, 2020). Informatics specialists and clinicians working in an effective alliance should control patient blood management systems.
In the future, transfusion medicine assigns a top priority to a digital foot printing, concentrating on the restrictive therapeutic guidelines, thus making sure that all blood transfusion cases are evidence-based. Sibinga (2021) emphasizes that alternative therapies should be taken into account to minimize risks for patients. The World Health Organization (WHO) calls for the creation of the integrated healthcare delivery that will allow lower costs and higher patient satisfaction (Sibinga, 2021). In turn, the 2030 UN Universal Health Coverage (UHC) goals include the points about the safety of the digital footprint and making the processes of blood transfusion lean (Sibinga, 2021). It is expected that automated and advanced AI and machine learning algorithms will predict blood ordering, and related logistics, minimizing human errors and technical issues.
Conclusion
To conclude, AI and machine learning provide promising perspectives for improving both the effectiveness and safety of blood transfusion and related patient satisfaction. Today, these technologies are able to determine the needs of a certain patient by constantly collecting blood data, processing it, and making suggestions about the volume, type, and other parameters of blood. In the future, there is a need to synthesize AI decisions with the efforts of care providers since only their integrated work can make the decisions regarding blood transfusion relevant, accurate, and timely. The adoption of AI technologies in the area of blood transfusion remains one of the key problems to be resolved.
References
Davenport, T., & Kalakota, R. (2019). The potential for artificial intelligence in healthcare. Future Healthcare Journal, 6(2), 94-98.
Levi, R., Carli, F., Arévalo, A. R., Altinel, Y., Stein, D. J., Naldini, M. M., & Celi, L. A. (2021). Artificial intelligence-based prediction of transfusion in the intensive care unit in patients with gastrointestinal bleeding. BMJ Health & Care Informatics, 28(1), 1-8.
Sibinga, C. T. S. (2020). Artificial intelligence in transfusion medicine and its impact on the quality concept. Transfusion and Apheresis Science, 59(6), 1-3.
Sibinga, C. T. S. (2021). Artificial intelligence and the future of transfusion medicine. Neuroscience, 2(2), 25-30.
Wang, Z., Zhe, S., Zimmerman, J., Morrisey, C., Tonna, J. E., Sharma, V., & Metcalf, R. A. (2022). Development and validation of a machine learning method to predict intraoperative red blood cell transfusions in cardiothoracic surgery. Scientific Reports, 12(1), 1-9.