Raising awareness may increase the likelihood of hematopoietic stem cell donation: analysis of a nationwide survey using Artificial Intelligence

Background: In Italy, the demand for allogeneic transplantation exceeds the number of compatible donors in the Italian Bone Marrow Donor Registry (IBMDR). This study aimed to explore the knowledge, beliefs, opinions, values, and feelings of the Italian population regarding stem cell donation. Methods: An online survey was shared via social media. Respondents were retrospectively identified as registered on the IBMDR (donor group) or never registered (non-donor group). Statistical analyses confirmed the relationship between knowledge level and willingness to donate. Six machine learning classifiers were trained using questionnaire responses to predict the probability of IBMDR registration. Results: A total of 1518 respondents participated. Characteristics identified in the non-donor group were a lower level of knowledge regarding donation needs (51.7% vs 24.4%, p<0.001) and negative feelings such as fear (Z=- 2.2642, p=0.02), confusion (Z=4.4821, p<0.001), and uncertainty (Z=3.3425, p<0.001). Higher knowledge predicted a greater likelihood of IBMDR enrollment. Machine learning analysis showed an AUC ranging from 0.65 to 0.81, depending on the classifier. Conclusions: The results underscore the need to improve strategies to raise awareness and knowledge of stem cell donation among the Italian population.